Abstract:Energy saving and emission reduction is the key to improve quality and efficiency, transformation and upgrading of intelligent agricultural machinery and equipment. Hybrid tractor is the key model for the development of the global new energy agricultural machinery and equipment industry, and it is also the key technology path to accelerate the transformation of Chinese agricultural machinery to high-end, intelligent and green. In recent years, domestic and foreign combined with local agricultural production characteristics, market demand-oriented, continued to spawn, iterative new technology and new products, and through multi-scene, multi-level demonstration and application, a complete technology research and development and application system was gradually formed, which accelerated the promotion of product maturity. On the basis of outlining the development history and development characteristics of hybrid tractors, the development status of key technologies in this field was summarized and analyzed from four aspects: design and development, power control, intelligence and test verification, which focused on analyzing the technological differences between domestic and foreign models, and the key issues restricting the development of the field were pointed out. Finally, under the background of policy guidance and technological innovation synergy, it looked forward to the development direction of the hybrid tractor industry from the perspective of agricultural modernization empowered by artificial intelligence, big data and other emerging technologies, and puted forward its future development trend in the optimized design of the whole machine, collaborative optimization and control, special performance test equipment, intelligent operation control and other aspects.

Abstract:The tractor serving as the core power equipment for achieving agricultural modernization and efficient production, has a technological development level that directly influences the efficiency, cost, and sustainability of agricultural operations. The industry is rapidly evolving toward high-end, intelligent, and green directions, with new power transmission technologies becoming the most critical and dynamic area of innovation in this transformation. A comprehensive review of the current research status of new power transmission technologies for tractors, both domestically and internationally was provided, which systematically addressed the field from three dimensions: theoretical research on new transmission systems, development of key technologies, and industrial applications. It was focused on diverse technical pathways, including power shift transmission, hydro-mechanical continuously variable transmission, pure electric drive systems, as well as diesel-electric hybrid, hydrogen hybrid, and methane hybrid power systems. It delved into the basic structure, working principles, and performance characteristics of these power transmission systems, and the development status of key technologies was analyzed by examining typical tractor products from China and other countries, along with their practical application outcomes and productization paths. The critical bottleneck issues that needed to be overcome in current new power transmission technologies were generalized and summarized, primarily including battery capacity and charging rates, control precision and stability, environmental adaptability of electric drive systems, multi-energy source management, and the production and storage of new energy sources. Finally, it outlined future development trends for new power transmission systems in tractors, proposing deep integration in three major directions: global energy efficiency utilization, full-scenario self-adaptive compatibility, and modular multi-functional expansion.

Abstract:The development of new energy power equipment for agriculture is of great significance for promoting the green and low-carbon transformation of agriculture, ensuring national food security and achieving agricultural modernization. A systematic review of domestic and international research progress in new power equipment for agricultural was provided. It focused on analyzing the current status and development trends of four major categories of equipment: new energy tractors, new energy micro-tiller, new energy work platforms, and new energy agricultural low-altitude aircraft. In the field of new energy agricultural machinery, tractors are developing in parallel toward pure electric, hybrid, and clean fuel options;new energy micro-tiller are evolving from “pure electric dominance” to “electric and hybrid parallel development and intelligent self-propulsion”. New energy work platforms are advancing along a technical path characterized by “electric drive as the main power source, supplemented by hydrogen/biofuels, and multi-form coordination of wheels, tracks, and wheeled legs”. New energy agricultural low-altitude aircraft are developing along a technical route featuring “electric drive as the main power source, hydrogen energy for extended range”. However, the new energy agricultural equipment industry is currently facing challenges such as inadequate adaptation of core technologies, an incomplete standard testing system, weak industrial chain coordination, lagging infrastructure, and an incomplete market service system.To this end, it was proposed that future efforts should be coordinated across multiple fronts: top-level policy design and standard testing, breakthroughs in key “three-electric” technologies and specialized chassis systems, industrial chain collaboration and pilot-scale validation, energy-supply infrastructure such as fast charging, battery swapping, and hydrogen fuel, and the completion of scenario-based demonstration and service systems. It was suggested to enhance the adaptability and reliability of new energy agricultural machinery to promote the application of electric agricultural machinery.

Abstract:Tractors are the most widely used power machinery in agricultural production. Conducting durability simulations of tractors under real-world operating conditions through laboratory test benches is of great significance for improving their reliability. A four-column wheel coupling test bench was utilized to explore the whole-machine durability testing methodology for tractors. An operational load acquisition system was established to collect load data under actual working conditions. Accelerated life testing was conducted by using three different identification methods: the time-domain damage retention identification method, the time-frequency analysis-based cumulative power spectral density identification method, and the wavelet transform-based envelope identification method. The effectiveness of these acceleration methods was validated through wheel coupling bench testing. The results indicated that, within the application scenarios and iterative approaches adopted, the time-frequency analysis-based cumulative power spectral density identification method demonstrated superior performance in signal compression, achieving a reduction ratio of 36.37%. The drive load spectra for the wheel coupling test bench were iteratively derived from both the original signal and the three accelerated signals. The loading results demonstrated that when the chassis and cab were selected as test targets, the damage reproduction tendency at designated measurement points was effectively achieved. Under the same test conditions but with different acceleration identification methods, the time-frequency analysis-based cumulative power spectral density identification method showed superior consistency in reproducing chassis damage, while the time-domain damage retention identification method performed better in replicating cab damage.

Abstract:In response to the issue that traditional communication network cabling methods are unable to meet the requirements of high efficiency, stability and functional expansion for information transmission in autonomous hydrogen fuel cell tractors, a multi-rate control area network (CAN) bus architecture was proposed based on the SAE J1939 protocol and ISO 11783 standard. Firstly, the data flow requirements for sensing, positioning, drive control, etc. were analyzed in combination with the autonomous operation process in the field. Then, using the principle of minimum complexity in graph theory, multiple topologies were compared and a gateway plus three-bus structure was determined. The CAN communication network nodes were divided based on the structure of the autonomous operation hydrogen fuel cell tractor and its working conditions, and the functions of each node were defined. For common network topologies, a mathematical model was established based on graph theory, and the total information volume per second N and the node information volume per second Z of each structure were analyzed and compared. A high-speed and low-speed dual-bus structure was adopted to design the communication network of the autonomous operation hydrogen fuel cell tractor, with bus rates selected as 500kb/s and 250kb/s. The communication message and system software design were developed based on the SAE J1939 protocol and ISO 11783 standard. The hardware functions and communication effects of the communication system were tested by using CANoe software and verified on the whole vehicle. The test results of the bus load rate were 28.22%, 19.20% and 22.85%, all below the 50% requirement. These results verified the feasibility and provided a feasible solution for the design of the communication network of autonomous operation hydrogen fuel cell tractors.

Abstract:Hilly and mountainous areas are important grain production bases in China. The farmlands in these regions are characterized by small plot sizes, steep slopes, and narrow roads. Traditional fuel-driven chassis have disadvantages such as slow power response, inflexible steering, and low levels of intelligence. To address these issues, an electric intelligent chassis suitable for light and simple rice and wheat harvesters was designed. This machine adopted a distributed design to achieve flexible steering and a wide track design to adapt to rice harvesting operations in hilly and mountainous areas. Motion condition analysis was conducted, and the parameter design of the electric drive system and battery system was completed, solving the problem of slow power response of fuel-powered chassis. The software and hardware design of the motion control system was carried out, enabling remote control operation or autonomous navigation of the chassis. Performance tests and operation parameter optimization tests were conducted. Taking the harvesting speed, threshing drum speed, and cleaning fan speed as test factors, and the breakage rate, impurity rate, and loss rate as evaluation indicators, a Box-Behnken test was designed. The results of the Box-Behnken test were analyzed by variance, and a regression model between the evaluation indicators and the test factors was established. With the goal of minimizing the breakage rate, impurity rate, and loss rate, the optimal solution for the harvesting speed, threshing drum speed, and cleaning fan speed was obtained: the harvester’s harvesting speed was 0.8m/s, the threshing drum speed was 1000r/min, and the cleaning fan speed was 2786.15r/min. Field tests showed that under this condition, the breakage rate was 0.42%, the impurity rate was 4.05%, and the loss rate was 0.87%. The machine operated smoothly during the operation period, and the operation indicators met the standards for mechanized rice and wheat harvesting. The test results indicated that the designed electric intelligent chassis could meet the operational requirements of the harvester.

Abstract:Aiming to address issues such as low energy utilization efficiency and insufficient speed stability in electric tractor traction operations, an energy-saving control method was proposed based on a four-motor dual-coupled power system (FDPS). The FDPS combined the characteristics of distributed drive for front and rear axles with coupled drive for single axles, supporting single-motor, dual-motor, triple-motor, and quad-motor drive modes. A longitudinal dynamics model for the working unit and a tire-soil interaction model were established, along with motor and transmission system models that characterized the efficiency and dynamic properties of the FDPS. A multi-drive-mode energy-saving control architecture matching traction workload was proposed. The upper layer employed a hybrid penalty function particle swarm algorithm to offline generate FDPS drive mode switching and multi-motor optimal torque distribution rules based on varying demand driving forces and operating speeds. The lower layer integrated a fuzzy PID algorithm to online identify the machine’s demand driving force with precision while tracking target operating speeds. A 900-second energy-saving control bench test was conducted by using field-measured traction resistance data. Test results demonstrated that the proposed method reduced energy consumption by 3.76% and lowered the root mean square speed error by 11.1% compared with the average torque distribution control strategy, enhancing the machine’s operational economy and speed stability. The proposed power system can serve as a universal power platform, and the energy-saving control method can provide a reference for multi-power-source drive control research in electric tractors.

Abstract:Aiming to investigate the impact of battery capacity configuration on system efficiency and life cycle cost (LCC) for high-horsepower hybrid tractors, thereby addressing the research gap in battery capacity optimization design for hybrid tractors, taking a 190kW series hybrid tractor as the research object, an LCC evaluation system encompassing equipment procurement, battery replacement, and fuel consumption was established. A two-level battery capacity optimization method combining enumeration and dynamic programming (DP) was proposed: the outer layer investigated the influence of battery capacity on LCC under time-varying weights of energy and battery prices through global search via enumeration method, while the inner layer employed DP algorithm for optimal control of energy management strategies in hybrid tractors. A tractor simulation platform was developed by using Matlab/Simulink to quantitatively analyze the coupling mechanism between battery capacity and energy management strategies on system efficiency throughout the tractor’s lifecycle. Simulation results indicated that within the 2.2~110A·h battery capacity range, increasing battery capacity benefited fuel consumption reduction and battery lifespan extension;the 24.2A·h configuration represented the fuel consumption inflection point, beyond which capacity increase yielded minimal efficiency improvement;when battery capacity reached 52.8A·h, no replacement was required based on cycle life, but the nonlinear battery capacity-LCC relationship showed 22A·h configuration minimized LCC to 3.9093 million yuan. The proposed “technical parameters-life cycle cost” optimization framework established theoretical paradigms and provided valuable references for hybrid system design in agricultural machinery.

Abstract:Aiming at the problems of poor terrain adaptability, inflexible movement and insufficient environmental friendliness of the current power equipment, a four-wheel multi-mode steering electric unmanned agricultural power equipment with an articulated frame and an independently driven steering wheel set was proposed to meet the operational needs of sowing, tillage, plant protection, transportation and other links in facility horticultural agricultural production. Firstly, the overall structure of the power equipment consisted of an articulated frame, an independently driven steering wheel set, a central suspension device and an electronic control system. It supported multiple steering modes, including two-wheel steering, four-wheel steering, on-the-spot steering and crab steering. Next, matching calculations were performed on key components of the power system, such as the drive motor, steering motor, PTO motor, and power battery. Subsequently, the dynamic motion characteristics of the agricultural power equipment under various steering modes were analyzed, and a multi-mode dynamic model was established. Finally, a hierarchical drive control method was applied in the mechatronic control system, and multiple field tests were carried out with driving speed, steering performance, climbing performance, stability of tillage depth, and endurance performance as evaluation indices. The test results showed that the equipment achieved an average driving speed of 9.67km/h, the minimum turning radius of 1715.5mm under four-wheel Ackerman steering, the maximum climbing grade was 15°, the coefficient of the tillage depth stability was 94.62％, and the battery runtime was 2.18~2.77h. The equipment demonstrated excellent steering characteristics, passability, and operational stability, providing theoretical and technical support for the design of advanced small- and medium-sized intelligent agricultural machinery equipment.

Abstract:In allusion to the efficient vibration energy recovery problem of a wheeled electric tractor, a variable stiffness spring was innovatively integrates into the research team’s existing vibration energy feedback device and a method was proposed to match the device’s resonance frequency with the road surface excitation frequency, thereby improving energy capture performance. Firstly, the range of road excitation frequency was analyzed based on the road grade and vehicle speed of the electric tractor. Subsequently, a nonlinear 10-degree-of-freedom dynamic model of the variable stiffness vibration energy feedback device was formulated. With the QR decomposition, implicit restart Arnoldi method (IRAM) and adaptive cross approximation algorithm (ACA), the modal frequencies of the device were solved, and the modal evolution law was analyzed by combining these methods with the finite element method. On this basis, the stiffness coefficient of variable stiffness spring was determined, and the frequency matching method for the variable stiffness vibration energy feedback device was proposed. Finally, fixed-frequency and swept-frequency bench tests were conducted to compare the energy capture performance of the fixed and variable stiffness vibration energy feedback devices. The results demonstrated that the variable stiffness spring structure significantly broadened the resonance band of the device, increased the maximum instantaneous output voltage to 52.2V, and boosted the average output power and electricity generation by 34.4% and 42.8%, respectively，at the same time, it met the vibration reduction requirement for the battery pack. The above research can provide theoretical and technical support for the power supply of electric tractors and it was of great significance to promote the development of its energy-saving technology.

Abstract:In order to realize the precise control of ploughing resistance and improve the traction efficiency of tractor in ploughing operation, a prediction method of ploughing resistance was proposed. A plowing resistance perception model based on the upper pull rod force was proposed, and the experimental verification was carried out. In view of the problem that the measurement results were unstable by only sensing the ploughing resistance through the upper pull rod force in the actual field operation, a ploughing operation parameter test platform was built, and the multi-source heterogeneous sensor data based on tillage depth, upper pull rod force, vehicle speed and wheel speed were obtained and the prediction samples were constructed. The wavelet threshold denoising (WTD) and sparrow search algorithm (SSA) were introduced into the least squares support vector machine (LSSVM), and the combined prediction model of ploughing resistance based on WTD-SSA-LSSVM was established and the model performance was verified. The results showed that the model prediction method had higher accuracy than that of the upper pull rod force. In addition, different prediction methods were compared. The R2, MAE, RMSE and MAPE of the test set obtained by the combined prediction model method were 0.97, 118.1N, 151.4N and 2.2%, respectively. Compared with prediction models of LSSVM and SSA-LSSVM, R2 was increased by 8.9% and 5.4%, respectively. MAE was decreased by 49.7% and 42.2%, respectively. RMSE was decreased by 46.7% and 39.1%, respectively. MAPE was decreased by 56.8% and 48.8%, respectively. Therefore, the method proposed had better prediction performance and was more suitable for the prediction of tractor plough resistance.

Abstract:Accurate tire-soil models are fundamental to optimizing the traction force control of hybrid tractors, thereby improving overall energy efficiency. Among these, precise estimation of the longitudinal and vertical forces is crucial for ensuring model identification accuracy. However, during operation, terrain variations and the unknown load changes caused by the interaction between the plow body and the soil lead to significant fluctuations in the longitudinal and vertical forces on each tire, which increases the difficulty of accurate estimation. To address this, a tire-soil interaction model identification method based on the fusion of multiple filters and joint observation was proposed, taking the distributed hybrid electric tractor (DHET) as research object. Firstly, the longitudinal force of the tire was estimated by using the Kalman filter (KF), and the vertical force was estimated by using the cubature Kalman filter (CKF) algorithm combined with the dynamic model. Then, based on the Brixius model and the vehicle’s longitudinal dynamics, an online identification device for the tire-soil model was constructed by using the unscented particle filter (UPF). The unscented transformation (UT) was used to design the particle proposal distribution, preserving higher-order information in the nonlinear tire-soil model, and the particle filter (PF) algorithm was applied to update the estimated parameters, completing the model’s online identification. Hardware-in-the-loop (HIL) test results showed that the method can quickly and accurately achieve identification under varying terrain conditions, with a root mean square error not exceeding 1.2. Experimental results indicated that, under two types of soil conditions, the identification model’s boundary traction coefficient error was within ±0015, with the front wheel test data distribution accounting for 84.45% and 88.16%, respectively, and the rear wheel data for 86.72% and 85.38%, verifying that this method maintained high accuracy and robustness under different soil conditions. The research findings can provide theoretical support for the optimal drive force distribution strategy of tractors and contribute to improving traction efficiency and fuel economy.

Abstract:The operation state of agricultural machinery is a key indicator for assessing the efficiency of agricultural mechanization and precise management. To realize the precise monitoring of the operation state of farm machinery by the data management platform, an operation state identification method was proposed based on the multidimensional features of the spatial track of farm machinery. Firstly, based on the big data management platform supported by the modern information technology of Internet of Things, the potential characteristics of the trajectory points in the operating space of agricultural machinery were studied, and the distribution laws of the characteristics such as speed, acceleration, steering rate and distribution density of trajectory points were analyzed. Secondly, based on the distribution characteristics of each feature and the demand for operation state recognition, a multi-strategy split-box processing method was used to quantitatively divide the features, and weight of evidence (WOE) and information value (IV) methods were introduced to quantify the influence weight of different features on the operation state of the farm machinery, so as to assess the impact on operation state the key features of the recognition ability were evaluated. Finally, based on the multidimensional key features of the spatial trajectory points of the farm machinery, the fusion algorithm of BP neural network and AdaBoost was combined to recognize the operation state of the farm machinery. The experimental results showed that the accuracy of the proposed algorithm model in the prediction of the operation state of agricultural machinery was as high as 97.3%, indicating that the recognition method based on the multidimensional features of agricultural machinery can accurately recognize the operation state of agricultural machinery.

Abstract:Aiming to address the issues of missing accuracy evaluation for tractor dynamic models and the lack of systematic, standardized assessment methods, an accuracy evaluation approach was proposed based on time-response error and grey relational analysis, grounded in the verification, validation, and accreditation (VV&A) theory of model simulation systems. A hierarchical validation index system comprising phase error, amplitude error, and shape error was constructed, and the grey relational degree method was applied for segmental evaluation of dynamic response curves. Four sets of idealized test cases with defined error components were designed and analyzed by using both the proposed method and traditional numerical error analysis techniques. The results demonstrated that the proposed method effectively decoupled and accurately identified multiple error characteristics in each test case, precisely reflecting the sources and distribution of errors, while its comprehensive evaluation accuracy outperformed traditional methods. Furthermore, taking a self-developed longitudinal dynamic model of a power-shift tractor as an application case, real vehicle data were collected under five road conditions: tilled field, untilled field, stubble field, concrete surface, and mixed road surface. The proposed method was used to compare and analyze simulation results against real vehicle data under various working conditions. The results indicated that the mean comprehensive accuracy score of the model was 90.95. Under high-speed and complex road conditions, scores for various error characteristics decreased significantly, with comprehensive results in some extreme conditions falling below 88, reflecting certain limitations of the model under different working conditions, which aligned with objective reality and verified the effectiveness and comprehensiveness of the proposed evaluation method.

Abstract:Hilly and mountainous regions are characterized by complex terrain and agricultural machinery with high centers of gravity, which often leads to instability and rollover risks during rail transport operations. To address this challenge, a hydraulic automatic leveling anti-instability system for a mountain rail transport platform was proposed. The system integrated an ADXL345 triaxial accelerometer for real-time attitude monitoring and employed a particle swarm optimization fuzzy PID (PSO Fuzzy-PID) controller, using the platform’s Z-axis tilt angle as the feedback variable. A simulation model was established in Simulink to compare the performance of conventional PID, fuzzy PID, and PSO Fuzzy-PID controllers. Simulation results demonstrated that the PSO Fuzzy-PID controller achieved a rapid response time of 0.181s with zero overshoot, and outperformed conventional PID and fuzzy PID controllers in terms of adjustment time, stability, and overall dynamic response. Static experimental tests were conducted under track inclinations of 4°, 7°, and 10°, and the PSO Fuzzy-PID controller yielded an average leveling error standard deviation of only 0.068°, with maximum error constrained within 0.45°. The average leveling time was reduced by approximately 6.3% and 15.6% compared with that of fuzzy PID and conventional PID, respectively. Dynamic experiments further revealed that the system maintained a response delay below 0.4s, with platform attitude fluctuations confined to -0.079° to 0.497°, thereby meeting the precise leveling requirement within ±10°. These results confirmed the effectiveness and reliability of the proposed PSO Fuzzy-PID based hydraulic leveling system. The system significantly enhanced the operational stability and safety of agricultural machinery transport on hilly rail platforms, providing valuable technical support for the intelligent development and practical application of agricultural transport equipment in mountainous regions.

Abstract:Aiming at the problem that the artificial setting of comprehensive coefficient is difficult to adapt to the complex and changeable working environment of soil conditions in the process of tractor electro-hydraulic suspension tillage depth control, a tillage depth control method based on optimal draft-position comprehensive coefficient was proposed. Firstly, the test platform of tractor electro-hydraulic suspension draft-position integrated tillage depth control and loading system was built and the system was modeled by system identification method. Then aiming at the problem of uncertain parameters of the equivalent transfer function of the system, a global sliding mode control algorithm was designed, and an arc-tangent function was introduced to reduce the chattering of the system. The full factor sample test was designed, and the comprehensive operation evaluation index was established by compromise programming method. The expected draft-position comprehensive coefficient under different working conditions was obtained, and the optimal draft-position comprehensive model was established based on the least squares support vector machine optimized by particle swarm optimization algorithm (PSO-LSSVM). Finally, the effect of the control system was verified by using the built control system test platform. The results showed that the proposed algorithm achieved a rise time of 1.06s without noticeable chattering. Compared with the fixed weight-depth control method with a comprehensive coefficient of 0.5, the proposed method reduced the depth deviation rate by 19.5% and 13.9% at target depths of 140mm and 220mm, respectively, while improving depth uniformity by 16.6% and 11.2%. This method can realize the adaptive adjustment of the comprehensive coefficient according to the soil specific resistance and tillage depth, and was more adaptable to the complex field operation environment, which provided a reference for the research of tillage depth control method.

Abstract:Accurate detection of the drive wheel slip in tractor ploughing operations is fundamental to achieving slip control, which helps improve the traction efficiency and ploughing quality of a tractor. To address the issue of ensuring the accuracy and real-time performance of slip rate detection under heavy load ploughing conditions, a slip rate detection method based on multisensor data fusion was proposed. A real-time data collection platform for tractor ploughing operations was established. Field experiments were conducted to obtain tillage speed, forward acceleration, tractor pitch angle, roll angle, traction resistance, tillage depth and corresponding slip rate under different operating conditions. A datasets was constructed and subjected to statistical analysis. Four machine learning algorithms—random forest (RF), support vector regression (SVR), radial basis function neural network (RBFNN), and artificial neural network (ANN)—were employed to construct the slip rate detection models. The hyperparameters of these models were optimized by using grid search and particle swarm optimization (PSO). The performance of the models was evaluated via three metrics: the coefficient of determination (R2), the root mean square error (RMSE), and the mean absolute percentage error (MAPE). The results indicated that, except for the RF model, which exhibited overfitting, the other three models achieved satisfactory slip rate predictions (R2>0.9). Among them, the ANN model optimized with particle swarm optimization (PSO) for initial threshold weight optimization outperformed both the SVR model with an R2 of 0.918, an RMSE of 2.6% and a MAPE of 8.9%, and the RBFNN model with an R2 of 0.903, an RMSE of 3.0%, and a MAPE of 8.8% on the test dataset, achieving an R2 of 0.937, an RMSE of 1.0%, and a MAPE of 7.6%. This method enabled reliable and accurate detection of tractor drive wheel slip rate, offering an approach for precise online detection and control of slip rate.

Abstract:Due to the complex working conditions of tractors, which are influenced by the working environment, soil conditions, and traction loads, random time-varying disturbances during power shifting process could easily cause shifting jerk and even damage the transmission. Therefore, an optimal control method for tractor power shifting process considering time-varying disturbance observer was proposed. Firstly, the dynamic characteristics of the tractor shift process were analyzed to establish the shifting process dynamics model. The shift time, sliding friction work, and shock degree were taken as the shifting quality indicators, and a quadratic index function of the shifting control system was constructed. Then a high-order disturbance observer (HDO) was introduced to estimate the time-varying disturbances and their derivatives during the shifting process, and the linear quadratic regulator (LQR) was derived and solved for the optimal control law of the shifting process based on the Hamiltonian function. Finally, a hardware-in-the-loop simulation platform for tractors was built based on dSPACE and Matlab/Simulink to verify the shifting process performance under different operating conditions. The results showed that taking the shift from 1LL gear to 1L gear as an example, the shifting times under the transportation and plowing working conditions were 0.72s and 0.85s respectively, which were reduced by 22.58% and 38.4% compared with that of LQR. After compensation by HDO, the maximum shifting jerks were 16.11m/s3 and 8.51m/s3, respectively, which were 5.29% and 2.04% higher than that of LQR, the shifting jerk difference was not obvious, but the sliding friction work control effect was significant, being only 4.61kJ and 8.82kJ, which were reduced by 38.04% and 59.17% compared with that of LQR. The method could effectively observe and suppress the time-varying disturbances during the shifting process, and improve the shifting comfort.

Abstract:Aiming to address issues with traditional mechanical adjustment methods for no-till precision planters in stubble fields, such as the ground wheel assembly’s inability to adapt to terrain changes, leading to unstable residue clearing performance and increased wheel slip, a hydraulically controlled terrain-following depth-limiting ground wheel assembly was designed. Its hydraulic system enabled rapid adjustment of wheel height while maintaining a constant position, ensuring stable furrow opener depth. The cushioning spring reduced ground-induced impacts transmitted to the planter, enhancing the stability of the residue clearing device. Through theoretical analysis, the main structural parameters of the terrain-following depth-limiting ground wheel assembly were determined. Simulation tests identified the optimal stiffness of the cushioning spring as 1.2×105N/m. Comparative field trials demonstrated that the hydraulically controlled terrain-following depth-limiting ground wheel assembly operated stably with lower slip rates. When the furrow opener depth qualification criterion was set at (50±10)mm, the qualification rates for opener soil penetration depth were 62% and 82% for the mechanically-adjusted ground wheel assembly and hydraulic-controlled terrain-following depth-control wheel assembly, respectively. The hydraulic assembly exhibited a 20 percentage point higher qualification rate than the mechanical assembly. Furthermore, the hydraulic terrain-following wheel demonstrated significantly lower slip rates. While both adjustment methods showed increasing slip rates with the rise of operating speeds, the hydraulic terrain-following wheel exhibited a smaller rate of increase.

Abstract:In response to the fact that a large amount of straw and sticky soil of oilseed rape direct seeding in the middle and lower reaches of the Yangtze River, from the perspective of combining oilseed rape seedbed preparation and deep fertilizer application operation, a preparation method of seedbed with the whole width of the compartment shallowly rotated and the seeding belt rotated and cut for deep loosening and deep application of fertilizer was proposed, and a rotary-cutting combined deep fertilizer application device was designed. Based on the characteristics of rice root distribution in the previous crop in the tillage layer, combined with the root growth characteristics of rapeseed plants, the fertilizer depth of 100~150mm and the deep loosening depth of 200mm were determined;based on the kinetic analysis of the stubble cutting and breaking disc to cut off the stubble force, and the notch slip-cutting curves were qualitatively analyzed, the radius of the disc, the angle of the slip-cutting and the number of the notches were determined;based on the plowing depth, key parameters of the deep loosening and breaking shovels were determined;and the key parameters of the rotary-cutting and breaking shovel were determined through the single disturbed area, and the row spacing of rotary-cutting combined deep fertilizer application device was obtained. Using discrete element software to simulate the operational performance of a rotary-cutting combined deep fertilizer application device, the results showed that when the fertilizer depth was 150mm, the seed bed belt was reasonably constructed, and the straw was evenly mixed and buried. The fertilizer depth was 156.60mm, with a depth stability coefficient of 87.46%, and a uniformity coefficient of 13.18%. The field performance test showed that when the fertilizer depth was 150mm, the stubble burying rate was 86.65%;the plowing depth of the seed bed belt and non-seed bed belt was 191.17mm and 112.50mm, respectively;the stability coefficient of plowing depth was 90.70% and 91.18%, respectively;the actual fertilizer depth was 149.44mm, and the qualified rate of fertilizer depth was 85.80%. The field sowing test showed that when the fertilizer depth was 150mm, the actual fertilizer depth was 151.56mm, the qualified rate of fertilizer depth was 85.83%, the coefficient of variation of fertilizer uniformity was 14.33%, and the seedling emergence rate under the treatment of rotary-cutting combined deep fertilizer device was 88.97%, which was higher than that of rotary-tillage mixed fertilizer and rotary-tillage without fertilizer, and the length of the main root in the flowering stage was increased by 22.24% and 38.26%, the diameter of the main root neck was increased by 0.60% and 3.57%, and rapeseed yield was increased by 14.02% and 37.50%, respectively. The results showed that the rotary-cutting combined deep fertilization device provided a suitable seedbed strip and fertilization effect for the growth of oilseed rape, which could provide a reference for the preparation of seedbed and deep fertilization for mechanized direct seeding of oilseed rape.

Abstract:Aiming to achieve orderly seeding of giant Juncao seed stems, enhance seeding efficiency and accuracy, and prevent damage to seed buds, a reciprocating roller precision seed dispenser for giant Juncao seed was designed. This device consisted of a reciprocating seed feeding mechanism, a roller seed dispensing mechanism, and seed guiding components. Based on the physical characteristics of giant Juncao seed stems in the Northwest region and the agronomic requirements for planting the stems, theoretical analyses were conducted on the processes of seed feeding, dispensing, and output of the device to determine the structures and parameters of key components. The working process of the seed dispenser was simulated by using the EDEM method to analyze the forces and motions acting on the stems, confirming that no damage occurred during the seed feeding phase. Using seed filling rate, clogging rate, and qualified planting distance as experimental indicators, a second-order regression orthogonal rotary combination simulation experiment was designed. The optimal structural parameter combination for the seed dispensing roller was determined to be: groove depth of 16mm, groove width of 24mm, and surface spacing of 7mm. The optimal structural parameters for the seed output port were as follows: side plate inclination angle of 45°, opening width of 45mm, and opening length of 30mm. To determine the best operational parameters for the seed dispenser, the qualified planting distance rate and the variation coefficient of planting distance were used as experimental factors in a second-order regression orthogonal rotary combination test. The results showed that when the rotation speed of the seed dispensing roller was 11.57r/min, the height from the seed discharge port to the ground was 191.56mm, and the inclination angle was 35.77°, the qualified planting distance rate reached 96.4%, with a variation coefficient of 9.98%. The working performance of the seed dispenser met the agronomic requirements for giant reed planting.

Abstract:Aiming to enhance the operational adaptability of transplanters in hilly and mountainous terrain, as well as improve the quality and stability of their performance on slopes, a leveling control system based on super-twisting sliding mode active disturbance rejection control (STSMC-ADRC) was proposed for a previously developed transplanter. This system enabled real-time lateral leveling control of the transplanter operating along contour lines. A mathematical model of the servo-electric cylinder control was established according to the overall structure and leveling control principle of the transplanter to describe and analyze the dynamic performance indicators of the system. Based on the mathematical model of the servo-electric cylinder, which provided the driving force for leveling the transplanter, a super-twisting sliding mode active disturbance rejection controller was designed. Lyapunov stability analysis was conducted for the designed controller in combination with the established control system. A co-simulation platform using Matlab/Simulink and ADAMS was constructed to compare the performance of the proposed STSMC-ADRC with classical PID control and active disturbance rejection control (ADRC) under static conditions. The results indicated that under static leveling conditions, the leveling time with ADRC was approximately 29.5% shorter than that with PID control, while the STSMC-ADRC strategy reduced the leveling time by approximately 46.3% compared with PID control. Finally, static single-side bridge tests and field dynamic tests were conducted. The dynamic test results demonstrated that the STSMC-ADRC control algorithm performed the best in experiments conducted at speeds of 0.2m/s, 0.3m/s, and 0.4m/s, with root mean square errors of 0.77°, 1.94°, and 2.32°, respectively, outperforming both ADRC and PID control algorithms.

Abstract:Most of the existing solanaceous vegetable grafting machines are semi-automatic models. Due to the significant differences in the growth morphology of seedlings, the automatic seedling picking process is difficult to achieve. Based on the working requirements of the fully automatic row grafting equipment for solanaceous vegetables, an automatic scion seedling picking device was designed. This device adopted flexible gripper hands, which had strong adaptability to the growth form of seedlings. It can automatically supply, clamp, cut and transfer the scion trays, achieving the automation of seedling extraction for grafting operations, thereby promoting the upgrading of the grafting machine from semi-automatic to fully automatic operation. In order to verify the working performance of the scion picking device, taking chili seedlings in the grafting period as the object, the success rate of picking as the test index, and the thickness of the elastic piece, the structure of the clamping hand, and the scion growth offset as factors, a three-factor and three-level orthogonal test for scion picking was carried out. The test results showed that the influence degrees of the three factors on the success rate of picking, from large to small, were as follows: gripper structure, growth offset, and elastic piece thickness. The optimal combination was gripper structure Ⅲ, growth offset of 0~14mm, and elastic piece thickness of 0.3mm. The scion picking test was conducted under the optimal combination conditions. A total of 150 scions from three tray seedlings were selected as the seedlings, after experimental verification, the success rate of picking was 98.7%, meeting the performance requirements of the scion picking operation of the grafting machine.

Abstract:Aiming to improve the simulation accuracy of the discrete element model for tea garden soil-organic fertilizer mixture, focusing on the combination of physical experiments and simulation to calibrate the parameters of tea garden soil and organic fertilizer, by using the slope method, rolling friction test, and coefficient of restitution measurement, the physical and contact parameters of soil and organic fertilizer were obtained, and the mixing angle was measured based on the cylindrical lifting method. The Plackett-Burman experiment was used to screen for significant influencing parameters, and the optimization interval was determined by combining the steepest slope test. A quadratic regression model was established by using the Box-Behnken response surface method to obtain the optimal parameter combination: soil-organic fertilizer recovery coefficient of 0.35, organic fertilizer rolling friction factor of 0.115, and organic fertilizer steel recovery coefficient of 0.606. At the same time, the applicable range of moisture content for the model (5%~10%) was clarified through experimental analysis.The verification test showed that the simulated stacking angle was 24.49°, with a relative error of 1.65% compared with the actual value. A prototype of a tea garden tillage and fertilization machine was built, field rotary tillage and fertilization experiments were conducted, and rotary tillage and fertilization experiments were simulated. The relative errors of soil coverage rate in field experiments and simulation experiments at different rotary tillage speeds (150r/min, 200r/min, 250r/min, and 300r/min) were 6.21%, 4.52%, 7.13%, and 7.72%, respectively. The soil-organic fertilizer distribution in simulation experiments was consistent with that in field experiments, verifying the accuracy of the soil-organic fertilizer mixture discrete element model and contact parameters.

Abstract:Aiming to address the issues of resource waste and environmental pollution caused by uniform fertilization in traditional methods, as well as the inadequate adaptability of existing variable-rate fertilization systems’lag compensation approaches, a prescription map-based variable-rate fertilization system for rice and wheat featuring spatio-temporal coupling compensation function was designed. The system employed an STM32F407 chip and CAN bus architecture for the control unit and central processor. Integrated with ArcMap GIS platform, it proposed a digital prescription map generation method based on agronomic expert decisions. The developed human-machine interface software supported GPS message reception/parsing, prescription map loading, and high-efficiency retrieval. Innovatively, a spatio-temporally coupled calibration model for fertilizer application adjustment was established through dynamic outlet positioning modeling and lag time self-learning algorithm, enabling automatic field calibration and dynamic compensation of system response time. Bench tests demonstrated that the self-learning algorithm achieved over 90.5% accuracy across working speeds (4~8km/h) and fertilization gradients (0~450kg/hm2，0~750kg/hm2，450~750kg/hm2). Field trials showed that with spatio-temporal calibration enabled, the average fertilization accuracy reached 97.4% within 50-meter segments—an 8.0 percentage points improvement over non-calibrated status—while maintaining only 2.5 percentage points accuracy fluctuation at 8km/h. This system provided a high-precision, low-cost intelligent solution for variable-rate fertilization in rice/wheat cultivation, offering technical support for China’s national policy of reducing fertilizer use while increasing efficiency.

Abstract:Aiming to address the key issues of low irrigation water use efficiency and backward management levels in grape cultivation in arid valley regions. It focused on collecting big data on the habitat and physiology of grape crops, integrating agronomic theoretical data, and an integrated decision-making system for water and fertilizer management was designed based on comprehensive multi-data coupling analysis. The system utilized standard Modbus protocol sensors for data collection and employed a LORA low-power wireless networking transmission module with freely settable timing. By integrating devices such as cloud boxes, gateways, and DTUs, the system established a soil and environmental Internet of things (IoT) monitoring and collection system, achieving continuous online monitoring and data collection of soil and the environment. The LORA wireless network transmission adopted the Modbus-RTU protocol for data transfer. The system incorporated a crop cultivation model built with an optimized fuzzy PID control algorithm based on the revised PM model. Through this model, the system can drive decision-making for irrigation and fertilization management, optimizing the agricultural production process. It achieved online control of integrated water and fertilizer irrigation equipment, reducing the need for manual operation. Experimental results showed that compared with traditional PID and fuzzy PID, the system’s overshoot was reduced by 13.6 percentage points and 3.6 percentage points respectively, the system response rate was increased by 58.2% and 26.6% respectively;water usage was reduced by 25.0% and 13.5% respectively, and fertilizer usage was reduced by 30.0% and 17.1% respectively. This system aimed to promote the upgrade of integrated water and fertilizer technology, transforming the decision-making mechanism from “theory-oriented, experience-guided, and fixed-value regulation” to a benchmark based on big data analysis and judgment, with the expectation of providing technical support for the realization of real-time and precise regulation of integrated water and fertilizer management in grape plantations in arid valley regions.

Abstract:Aiming to address the needs of modern agriculture in monitoring the overall crop growth status and developmental trends across various growth stages, and enable rapid, non-destructive crop assessment, a handheld crop normalized difference vegetation index (NDVI) sensor based on an active light source was developed. The sensor integrated narrow-band LEDs at 650 and 850nm, along with Si PIN photodetectors, ensuring stable light output and precise signal acquisition. NDVI values were measured by using signal modulation and demodulation techniques across 60 wheat canopies subjected to four different fertilization levels. Experimental results demonstrated excellent light source stability, with standard deviations at both wavelengths maintained within 0.9mV. In optical calibration experiments conducted at varying measurement heights, the results exhibited a strong linear relationship to the calibration parameters, with correlation coefficients (R2) exceeding 0.97. When the measurement height ranged from 20cm to 40cm, NDVI outputs remained largely stable, with root mean square errors (RMSE) of 0.025 and 0.015, respectively. Moreover, the sensor effectively mitigated the influence of ambient light fluctuations, ensuring consistent measurement results. A comparative analysis with a reference instrument demonstrated strong agreement in NDVI measurements, with R2 of 0.79 and RMSE of just 0.031. These results confirmed that the developed handheld crop growth sensor delivered high reliability and accuracy in practical applications, providing robust technical support for crop growth monitoring and a sound scientific foundation for modern agricultural production management.

Abstract:In order to improve the online yield measurement accuracy of grain combine harvester and suppress the interference of body tilting and mechanical vibration during the yield measurement process, a pressure-type grain yield measurement system based on tilting angle and vibration compensation strategy was developed. The system consisted of a pressure-type grain yield measurement device, a GNSS positioning module, a data processor and a host computer. Firstly, a mathematical model of grain yield monitoring was constructed based on the functional relationship between grain yield and sensor pressure, inclination, vibration and humidity, and a test rig for grain yield monitoring was built to optimize the structural parameters of the device through orthogonal test combined with response surface analysis. The effects of the number of pressure sensors, the mounting position of the vibration absorber and the inclination angle of the device on the measurement error were investigated, and the optimal parameter combinations were determined to be five pressure sensors, the 3rd reference position (175mm from the left slot wall) for the vibration absorber and the 21°inclination angle of the device, respectively. The results of univariate (flow rate change) and multivariate (flow rate, inclination angle and vibration change) tests showed that the yield measurement error of 20 groups of tests under different working conditions was stably controlled within 3%;the results of field tests showed that the average yield measurement error in actual operation was 3.41%, with a maximum of no more than 4%, which verified that the system can effectively inhibit the interference of the tilting of the vehicle body and the vibration of the machinery on the accuracy of the measurement of the yield and provide an effective technical support to obtain the information of the distribution of the yield of the precision agriculture. The system can effectively suppress the interference of body tilt and mechanical vibration on yield measurement accuracy, and provide effective technical support for the acquisition of precise agricultural yield distribution information.

Abstract:Aiming to address the issue that the threshing and separation device of traditional soybean combine harvesters cannot adjust the threshing gap in real-time, resulting in uneven material distribution in the threshing chamber and consequently high rates of soybean grain breakage, unthreshed grain loss, and entrainment loss, an adjustment system was designed based on the realtime matching of feed rate and threshing gap. Using the feed rate, threshing section gap, and separation section gap as experimental factors, and the grain breakage rate, unthreshed grain rate, and entrainment loss rate as evaluation indicators, bench tests, adaptive threshing gap adjustment tests, and comparative tests were conducted. The experimental results indicated that the average adjustment errors of the control system for the threshing and separation section gap were 0.49mm and 0.47mm, respectively. Under the test combinations where the feed rates were 0~2kg/s, 2~3kg/s, and 3~4kg/s, with corresponding threshing section gaps of 15mm, 23mm and 30mm, and separation section gaps of 15mm, 20mm and 27mm, the grain breakage rate, unthreshed grain rate, and entrainment loss rate were 2.73%, 0.30% and 0.42%, respectively. Compared with the constant gap tests, these values were reduced by 0.39, 0.27 and 0.28 percentage points, respectively. All evaluation indicators from the adaptive threshing gap adjustment tests outperformed those from the comparative tests. The research result can support the development of adaptive threshing gap control systems and the intelligent of the soybean combine harvester.

Abstract:Aiming at the saline peanut harvesting process in the traditional shovel pole type excavation and harvesting device exists in the soil crushing effect is poor, the root and pod part of the soil content rate is high, the excavation operation resistance is large and has other problems. Combining saline soil characteristics and peanut planting agronomic requirements, a cutting and squeezing combined excavation device was designed. A mechanical model for the cutting and squeezing forces of the digging device and a mechanical model for the fragmentation and separation of the peanut root-pod-soil aggregates were established. The main factors affecting the operational performance of the digging device were identified as the blade surface inclination angle of the digging shovel, the blade edge angle of the digging shovel, and the radius of the squeezing roller. The value ranges for these parameters were analyzed and determined. Simulation experiments were conducted by using EDEM software to investigate the dynamic changes in peanut root-pod-soil aggregates during the operation of the digging device. Single-factor experiments were used to determine the influence level of each factor on operational resistance and the soil adhesion rate on the root-pod system. Using operational resistance and soil adhesion rate as evaluation indicators, a quadratic regression orthogonal rotational composite experiment was carried out by employing the central composite design (CCD) method. The optimal operational parameter combination was determined as digging shovel blade surface inclination angle of 22°, digging shovel blade edge angle of 50°, and squeezing roller radius of 110mm. Under these parameters, the soil adhesion rate on the root-pod system was 48.86%, and the machine operational resistance was 15696.3N. Field trial results demonstrated that the cutting-squeezing combined peanut digging device achieved a post-operation soil adhesion rate on the root-pod system of 53.61% and an operational resistance of 15965.57N. Compared with the original digging device, this represented a reduction of 21.58 percentage points in the soil adhesion rate on the harvested root-pod system and a reduction of 5.57% in operational resistance, indicating superior performance.

Abstract:In the plot combine harvester, the impurity removal system based on the Coanda effect serves as a novel secondary cleaning device, featuring advantages of simple structure and compact spatial footprint. However, in practical applications, relying solely on the Coanda effect generated by curved surfaces can only achieve partial impurity separation, as most impurities fail to be effectively removed under airflow, resulting in suboptimal cleaning efficiency. To enhance the device’s performance, strengthen the Coanda effect intensity at the end of the curved surface, and mitigate airflow diffusion, an auxiliary plate structure was introduced to the original design. Using the CFD-DEM method, the particle deflection angles of wheat grains and husk impurities were obtained for the configuration without the auxiliary plate. These angles were then utilized to determine the variation range of the auxiliary plate’s free-end angle (4.89°~27.12°) through geometric relationships. Numerical simulations were employed to analyze the flow field structure within the Coanda effect zone of the device, identifying the height range (0.112~0.128m) where abrupt velocity changes occurred at the end of the Coanda effect. Based on simulation results, the optimal length of the auxiliary plate and the positioning of its fixed end were further refined. Bench tests demonstrated that the highest cleaning efficiency was achieved when the auxiliary plate had a free-end angle of 25°, a length of 155mm, and a fixed-end height of 0.122m, and the impurity content in the grain is 0.998%, conclusively validating its significant role in improving device performance.

Abstract:In order to solve the problem of lack of accurate modeling and parameter support for multi-component heterogeneous particles in the clearing link during the combined millet harvesting process, a research on the modeling and parameter calibration of grain threshing mixture was carried out. Based on the compositional characteristics of millet threshing mixture, three types of typical particle models were constructed for seeds, millet panicle clusters and stalks, and the discrete element method combined with the auto-filling technology was used to model spheres, heteromorphs and cylinders, respectively, and the modeling morphology was optimized through the analysis of the number of filled spherical surfaces and the degree of morphological differences. Based on the determination of physical parameters of particles, using the angle of repose as an index, the parameter calibration method combining Plackett-Burman test, the steepest ascent test and Box-Behnken response surface optimization was adopted to realize the correction of contact parameters of seeds, millet panicle clusters and short stalks, and the average errors of simulation and actual angle of repose were 1.94%, 3.72% and 1.16%, respectively. In order to verify the accuracy of the model, high-speed camera validation tests were carried out, and the particle trajectories were extracted by using TEMA software and compared with the simulation results. The results showed that the overlap of the three types of particle trajectories was greater than 0.75, and the obtained mean absolute error and root mean square error were at a low level, indicating that the simulation trajectories were highly consistent with the measured results in terms of amplitude and trend, which verified the accuracy, and the model constructed was highly consistent with the measured results. The results showed that the modeling method and calibration parameters constructed had good physical consistency and engineering applicability, which can provide effective support for the particle dynamics simulation and mechanism research of millet cleaning process.

Abstract:Aiming to address the challenge faced by sugarcane harvester operators who rely solely on visual judgment and experience to determine cane tops’positions, making it difficult to adjust cutting height accurately and in real-time, an intelligent positioning system for sugarcane harvester top cutters was proposed based on dual-modal and fuzzy adaptive PID control. Additionally, a cluster-based knife adjustment strategy was developed based on sugarcane cluster feature recognition. The system firstly utilized the YOLO v8 instance segmentation model to detect cane tops, with a depth camera capturing depth data of sugarcane tops in real time and converting pixel coordinates into camera coordinates for height measurement. To validate the depth camera’s field performance, multiple field experiments were conducted. Results showed that within a range of 50~100cm from the sugarcane, the average relative error of the depth camera ranged from 0.189% to 0.949%. Subsequently, fuzzy rules were designed, and a microcontroller integrated with a fuzzy adaptive PID algorithm was used to control the servo motor’s operating speed. The maximum rising response speed was approximately 289.36mm/s, and the descending response speed was about 273.16mm, with a measurement error within ±1mm. PID comparison experiments showed that fuzzy PID control, compared with traditional PID control, reduced the response time by 0.12s, bringing it down to 1.24s. The overshoot was decreased from 12.80mm to 4.63mm, the number of overshoots dropped to one, and the steady-state error stabilized within ±2mm. Finally, dynamic tests demonstrated that the system’s average recognition time was 0.08s, showcasing excellent real-time performance.

Abstract:In order to solve the problems of low adaptability of the tool to the banana crown and the low success rate of the banana in the process of intelligent de-handing, a type of end effector was designed according to the morphological parameters of the banana bunch and the crown and the production requirements of the de-handing. The cutting arc of the adjustable opening-closing de-handing tool was used to adapt to different sizes of the banana crown to achieve continuous banana de-handing. The key parameters and main structure of the de-handing end effector were designed. The mechanical analysis model of the process of cutting banana crown by the de-handing tool was established, and the strength simulation was carried out. The results showed that the maximum stress of the de-handing tool during the de-handing process was 58.147MPa, which was less than the yield strength of the tool material 205MPa. The test platform of robot de-handing was built, and the significant factors affecting the de-handing effect were sorted as the cutting angle of de-handing, the blade angle and the de-handing speed. In order to obtain the optimal parameter combination, taking the peak cutting force and incision quality as the evaluation indices, the combination test of three-factors and three-levels were carried out by using Design-Expert software, and the mathematical regression model between the experimental factors and the evaluation indexes was obtained. The multi-objective optimization of the regression model was carried out, and the optimal parameter combination was obtained as follows: the de-handing speed was 2m/s, the cutting angle of de-handing was 5.77°, and the blade angle was 5°. The results of the optimal parameter combination verification test showed that the peak cutting force was 31.49N, the incision quality was 84 points, the relative error with the prediction results of the regression model was less than 5%, and the success rate of the de-handing was 92%, indicating that the opening-closing end effector of banana de-handing can effectively adapt to the banana crown’s morphology and obtain better incision quality.

Abstract:The existing residual plastic film recovery techniques in Xinjiang’s cotton fields primarily focus on mechanical properties, neglecting the dragging effect of root stubble on the film during collection. This results in residual film readily becoming entangled in the root stubble. It confirmed the operational parameters of the pick-up teeth by analyzing the recovery of residual film when picking up teeth in conditions containing root stubble. Employing a coupled finite element method (FEM) and smooth particle hydrodynamics (SPH) algorithm, the tines, stubble, and residual film were modeled as FEM meshes, while the soil was represented as SPH particles. This enabled simulation of the shear strain, peak stress, pick-up height, and tensile stress exerted by root stubble on the residual film. The recovery of residual plastic film in the presence of cotton root stubble was simulated and the recovery efficiency was analyzed. Field trial results indicated that when root stubble height ranged from 5cm to 9cm, the average recovery rate of residual film reached 91.25%, with a residual film retention rate on root stubble of 4.30%. When root stubble height was 10~12cm, the average residual film recovery rate across plots was 82.66%, with 12.55% of residual film remaining on the root stubble. Plots with root stubble heights of 5~9cm exhibited an 8.59 percentage point increase in recovery rates compared with plots with root stubble heights of 10~12cm, alongside an 8.25 percentage point reduction in residual plastic film on stubble. In trial plots with root stubble heights of 5~9cm, the residual film retention rate on root stubble was 45.34% lower than in plots with root stubble heights of 10~12cm. The research findings can provide theoretical support for the correlation between root stubble and plastic film remnants, as well as for enhancing the recovery rate of plastic film remnants.

Abstract:In recent years, film mulching technology has been widely used in agriculture. After cotton harvest, large amounts of residual film need to be recycled, simultaneously, cotton stalks crushing needs to be carried out, in Xinjiang, China. The existing integrated machine for straw crushing and residual film recycling in China can be mainly divided into drum-type and chain teeth type. Considering the current drum-type integrated machine, which has a low residual film recovery rate, and high impurity content in the residual film, the rubbing-type residual film baling mechanism was prone to throw out the residual film and was easy to deform the baling belt and other problems, an integrated machine for straw crushing and residual film recycling and baling was developed. Through principle analysis and calculation, the key parameters of the hammer blade straw crushing mechanism and the screw conveying mechanism were determined;the residual film picking mechanism and the wheel drive mechanism were improved, and the dimensions and rotation speed of the film pick-up drum, the transmission ratio between the wheel and the film pick-up drum, the dimensions of the film pick-up teeth and the space between adjacent film pick-up teeth were determined;through theoretical analysis, the structural parameters and operating parameters of the screw impurity removal mechanism and the film removal mechanism were determined;by analyzing the process of residual film baling, the parameters such as the angle between the baling belt and the horizontal plane, the linear speed of the baling belt, and the number and space of the support rollers were determined. The field test demonstrated that the optimal operating parameters of the machine were as follows: the machine’s advancing velocity was 10km/h, the depth of the film-picking teeth was 25mm, and the height of the stubble was 10cm. Validation test results demonstrated that under the aforementioned optimal operating parameters, the film pickup rate was 89.78%, the soil content was 11.15%, and the pass rate of crushed straw was 94.79%.

Abstract:Aiming to address the challenges of high labor intensity, low operational efficiency, severe tobacco leaf damage, and lack of suitable harvesting equipment in tobacco-growing areas such as Yunnan, Guizhou, and Hunan provinces, an optimized chassis for a tracked self-propelled semi-automatic tobacco harvester was designed, specifically for hilly regions. The machine featured a gantry frame that enabled simultaneous harvesting three rows of tobacco. To reduce mechanical damage to the tobacco leaves and meet the needs of stable operation at low-speed, a dual-motor independent drive system and a mechanical automatic row-following device were designed. Additionally, based on CAN bus technology, an integrated vehicle control system combing manual and remote control was developed to meet control needs in various scenarios. Stability analysis and performance testing of the machine were carried out. The results showed that the maximum speed of the harvester chassis was 5km/h with minimum turning radius of 903mm. And it could handle slopes with up to 23°, and the lateral ultimate tipping angle was 24.7°. There was no side-slip or rollover when operating on contour lines with a lateral tilt angle of 15°~20° on sloping land. Moreover, it could stop for more than 5 minutes on both upward and downward directions along longitudinal slopes ranging from 10° to 15° without any issues. During field harvesting, the machine exhibited smooth running characteristics with ample power reserves allowing free speed adjustments between 0km/h and 4.5km/h. This highlighted its strong adaptability to manual tobacco leaf harvesting operations. The average mechanical damage rate to tobacco leaves was 0.73 leaves per operation per hectare, and the average mechanical damage rate to tobacco plants was only 0.083 plants per operation per hectare. It showed that the developed tracked self-propelled semi-automatic tobacco leaf harvester chassis could meet the requirements of manual and machine-operated harvesting in narrow fields and gentle slopes, which was of significant importance for promoting the full mechanization of tobacco production.

Abstract:In order to explore the suitability of abandoned cultivated land for recultivation and improve the utilization rate of cultivated land, taking Lishu County, Jilin Province as the study area, firstly based on the Google Earth Engine platform and the 2022 Sentinel-2 image data of Lishu County, a total of 46 features such as spectral features, vegetation features, texture features, and terrain features were calculated, and random forest algorithm was used to construct the optimal feature set, and finally 39 features were screened to participate in the classification, so as to obtain the scope of abandoned cultivated land. Then the AHP-CRITIC-TOPSIS model was used to construct an evaluation model of the suitability of abandoned cultivated land for recultivation from three dimensions: natural, soil and socio-economic conditions, and the suitability was divided into four grades: highly suitable, moderately suitable, barely suitable and unsuitable, so as to put forward a feasible scheme for the rehabilitation of abandoned cultivated land. The results showed that the accuracy of the classification results of land use types in Lishu County in 2022 was relatively accurate, and the overall classification accuracy and Kappa coefficient were above 95%, indicating that the classification results were consistent with the actual land types and the overall classification effect was good. The area of abandoned cultivated land not suitable for recultivation in Lishu County was 1673.73hm2, which was mainly located in the southeast of the study area, and the area of abandoned cultivated land suitable for recultivation was 1071.67hm2, which was mainly located in the low-altitude plain area in the northwest of the study area. At present, the implementation of the national cultivated land protection policy indeed improved the use of cultivated land in rural areas, but the problems of abandonment and low utilization efficiency of rural cultivated land were still prominent, and the results can provide a scientific basis for improving the quality of cultivated land and ensuring food security in Lishu County and surrounding areas.

Abstract:The human-land-water systems within a river basin are interconnected and exhibit mutually reinforcing relationships. Investigating their coupling coordination degree and influencing factors is crucial for achieving regional high-quality development and ecological civilization construction. Taking the nine provinces in the Yellow River Basin as the study area, the entropy method and a comprehensive evaluation method were employed to measure the index levels of the human system, land system, and water system within the Yellow River Basin from 2013 to 2022. Building upon this foundation, the coupling coordination degree model and fixed effects (FE) model were utilized to analyze and evaluate the spatio-temporal differentiation characteristics of the human-land-water systems coupling coordination degree and its influencing factors at both dual (human-land, human-water, land-water) and ternary (human-land-water) levels. The results indicated that from 2013 to 2022, the overall human-land-water systems in the nine provinces showed steady progress, although the development speeds of the subsystems varied, mostly exhibiting a fluctuating upward trend. Significant spatial disparities existed in the coupling coordination degree across the nine provinces during 2013—2022. By 2022, five provinces (Inner Mongolia, Ningxia, Shanxi, Henan, Shandong) reached an intermediate coordination level, while the other four provinces still had considerable room for development. The coupling coordination degrees of the dual systems (human-land, human-water, land-water) in all provinces also generally showed a year-by-year strengthening trend. Technological investment and rural governance investment positively influenced the coupling coordination degree. Industrial structure upgrading, the degree of openness, fixed asset investment, and government regulatory capacity exerted varying degrees of influence on the coupling coordination degree. In conclusion, against the backdrop of advancing the major national strategy for ecological protection and high-quality development in the Yellow River Basin, differentiated pathways for synergistic human-land-water development should be formulated for the upper, middle, and lower reaches. This would facilitate the rational coordination of multi-system development relationships within the basin and promote healthy and sustainable regional development in the future.

Abstract:Accurately and rapidly monitoring the spatio-temporal distribution of aboveground biomass in crops is crucial for assessing growth conditions and managing precision irrigation in agricultural fields. Spectral indices have been widely utilized for estimating aboveground biomass;however, the spectral saturation effect under high coverage significantly impacts model performance. Furthermore, the effectiveness of integrating complementary and synergistic effects of spectral, temperature, and texture information from UAV remote sensing for estimating maize aboveground biomass across different growth stages and irrigation treatments at the farmland scale remains unclear. Additionally, the influence of meteorological factors on the interannual performance of remote sensing models for aboveground biomass estimation requires further investigation. Based on data from experimental sites in Inner Mongolia in 2018, 2019, and 2021, the spectral indices, temperature indices, and texture information obtained from UAV-based multisource remote sensing system, along with meteorological parameters such as reference evapotranspiration and vapor pressure deficit, were used as input features. Five modeling methods—stepwise regression, random forest regression, adaptive boosting regression, support vector regression, and one-dimensional convolutional neural network regression— were employed to establish a multisource feature fusion remote sensing estimation model for aboveground biomass of maize in large fields. The results showed that compared with single-type remote sensing information, the aboveground biomass estimation model integrating spectral, temperature, and texture information had better accuracy. Among them, the one-dimensional convolutional neural network regression model had relatively better accuracy (R2=0.87, RMSE=295.77g/m2). After introducing meteorological parameters, there was no significant improvement in the accuracy of the aboveground biomass estimation model by using drone multisource remote sensing information, indicating redundancy between meteorological parameters and drone multisource remote sensing information. Considering the cost of drone-mounted equipment and the operation efficiency, the aboveground biomass estimation model based on spectral indices and meteorological parameters using random forest regression also had satisfactory accuracy (R2=0.85, RMSE=296.74g/m2), suggesting that meteorological parameters can serve as a supplement to spectral indices. The performance of the one-dimensional convolutional neural network model for aboveground biomass estimation varied across different growth stages, indicating that the accuracy and reliability of the model need to be comprehensively assessed based on the crop growth stage. The research result can provide technical support for promoting the application of drone multisource remote sensing technology in precision irrigation management in farmlands.

Abstract:The watermelon breed testing system combining machine vision and hyperspectral imaging was developed to enhance efficiency, reduce labor, and improve consistency in quality evaluation. The system included an automatic image acquisition device, featuring a Gige camera and a weighing platform, to capture watermelon phenotypic data. It used hyperspectral imaging on selected regions of interest to model SSC distribution within the watermelon, with measurements of transverse and longitudinal diameters achieved through minimum external rectangle fitting and rind thickness estimated using Canny edge detection. Convolutional smoothing (SGS) algorithm was used to preprocess the spectral data by combining three algorithms, namely multivariate scattering correction (MSC), standard normal transform (SNV), and unit vector normalisation (UVN), respectively, and then the best preprocessed spectral data were filtered by competitive adaptive reweighting algorithm (CARS), successive projection algorithm (SPA), and one-time combined dimensionality reduction algorithm (CARS+SPA) for feature wavelength screening, and finally the screened spectral data were used to build PLSR models and analyzed for comparison. The results showed that the system had the highest accuracy of 98.68%, 98.82% and 93.81% for the measured values of transverse diameter, longitudinal diameter and rind thickness of watermelon, respectively, with the root mean square error of 2.43mm, 2.08mm and 1.63mm, respectively, as compared with the manual measurements. The best model for predicting watermelon brix was (SGS+UVN)-CARS-PLSR, with prediction correlation coefficients and of 0.9204 and 0.9127, root mean square errors RMSEC and RMSEP of 0.3760°Brix and 0.4668°Brix, respectively, and a relative analytical error of prediction RPD of 2.49.

Abstract:Near-infrared spectroscopy has been widely used for the rapid analysis of the physical and chemical properties of straw, offering advantages such as fast detection speed, high efficiency, and non-destructive testing. However, in practical applications, differences in straw particle size can reduce the detection accuracy and stability of the model. For rice straw samples with varying geometric particle sizes, near-infrared spectra were obtained in the 10000~4000cm-1 wavelength range. Polynomial fitting (PF) was employed to remove background signals from the spectral data of the straw samples. A one-way analysis of variance (ANOVA) was used to perform statistical testing (P<0.05) on the spectral feature signals of the geometrically sized straw samples after polynomial fitting. Principal component analysis (PCA) and Biplot analysis were employed to assess the contribution of spectral feature signals to the spectral differences among straw samples with different geometric particle sizes. Pearson and Spearman correlation analyses were used to screen for key spectral feature signals. Based on the quadratic polynomial fitting mathematical model, it was assumed that absorbance was composed of particle size-related spectral information (PSRSI) and non-particle size-related spectral information (nPSRSI). A particle size regression correction method (PSRCM) was proposed based on the absorbance and particle size of key spectral feature signals. The results indicated that the spectral feature signals at 5180cm-1 of the geometric particle size straw samples exhibited extremely significant differences (P≤0.05, P≤0.01, P≤0.001) and showed a highly significant correlation with particle size (Pearson’s r was 0.96). The quadratic polynomial mathematical model performed optimally, with an average coefficient of determination R2 and root mean square error (RMSE) of 0.53 and 0.0021, respectively. PSRCM effectively corrected the spectral differences caused by characteristic signals at 6824cm-1, 5360cm-1, and 5180cm-1, with correction effects significantly superior to those of standard normal variate transformation (SNV) and multiplicative scatter correction (MSC). The research result can provide a theoretical basis for the transfer correction method of straw near-infrared spectroscopy models, which can be applied without the need for standard samples.

Abstract:Granulation is a common post-harvest disease of citrus fruits, particularly in pomelos. When this disease occurs, it will affect the taste of pomelo fruits and even make them lose their edible value. Pomelos are large, thick-skinned fruits which are difficult to visually detect their internal granulation. The rapid, non-destructive detection of granulation of pomelos was explored by using near-infrared transmission spectroscopy combined with one-dimensional convolutional neural networks. A self-developed visible-near-infrared full-transmission spectroscopy measurement system was used to collect data from red-flashed pomelos. Data were collected in the 562.03~1110.16nm range under short integration mode. Without spectral preprocessing, the Kennard-Stones algorithm divided samples into training and testing sets. Subsequently, a one-dimensional convolutional neural network model for pomelo dryness identification was established. The network had a seven-layer structure. Testing showed that the model achieved a classification accuracy of 98.15% for granulation levels. Compared with traditional models such as backpropagation neural network (BPNN), kernel extreme learning machine (K-ELM), and extreme gradient boosting tree (XGBoost), which used raw spectral data and polynomial smoothing processing (SG) and competitive adaptive reweighted sampling (CARS) to extract feature variables. Finally, data augmentation further validated the model’s stability. The proposed convolutional neural network model effectively identified internal granulation without complex preprocessing. It can select effective information during the spectral processing, greatly improving the model identification efficiency. The researh provided a reference for the application of deep learning methods to achieve rapid and accurate non-destructive identification of the internal granulation level of pomelo.

Abstract:Cucumber downy mildew is a fungal disease that severely threatens cucumber production and quality. Quantitative detection of sporangia and conidiophores is crucial for early disease prevention. However, traditional horizontal bounding box detection methods cannot accurately detect these features due to their diverse morphology and orientations. Therefore, an improved YOLO v8sOBB detection method was proposed by introducing the convolutional block attention module (CBAM) and the lightweight shared convolution detection head (LSCD) module. The aim was to enhance the detection efficiency and accuracy of sporangia and conidiophores of cucumber downy mildew. By incorporating CBAM, the model’s ability to identify key features was enhanced, allowing it to focus more on critical regions in microscopic images and improve the detection of small targets. The LSCD integrated multi-scale features through shared convolution operations, enhancing the model’s detection performance for targets of different sizes while reducing computational costs, making it suitable for resource-constrained environments. The rotated bounding box technique accurately captured sporangia and conidiophores’ inclination and rotation postures. Experimental results showed that, compared with the original YOLO v8s-OBB model, the improved YOLO v8s-OBB model not only reduced the model size but also achieved superior detection performance for sporangia and conidiophores of cucumber downy mildew, with precision, recall, and mAP@0.5 reaching 96.0%, 90.1%, and 96.5%, respectively. The improved YOLO v8s-OBB model outperformed advanced rotated object detection models such as S2ANet, H2RBox, and R2CNN in detection accuracy. The research result can validate the effectiveness of the improved model in practical applications and provide technical support for the early diagnosis of cucumber downy mildew.

Abstract:Wheat, as one of the world’s essential food crops, requires timely diagnosis and control of diseases and pests to significantly reduce yield losses, which is crucial for global food security. However, deep learning methods typically rely on large amounts of training data and high-performance computing resources, which pose limitations in scenarios with small sample sizes and limited resources. To address these issues, a knowledge-enhanced lightweight multi-modal wheat diseases and pests identification model named multi-modal blend convolutional neural network (Blend-CNN) was proposed. The model was structured around a dual-branch convolutional neural network framework. It utilized an EfficientNet backbone network to extract image features of diseases and pests and incorporated a multi-branch TextCNN backbone network to extract textual features of diseases and pests descriptions, thereby obtaining more feature information and improving identification accuracy. Additionally, an innovative convolutional network-based multi-modal fusion method was introduced, allowing the model to optimally integrate information from both modalities globally. Furthermore, to mitigate the accuracy loss common in traditional multi-modal methods, the gradient blend loss function was developed. Finally, to verify the model’s effectiveness, comparative experiments were conducted on a constructed dataset containing 880 samples. The results demonstrated that the proposed model achieved the highest identification accuracy of 96.95% on this dataset, compared with other models, it had fewer parameters, lower complexity, and it was more lightweight which was applicable for edge devices, offering theoretical support for wheat diseases and pests identification in scenarios with limited sample sizes and resources.

Abstract:Aiming to address the issues of small target size, varying pest postures, and low detection efficiency, and suboptimal performance of trapping devices during live rice pest identification, an intelligent recognition method was proposed based on an improved YOLO v8n model integrated with a one-to-three live targeted trapping device. By optimizing the structural design of the traditional targeted recognition equipment, a one-to-three live targeted trapping device was developed to enhance operational efficiency. For the construction of the dataset required for intelligent recognition, a frame-by-frame image clarity comparison and sorting method was proposed to improve data quality. Additionally, an adaptive module was introduced into the improved CBAM attention mechanism to enhance the model’s focus on target regions, and a high-resolution P2 layer was added to the AFPN structure to improve small target recognition capability. Comparative experiments demonstrated that the improved YOLO v8n model achieved a mean average precision (mAP@0.5) of 95.65% on the rice leaf folder dataset. Ablation experiments further verified that the enhanced CBAM module and the introduction of the P2 layer in the AFPN structure significantly improved model performance, resulting in optimal overall recognition capability. Field trials confirmed the practical effectiveness of the improved YOLO v8n model, achieving a recognition rate of over 90% for rice leaf folder pests across 60 mu (approximately 4 hectares) of rice fields when combined with the targeted identification device, demonstrating good robustness and stability. In summary, the intelligent recognition method for live rice pests based on YOLO v8n the improved model proposed effectively enhanced recognition accuracy and efficiency, providing a scientific basis for precise pest identification and control in rice fields, with strong practical application value and promotion potential.

Abstract:Gossypol, a toxic compound unique to the Gossypium species, holds significant research value in agriculture and pharmaceuticals. Pigment glands, the primary carriers of gossypol, exhibit phenotypic traits that strongly correlate with gossypol content, making them effective indicators for its evaluation. However, the detection of pigment glands primarily relies on manual analysis, which is both time-consuming and inefficient. To address these limitations, Gland-MSConNet, a novel model designed for the accurate segmentation of cotton leaf pigment glands in complex backgrounds was proposed. The model integrated the convolutional bidirectional Mamba (ConBimamba) and multi-scale attention aggregation (MSAA) modules, significantly enhancing its ability to capture multi-scale features and improving robustness in challenging scenarios. Additionally, the incorporation of sub-pixel convolution facilitated efficient upsampling, preserving fine-grained details. Experimental results demonstrated that Gland-MSConNet achieved a segmentation mIoU of 89.83% and a recall rate of 95.03%, representing improvements of 4.66 percentage points and 5.78 percentage points, respectively, over the U-Net model. These findings highlighted the model’s high segmentation accuracy and robustness. The Gland-MSConNet model provided a strong technical foundation for the segmentation and phenotypic quantification of cotton leaf pigment glands, offering a reliable basis for the automated evaluation of gossypol content.

Abstract:Cucumber is one of the most important vegetables and economic crops in the world. The occurrence of fungal diseases in cucumbers seriously threatens the safety of cucumber production, with powdery mildew being one of the most common fungal diseases. With the rapid development of computer technology, more and more deep learning algorithms are being applied to identify powdery mildew fungus. However, existing algorithms suffer from low accuracy in recognizing small and occluded targets, as well as insufficient localization precision. To address this issue, the parallelized patch-aware attention (PPA) module was firstly introduced into the backbone network of YOLO v8s. By employing a parallel multi-branch structure and attention mechanism, it effectively captured multi-scale features of small targets, preserved critical information during multiple downsampling processes, and enhanced the performance of small target detection. Additionally, the global-to-local spatial aggregation (GLSA) module was introduced into the neck, which combined global contextual information with local detail features, significantly improving the model’s feature representation capability. This module enhanced the detection performance for small targets and complex scenes by better capturing multi-scale features. Experimental results showed that PG-YOLO v8s significantly improved powdery mildew fungus detection performance compared with YOLO v8s. The network achieved high precision in detecting powdery mildew fungus, with notable improvements in the detection accuracy of small and occluded targets. The research result can provide a high-throughput method for detecting powdery mildew fungus, enabling precise early detection and guiding early intelligent decision-making in cucumber production. This approach can help to improve disease control efficiency, ensure cucumber yield and quality, and it was of great significance for the sustainable development of agricultural production.

Abstract:Addressing the challenges of apricot mushroom grading, which relies heavily on manual labor, is inefficient, and prone to subjectivity. An automated grading and detection method for apricot mushrooms named FCML-YOLO v8 was proposed. Building upon the improved YOLO v8n-seg, the network sequentially integrated the feature-focused diffusion pyramid network (FDPN), characteristic attention fusion module (CAFM), and multi-level channel attention (MLCA) mechanism. By combining the FCML-YOLO v8 model with subordinate judgment, mask merging, and center screening methods, accurate assessments of apricot mushroom size and disease spot conditions was achieved. Additionally, through image refinement and curve-fitting techniques, it effectively quantified the curvature of apricot mushrooms. Furthermore, by leveraging CIE XYZ color space values, the color characteristics of apricot mushrooms were precisely analyzed. By integrating these four indicators, the automated grading and detection of apricot mushrooms were successfully implemented. Experimental results demonstrated that the FCML-YOLO v8 model achieved a precision of 94.41%, a recall rate of 88.48%, and an mAP@0.5 of 92.04% in bounding box prediction. In mask detection, it achieved a precision of 90.52%, a recall rate of 84.14%, and an mAP@0.5 of 86.79%. Compared with the original YOLO v8n-seg model, the mAP@0.5 of FCML-YOLO v8 was improved by 6.39 and 4.79 percentage points, respectively, in bounding box prediction and mask detection. In the constructed apricot mushroom grading and detection experimental setup, the FCML-YOLO v8 model achieved an mAP@0.5 of 92.98%, fully meeting the requirements of industrial-grade applications.

Abstract:The integration of multi-source heterogeneous agricultural information not only help understanding and unveiling the pathways of diseases infection and prevention but also offers vital data support for research on diseases prescription recommendations. Aiming at the problems of fusion, alignment and heterogeneity in the analysis of multi-source heterogeneous agricultural data such as electronic medical records, the knowledge graph of crop diseases prescription was constructed, and the diseases association on this basis was visually analyzed. Initially, starting from the principle of the disease triangle, the critical roles of pathogens, hosts, and environment in the process of disease infection and control were analyzed, constructing the ontology layer of the crop diseases prescription knowledge graph with 18 ontology concepts, 17 relationships, and 6 attribute edges based on data characteristics. Subsequently, the entity layer of the knowledge graph was built by integrating rule-matching and deep-learning knowledge extraction methods, encompassing 1121 entity instances and 8292 relationship instances. Lastly, identification of Top20 key nodes based on degree and betweenness centrality, along with Top5 disease-prevention product association predictions using the Adamic-Adar index, were conducted to visually analyze the associations between key entities and attributes, including diseases, symptoms, and prevention products. Six rules were established to enhance the recommendation and information retrieval functions of disease prevention products from three perspectives: prevention scheme selection, green subsidy filtering, and related information inquiry. The research result can provide reference for electronic medical record data mining and association analysis of crop diseases.

Abstract:The planting density of rice in the early seedling stage is closely related to yield estimation and field management. With the development of computer vision technology, the use of drone image recognition technology to automatically count rice is gradually replacing traditional manual sampling and counting methods. However, images taken by drones have problems such as irregular rice seedling shapes, overlapping rice seedlings, scale changes, and background occlusion, which lead to deviations in feature learning, which in turn affects the accurate estimation of rice density and the efficiency of computational prediction. Significant relative size differences among rice seedlings will also become another important factor affecting counting accuracy. In response to these problems, a rice seedling counting algorithm (rice counting network, RCN) was proposed based on an improved multi-column convolutional neural network. Based on the size and shape of the target, an improved Gaussian density kernel was proposed, which combined the similarity information of the rice image and the RGB image of the center point to make the density kernel more suitable for the texture of the rice seedlings. The number of rice ears of different sizes and shapes was quantified, and the scales were divided finely to calculate the receptive field sizes of different features that needed to be fused. The fusion expansion convolution module was used to increase the receptive field size and enhance the model’s multi-scale capabilities. On this basis, appropriate attention mechanisms were added to the pathways at different scales to enhance attention to reasonable information about rice seedlings at different scales. It showed obvious effectiveness in scenes with complex backgrounds, multi-scales and dense occlusions. On the RSD2023 data set captured at a height of 5 m, RCN’s mean absolute error MAE=3.10holes/m2 and mean square error MSE=4.03holes2/m4, which were better than that of the mainstream algorithms such as MCNN, CSRNet, SCAR and TasselNetV2

Abstract:Deep learning methods are widely applied in agricultural vision tasks. Determining harvesting parameters is a primary task for the mechanized harvesting of Camellia oleifera fruits, and automated fruit detection and object counting based on deep learning are crucial for accurately setting these parameters. However, factors such as occlusion and illumination variations in complex orchard environments directly affect the efficiency and accuracy of detection methods. An efficient, lightweight model named YOLO-LGC was proposed based on YOLO 11n, aiming to enhance target detection accuracy and spatial localization capability in complex orchard environments. Firstly, the LNN ghost dynamic convolution (LGC) module was introduced into the backbone network to optimize feature extraction through adaptive kernel selection, thereby improving detection accuracy in complex environments. Secondly, the programmable gradient information (PGI) mechanism was adopted to dynamically adjust gradient propagation, addressing feature conflicts caused by occlusion and illumination changes. Finally, the Dynamic UpSample module was integrated, guided by PGI, to enhance feature retention and improve spatial localization accuracy. Combined axis-aligned bounding boxes (AABB), an inertial measurement unit (IMU), and Kalman filtering to visualize the spatial distribution of Camellia oleifera fruits at heights of 0~1.5m, 1.5~1.8m, and above 1.8m. By analyzing the spatial relationships of the fruits, more accurate positional information was provided. Experimental results showed that compared with YOLO 11n, YOLO-LGC improved detection precision from 82.0% to 83.0%, recall from 76.4% to 78.0%, mAP@0.5 from 85.1% to 87.6%, and FPS from 229.8f/s to 299.7f/s, with computational complexity slightly increased to 6.7×109. While maintaining the advantage of high inference speed, YOLO-LGC enhances detection performance, outperforming existing mainstream detection methods. The spatial distribution error rate was less than 5%, meeting practical detection requirements. The findings can provide a basis for matching mechanized harvesting parameters for Camellia oleifera fruits.

Abstract:Aiming at the problems such as occlusion and large deformation of cow images in the cowshed under a wide field of view, which cause difficulties in individual identification, a method for individual identification of cows in the cowshed under a wide field of view based on the DAM-ResNet model was proposed. The Mask R-CNN model was used to segment four types of dairy cow parts: the back, the left side of the trunk, the right side of the trunk and the buttocks. Based on the ResNet34 residual network, the second-generation deformable convolution was introduced to enhance the extraction of deformed image features of cow patterns in the segmentation results. Integrating the AFF attention feature fusion module into the residual structure to achieve accurate recognition of images of small target cows at a distance. The fine-grained classification loss function, interchannel loss (MC-Loss), was adopted to improve the recognition accuracy of the model for cows with similar patterns. The multi-part dataset of dairy cows was constructed by using the segmented images of dairy cows, and the DAM-ResNet model was trained. Individual cow recognition tests were conducted on a dataset of 12864 images of 57 dairy cows. The results showed that the recognition accuracy rates of the back, left trunk, right trunk and buttocks of dairy cows were 96.13%, 96.56%, 96.94% and 93.14%, respectively, which were 2.76, 2.88, 2.92 and 4.25 percentage points higher than those of the original ResNet model. The recognition accuracy rates of the method proposed within the ranges of 10~20m, 20~30m and 30~40m were 97.54%, 90.72% and 82.17%, respectively. The research results can provide technical support for intelligent dairy cow breeding.

Abstract:With the intelligent development of animal husbandry, computer-vision-based individual cattle recognition is playing an increasingly important role in modern cattle farming. However, in practical applications, the rapid changes in the angles of cattle faces, frequent fluctuations in the number of cattle, and limitations in model recognition speed severely affect the accuracy and efficiency of recognition. To address these challenges, an individual recognition method for Simmental cattle was proposed based on improved YOLO v8 keypoint detection. The method consisted of two stages: in the cattle face detection stage, the improved YOLO v8 model automatically detected keypoints in cattle face images and performed correction alignment and background cropping to make the angles of the cattle face images more consistent. In the cattle face recognition stage, depthwise separable convolution was firstly introduced to reduce the number of parameters, and then residual connections and attention mechanisms were utilized to enhance the model’s feature extraction capabilities, thereby improving recognition accuracy and speed. To validate the effectiveness of the method, a dataset of cattle face images containing 159 Simmental cattle was constructed. Experimental results showed that the proposed model had significant improvements in the number of parameters and inference speed, with processing time for a single image reduced to within 39ms and a recognition accuracy of 95.8%, an average improvement of 4.8 percentage points compared with that of other models. Additionally, the proposed method supported real-time updates to the cattle database, maintaining high accuracy even after adding new cattle face images.

Abstract:Given the current research on beef cattle behavior, which mainly focused on basic behavior recognition and lacked local perception of complex behaviors, the identification of fine behaviors under basic behaviors of beef cattle was researched. A multi-behavior detection and statistical method based on YOLO v8 for beef cattle was proposed. Cattle behavior images were collected by cameras to build a comprehensive dataset that included fundamental behaviors such as standing, lying down, feeding, and drinking, as well as fine behaviors like licking, walking, searching, and tail flicking. YOLO v8n-P2 was selected as the basic model to enhance the ability of the model to detect calves;the feature extraction structure of C2F-PPA was designed;the YOLO v8 detection head was improved by interactive strategy, and the TDADH was constructed. MPDIoU loss function was used to address limitations associated with CIoU. Subsequently, statistical analysis of cattle behaviors based on detection results throughout the day was conducted;these results were visually displayed for various occlusion environments. The experimental results showed that the PTPM-YOLO v8n model achieved precision rate of 91.0%, a recall rate of 87.9%, and a mAP@0.5 score of 94.3% in recognizing all eight behaviors tested. Compared with the original model YOLO v8n, the mAP@0.5 of PTPM-YOLO v8n was increased by 3.0 percentage points, and the parameter number was decreased by 21.9%. which identified all behaviors and basic behaviors, the mAP@0.5 was increased by 3.0 and 2.4 percentage points, respectively. The method presented can accurately identify fine behaviors of beef cattle under farming conditions, providing a reference for multi-behavior monitoring of beef cattle.

Abstract:In intelligent cattle farming, the application of facial recognition technology for cattle is crucial. However, due to the complex and variable farming environment and the non-planar characteristics of cattle faces, facial recognition technology often faces challenges such as insufficient recognition accuracy and poor robustness in practical applications. To address these issues, a FaceNet based on multi-scale fusion network (FMF) was proposed, which identified cattle through facial features. Firstly, the MSRCR method was used for color restoration preprocessing of the input cattle face images to reduce the impact of lighting on the FMF algorithm. Subsequently, a more lightweight MobileNetV3 was introduced into the main feature extraction network to reduce the model parameters and computation while ensuring high feature extraction capability. Finally, a composite dual-branch adaptive attention mechanism (M_CBAM) was proposed for multi-scale feature fusion. M_CBAM adjusted the weighting coefficients based on important features of the feature maps, adaptively weighting serial CBAM and parallel CBAM weights, then performed multi-scale feature fusion of local fine features and global features of cattle faces, improving facial recognition accuracy. To explore the effectiveness and real-time performance of the proposed algorithm, ablation experiments were conducted on a self-made cattle face dataset, and the results were compared with current mainstream recognition algorithms. Finally, the algorithm was deployed on the Jetson AGX Xavier embedded platform for application testing. The test results showed that the proposed algorithm achieved an accuracy of 93.86% and an FPS of 30.02 f/s on an open test set collected from other dairy farms. Under the condition of fast model inference speed, the recognition accuracy was significantly better than that of the original network and comparison networks.

Abstract:The accurate measurement of granular fertilizer mass flow rate presents a significant challenge for achieving precision fertilization. Due to the complex dynamics of granular flows, traditional measurement methods often struggle to handle the resulting signals, which are typically characterized by significant noise and non-stationarity. To address this issue, a novel pre-processing algorithm for granular fertilizer flow signals acquired by a microwave sensor was proposed. Initially, a model of the fertilizer discharge pipeline was constructed and simulation experiments were conducted. These simulations determined that under conditions of a 400mm pipeline length and 30mm diameter, the impact of particle collisions within the pipeline on velocity was minimized. Based on these simulation results, a signal acquisition platform for granular fertilizer mass flow was established. Through analysis of the acquired flow signals, a hybrid denoising algorithm combining variational mode decomposition (VMD) and discrete wavelet transform (DWT) was innovatively proposed. Experimental results demonstrated that after processing with this algorithm, the signal-to-noise ratio of the signal was increased by 7.3757dB, and the root mean square error was decreased by 57%. The algorithm effectively separated the noise component from the signal, thereby significantly improving the measurement accuracy of the granular fertilizer mass flow rate. Finally, real-world vehicle tests were conducted. The processed flow signals exhibited a maximum relative error of 13.84% and a minimum relative error of 1.36%. The results confirmed that the experimental platform developed and the proposed denoising algorithm provided reliable technical support for precision fertilization.

Abstract:In order to improve the accuracy of unmanned agricultural machine’s perception of obstacles in the farm environment, to solve the problem that visual detection is easily affected by light and millimeter-wave radar detection is easily affected by vehicle bumps, etc., as well as the problem that the visual target detection algorithm has a large number of parameters, a large amount of computation, and a large volume of the model under the complex field, this paper proposes an obstacle-avoidance target detection method for unmanned agricultural machines with the fusion of visual and millimeter-wave radar information. Part of the radar target data from millimeter-wave radar is first filtered, and a target tracking algorithm based on adaptive extended Kalman filtering is proposed. Then a farm environment obstacle dataset is produced and a target detection network based on improved YOLO v5s is constructed. Subsequently, the mapping of radar points into the image pixel coordinate system is realized by time-stamp alignment and coordinate transformation with direct linear calibration method. Finally, the fusion model of obstacle detection information from millimeter-wave radar and visual sensors is constructed through the decision-level fusion method and target matching strategy, and the experimental results show that the mean average accuracy of the improved YOLO v5s is 97.0%, which is similar to that of the original model, but the number of parameters, the amount of computation, and the size of the model are only 40.2%, 39.2%, and 38.2% of the original YOLO v5s model, respectively. Compared with YOLO v4-Tiny, YOLO v7-Tiny, YOLO v4 and YOLO v7 models, it can better balance the detection accuracy and speed. The results of multi-scene tests show that the fusion method proposed in this paper improves the recognition accuracy by 2.67 percentage points and 15.07 percentage points compared with radar and camera in daytime tests, and the fusion detection method can effectively make up for the failure of the camera in nighttime tests, and has better robustness and accuracy than the single-sensor algorithm, and the fusion obstacle avoidance system effectively realizes the parking avoidance of the unmanned agricultural machine.

Abstract:Aiming to address the core challenges faced by agricultural robots in autonomous navigation and path planning under complex farmland environments—particularly in irregular plots and soft, sticky paddy fields where common issues included low coverage rate, excessive turning redundancy, and poor path continuity, a full-coverage path planning algorithm tailored for convex polygonal farmlands and compatible with dual-helix driven agricultural robots was proposed. The algorithm integrated both the geometric characteristics of the farmland and the kinematic constraints of the robot. The working area was initially divided into a central parallel-working zone and a peripheral contour-following zone, corresponding respectively to strip-based traversal paths and boundary-parallel strip traversal paths. The complete path was further segmented into working and non-working paths, including center-region connection paths, inter-region transitions, contour-parallel region transitions, and entry/exit paths. The methodology for generating each type of path was detailed. An optimal working direction was determined via the “long-side-first” principle for the central region, and the Warnsdorff rule was applied to optimize the sequence of strip traversal. Dubins and Reeds-Shepp path models were introduced to enhance continuity and feasibility during turns. Considering metrics such as path length, computation time, and coverage rate, the algorithm was implemented on a dual-helix driven robot and tested in simulations and field experiments across four typical farmland types: rectangular, trapezoidal, irregular quadrilateral, and irregular polygonal plots. Experimental results demonstrated that the proposed algorithm achieved a coverage rate between 96.69% and 97.80%, with a path overlap rate controlled within 1.77%~2.30%. The generated paths were continuous and boundary-safe, indicating strong execution stability and environmental adaptability. The research result can provide a robust and adaptable path planning solution for agricultural robots operating in complex environments, with promising engineering feasibility and application potential.

Abstract:Traditional visual navigation methods struggle to handle interference from varying lighting conditions and weed occlusion in complex pear orchard environments. An improved navigation line extraction method was proposed based on YOLO v8 to address this challenge. The process integrated a multi-scale large kernel attention (MLKA) module into the YOLO v8 model to enhance the perception of trunk features. A multi-frame feature point fusion mechanism was designed. By recording and utilizing feature points detected in five consecutive frames, this mechanism effectively compensated for the problem of insufficient feature points in single-frame detection. Additionally, the random sample consensus (RANSAC) algorithm was introduced to denoise the feature points of the left and right tree rows, respectively, and the least squares method was used for the two side tree rows’ line fitting. The navigation line was ultimately generated by calculating the angle bisector of the fitted lines from both sides of the tree rows. Experimental results showed that the improved model achieved a precision of 89.8%, a recall of 79.9%, and a mean average precision (mAP50-95) of 55.1% in tree trunk detection tasks under various lighting conditions and weed occlusion scenarios. The navigation line generated by combining multi-frame feature point fusion with RANSAC denoising exhibited an average angular deviation of only 1.17° from the manually annotated reference navigation line, an average position deviation of 20.40 pixels, and an average root mean square deviation of 0.27. The research result can provide a low-cost and highly adaptable technical solution for visual navigation in pear orchard environments.

Abstract:Intercropping of grains and oils, along with plastic film coverage, serves as an effective agricultural practice adapted to specific natural conditions. This practice holds significant importance in dryland agricultural production in the northern regions. The application of these measures is bound to influence the changes in the carbon cycling processes of farmland soil. However, the variations in productivity and carbon footprint under intercropping and plastic film coverage, along with the associated ecological effects, remain unclear. Totally five treatments, including maize mulching (FMM), maize without mulching (NMM), soybean monoculture (SS), maize-soybean intercropping + maize mulching (FMS), maize-soybean intercropping + maize without mulching (NMS) were set up. Based on the life cycle assessment (LCA) method, the characteristics of greenhouse gas emissions, content of soil organic carbon, corn yield, and other factors from farmland soil under maize-soybean intercropping and plastic film coverage were evaluated. It also comprehensively assessed the impact of these practices on farmland productivity and carbon footprint through indicators, such as land productivity and carbon sustainability index. The results indicated that the relative yield of FMS treatment was increased by 57% to 60% compared with that of monoculture treatment, with a land equivalent ratio of 1.08 over the past two years, effectively enhancing farmland productivity. Simultaneously, the soil organic carbon loss in the FMS treatment was reduced by 7.8% to 46.57% compared with that of other treatments. It exhibited the lowest carbon footprint per unit area (CFA), carbon footprint per unit maize equivalent yield (CFMEEY), and carbon footprint per unit total income (CFE), effectively preserving soil organic carbon while reducing greenhouse gas emissions, thereby alleviating environmental burdens and carbon emissions. Although FMS treatment reduced some crop yield and economic benefits compared with monoculture systems, it demonstrated relatively lower soil greenhouse gas emissions and organic carbon loss during the production process. It also had a higher carbon sustainability index compared with NMS treatment, achieving effective management of soil carbon emissions. FMS as an environmentally friendly planting model, it can realize low carbon and high yield, thus providing guidance for agricultural producers in selecting suitable planting systems.

Abstract:In order to optimize the irrigation system of tomato under film drip irrigation in solar greenhouse in Northwest China, two experiments in spring and autumn in 2023 were conducted, four irrigation amounts (W1: 125% ETc, W2: 100% ETc, W3: 75% ETc, W4: 50% ETc, ETc is crop evapotranspiration) and three irrigation frequencies (D1: every 4 days, D2: every 7 days, D3: every 10 days) were set, and the effects of irrigation amount and irrigation frequency on soil moisture content distribution, root architecture, photosynthetic fluorescence characteristics, yield quality and water productivity of tomatoes were systematically evaluated. The results showed that irrigation amount and irrigation frequency had significant effects on root architecture, photosynthetic fluorescence characteristics, yield quality and water productivity (P<0.05). High-frequency irrigation (D1) significantly increased the soil water content in the 0~60cm soil layer, and the soil moisture content in autumn was higher than that in spring. Appropriately increasing irrigation amount and high-frequency irrigation was beneficial to the growth of tomato roots, improving the photosynthetic fluorescence parameters of tomato leaves, and thus promoting the improvement of photosynthetic efficiency. The total root length and root surface area of W2D1 treatment were significantly higher than those of other treatments（P<0.05), while excessive irrigation was not conducive to root growth. The net photosynthetic rate, maximum photochemical efficiency of the photosystem, actual photochemical quantum yield of PSⅡ and relative content of leaf chlorophyll in the W2D1 treatment were maximized. Appropriately increasing irrigation amount and high-frequency irrigation could increase yield (5.74%~57.78%). High-frequency irrigation could improve water productivity, increase soluble solids content, vitamin C content, soluble sugar content, and titratable acid content in tomato fruits, thereby improving tomato quality. A comprehensive evaluation of underground root characteristics, leaf photosynthetic fluorescence parameters, yield and quality, and water productivity by using the TOPSIS multi-objective comprehensive evaluation method revealed that the W2D1 treatment had the highest relative similarity in comprehensive indicators, resulting in the optimal comprehensive evaluation. The research results can provide a theoretical basis and technical support for the formulation of efficient water-saving irrigation system for greenhouse tomatoes under mulched drip irrigation in Northwest China.

Abstract:Aiming to investigate the effects of reduced nitrogen fertilizer with topdressing on growth and osmoregulatory substances in peanut under mulched drip irrigation, a split-plot experiment was conducted at the Aerxiang Irrigation Experimental Station in Liaoning Province from May to October in 2023 and 2024. The effects of different nitrogen application modes (N0, no nitrogen;N1, 25% nitrogen reduction with single application;SN1, 25% nitrogen reduction with three times application;and N2, conventional nitrogen rate with single application) on main stem height, root morphology, osmoregulatory substances, yield, and water use efficiency of peanuts under mulched drip irrigation (IMDI) and drip irrigation (ICK) were analyzed. Results indicated that in 2023 and 2024, the IMDI treatment consistently exhibited greater main stem height than ICK treatment across all nitrogen modes. SN1 obtained the highest main stem height, and followed by N2 and N1, N0 had the lowest main stem height. During the flowering and pod setting stage, the IMDISN1 treatment achieved the optimal root morphology. In 2023, at the flowering stage, compared with IMDIN1 and IMDIN2 treatments, the IMDISN1 treatment increased root length by 43.9% and 36.6%, respectively. In 2024, at the flowering stage, compared with the IMDIN1 and IMDIN2 treatments, the IMDISN1 treatment enhanced root surface area by 52.2% and 17.3%, and root volume by 49.5% and 25.0%, respectively. Compared with the ICKN2 treatment, the IMDISN1 treatment increased root length and root volume by 41.8% and 45.5%, respectively. Under drought stress, SN1 treatment had the highest proline content, and followed by N1 and N2 treatment, the N0 treatment had the lowest proline content. During the pod setting stage in 2024, compared with the IMDIN1 and IMDIN2 treatments, the IMDISN1 treatment increased proline content by 18.1% and 34.8%, respectively. During the pod setting stage in 2024, the N2 treatment had the highest soluble protein content and followed by the SN1 and N1 treatment, the N0 treatment had the lowest soluble protein content, the IMDIN2 treatment had 27.4% and 40.6% higher soluble protein content than the IMDISN1 and IMDIN1 treatments, respectively. During the flowering stage in 2024, the SN1 treatment obtained the highest soluble sugar content, and followed by the N1 and N2 treatment, the N0 treatment had the lowest soluble sugar content, compared with the IMDIN1 and IMDIN2 treatments, the IMDISN1 treatment increased soluble sugar content by 30.3% and 40.0%, respectively. The IMDI treatment had higher peanut yield than the ICK treatment, but the IMDISN1 treatment had similar peanut yield with ICKN2 treatment. Compared with the ICKN2 treatment, the IMDISN1 treatment increased water use efficiency by 32.3% (2023) and 41.1% (2024), respectively. Thus, reducing nitrogen rate by 25% with three split applications under mulched drip irrigation (IMDISN1) could improve water and nitrogen utilization, enhance osmoregulation during the critical water requirement stages, and ultimately promote peanut yield. These findings could provide theoretical basis for water-saving and nitrogen reduction in peanut production in sandy regions.

Abstract:In the spring (April), summer (July), and autumn (October) of 2019 and 2021, samples were collected six times from 99 points in the Ningxia section of the Yellow River Basin. These 99 points were categorized into eight hydrological networks: the main stream of the Yellow River (S1), Qingshui River (S2), Kushui River (S3), drainage ditch on the north bank of the Yellow River in the Weining section (S4), drainage ditch on the south bank of the Yellow River in the Weining section (S5), drainage ditch on the west bank of the Yellow River in the Qingshi section (S6), drainage ditch on the east bank of the Yellow River in the Qingshi section (S7), and urban landscape rivers and lakes (S8). The spatial and temporal distribution characteristics of fluorine in sediments were analyzed, the ecological risk of sedimentary fluorine was evaluated, and its impact on fluorine pollution in surface water was discussed. The results showed that the average contents of water-soluble fluorine (Ws-F), exchangeable fluorine (Ex-F), total fluorine (T-F) in sediments from the Ningxia section of the Yellow River Basin were 9.42mg/kg, 3.23mg/kg, and 517mg/kg, respectively. The evaluation results of ERbc and ERbf can better reflect the actual ecological risk. The contents of Ws-F, Ex-F, and T-F in the sediments of S3, S2 and S7 and the ecological risk of fluorine were all high, while those of S1 were low. Ws-F was an important factor that directly affected the F- content in surface water. T-F indirectly affected the F- content of surface water by transforming into Ws-F. The strong release capacity of fluoride in sediments of S3, S2 and S7 promoted the increase of fluoride content in surface water, which should be taken as the key control area. The dilution effect of S1 and the weak release ability of fluoride in sediments of S6 reduced the fluoride content in surface water.

Abstract:An indoor and a field experiment were conducted to study the water-stable large aggregates, soil organic carbon content, saturated hydraulic conductivity, hydrogen isotope distribution, physiological characteristics of wheat photosynthesis and water utilization under long-term application of organic fertilizer and chemical fertilizer, thus ascertaining soil moisture distribution and water use mechanism of long-term application of organic fertilizer. The results showed that the organic fertilizer significantly increased the proportion of soil macro-aggregate, soil organic carbon content, saturated hydraulic conductivity, and thus led to the high stability of soil structure. The abundance of hydrogen and oxygen decreased gradually with the deepening of the soil layer. In the jointing period, the long-term application of chemical fertilizer mainly used the water in 0~60cm soil layer, and the yield of organic fertilizer mainly used the water of 0~20cm soil layer, while the contribution rate of soil moisture of 0~20cm soil layer was significantly higher than that of other soil layers, reaching more than 50%. In the heading period, the long-term application of chemical fertilizer and organic fertilizer mainly used the water in 0~20cm soil layer, and the contribution rate of soil moisture in this soil layer was still more than 50%. In the filling period, the long-term application of chemical fertilizer mainly used the water in 0~20cm soil layer. The long-term application of organic fertilizer mainly used the water in 20~40cm soil layer, and its contribution rate reached 50.2%. With the advancement of wheat growth period, the soil water storage decreased gradually. Long-term application of organic fertilizer was more conducive to reducing the ineffective loss of soil water in the early stage of wheat growth, which improved the soil water storage, and then promoted the supply of water in the late stage of wheat growth. Compared with application of chemical fertilizer, long-term application of organic fertilizer was more conducive to improving the photosynthetic rate, stomatal conductivity and leaf water use efficiency of wheat in different growth periods. With the advancement of wheat growth period, its aboveground biomass gradually increased and above-ground biomass of organic fertilizer was higher than that of chemical fertilizer. Ultimately, long-term application of organic carbon content significantly improved wheat yield and water use efficiency. Corcorrelation analysis showed that the water supply of 40~60cm and 80~100cm soil layers at jointing stage, and 20~80cm soil layer at filling stage of wheat were more conducive to improving wheat yield, while soil supply of 40~60cm, 80~100cm at jointing stage, 60~100cm at heading stage, and 80~100cm at filling stage of wheat were more favorable to improve the water use efficiency. It concluded that the long-term application of organic fertilizer improved the soil structure, enhanced the storage, conservation and utilization capacity of precipitation in the soil, improved the photosynthetic physiological characteristics of wheat, and promoted the increase of yield and water saving and efficiency.

Abstract:In order to address the problems of tailwater treatment in fishpond aquaculture in the Ningxia region and the scarcity of water resources in rice-crab co-culture systems, the bacterial community structure and soil environmental factors of a fishpond-rice-crab integrated co-culture model (RS) were compared and contrasted with the traditional Yellow River-irrigated rice-crab co-culture model (CK) as control. The primary environmental factors influencing alterations in the organization of the bacterial population were also identified. The composition and structural variations of soil bacterial communities under both CK and RS models from April to August were examined, together with their associations with soil physicochemical markers, using 16S rRNA high-throughput sequencing techniques. The findings demonstrated that the RS model changed the species richness of soil bacteria while decreasing soil pH value and raising ammonium nitrogen and accessible phosphorus levels. The greatest Chao1 and Shannon diversity indices appeared in April, while the lowest appeared in May. Proteobacteria, Bacteroidota, and Firmicutes were the most prevalent bacterial phyla from April to August, with Pseudomonas being the most prevalent genus. The cumulative relative abundance of Pseudomonas was decreased by 7.61 percentage points, Thiobacillus was increased by 4.94 percentage points, and Unspecified_Bacteria was increased by 1.83 percentage points as compared with the CK model;changes in other taxa were within ±1 percentage point. Redundancy analysis showed that the most important variables influencing the top ten bacterial phyla were pH value, total nitrogen, ammonium nitrogen, total potassium, and organic matter concentration. Total potassium exhibited the highest explanatory power (23.67%) among them, exhibiting negative relationships with other phyla and positive correlations with Proteobacteria. In general, the top ten phyla showed a positive correlation with nitrate nitrogen concentration. As a result, the fishpond-rice-crab integrated co-culture paradigm optimized the makeup of the bacterial community, boosted bacterial species diversity, and greatly improved soil environmental conditions and structure.

Abstract:This Meta-analysis integrated 102 published domestic and international studies on biochar and humic acid application from 2019 to 2024. It examined their effects on crop yield and physicochemical-biological indicators of the 0~20cm soil layer across three dimensions: application type (X1), application rate (X2), and application duration (X3). Results indicated that biochar and humic acid application significantly increased crop yield, soil pH value, porosity, total nitrogen content, organic matter content, and alkali-hydrolyzable nitrogen content （P<0.01), while significantly reducing bulk density, electrical conductivity, available phosphorus content, and urease activity. The BRT model revealed the contribution order of factors as X2>X3>X1. The most pronounced yield increase occurred at application rates exceeding 40t/hm2 and application durations exceeding 4 years, with respective yield increases of 20.43% and 18.4%. Application rates of 0~20t/hm2 significantly increased available phosphorus, alkali-hydrolyzable nitrogen, urease content, and catalase activity by 28.29%, 10.01%, 12.37%, and 2.69%, respectively. Application rates of 20~40t/hm2 increased pH value (3.77%) and nitrate nitrogen content (47.85%);application rates exceeding 40t/hm2 significantly reduced bulk density (8.66%), while increasing porosity (15.75%), total nitrogen content (20.31%), and organic matter content (40.51%). Application duration of 0~2 years significantly reduced electrical conductivity (17.92%) and ammonium nitrogen (2.70%), but increased soil organic matter content (33.60%), available phosphorus content (26.07%), nitrate nitrogen content (6.88%), and catalase activity (4.51%);Application duration of 2~4 years increased soil pH value (16.80%), porosity (4.92%), and urease activity (17.43%);Application periods exceeding 4 years significantly increased alkali-hydrolyzable nitrogen content (22.89%) and total nitrogen content (22.49%) （P<0.05). The research quantitatively identified optimal application strategies for biochar and humic acid, providing theoretical foundations for optimizing their agricultural applications and practical guidance for farmland soil improvement.

Abstract:High-throughput sequencing technology, non-metric multidimensional scaling (NMDS), neutral community model (NCM) and structural equation modelling (SEM) were used to analyze sediment physicochemical properties, bacterial structural features, community assembly, and driving mechanisms in order to examine the bacterial community structure characteristics and community assembly processes of sediments in the Second Drainage Channel of Yinchuan City. The findings indicated that whereas total nitrogen (TN) and ammonium nitrogen (NH4N) concentrations did not substantially correlate with α-diversity indices, electrical conductivity (EC) did considerably positively correlate with Chao1, Observed, Shannon, and Simpson diversity indices. Thiobacillus, Bacteroidetes_vadinHA17, and Dechloromonas were the major genera in the bacterial phyla Proteobacteria and Desulfobacterota. Results from PER-SIMPER, DNCI, and the NCM model consistently showed that stochastic processes dominated the bacterial community assembly between August 2021 and August 2022. EC, NH4N, and total phosphorus (TP) were the main drivers of the neutrality model’s stochasticity (R2) and the stability of the co-occurrence network structure, according to SEM analysis. Co-occurrence network stability was specifically directly and significantly impacted negatively by EC （r=-0.369) and positively by TP （r=0.504). The research revealed the bacterial assembly methods, co-occurrence network topology, and the driving impacts of physicochemical parameters on R2 and network stability, in addition to advancing the understanding of structure characteristics of bacterial communities in aquatic sediments.

Abstract:In order to improve the performance and accuracy of discrete element method simulation optimization in the processes of blanking, conveying, and mechanical spreading of the shrimp feed pellet, the intrinsic parameters of the shrimp feed pellet, the contact parameters between feed, the contact parameters between feed and equipment were systematically calibrated. Taking Tongweis columnar sinking shrimp feed pellet as research object, the DEM simulation parameters of the shrimp feed pellet were calibrated by combining physical and simulation experiments. Firstly, by using the physical test method, the density, Poisson’s ratio, elastic modulus and shear modulus of the shrimp pelleted feed were measured to be 1177.68kg/m3, 0.26, 114.77MPa and 45.58MPa, respectively through instruments such as high precision digital calipers, electronic balances and pressure testing machines. Through inclined plane method and rebound method, the static friction coefficient, rolling friction coefficient, and restitution coefficient between the shrimp feed pellet and 304 stainless steel were determined to be 0.607, 0.097 and 0.35, respectively. The average stacking angle of the shrimp feed pellet was found to be 26.93°, and the feed’s area coverage on a steel plate was 13.32% through stacking tests. Secondly, based on the Plackett-Burman experiment, the factors significantly affecting the physical stacking angle and area proportion were screened out to be the static friction coefficient between feed, the rolling friction coefficient between feed, and the static friction coefficient between feed and 304 stainless steel. Based on this foundation, the steepest ascent experiment was applied to converge the optimization range, and the parameter optimization was carried out by fitting the regression equation between the significant influencing factors and the response values through the Box-Behnken Design experiment. The optimal parameters combination were obtained as follows: the static friction coefficient between feed was 0.207, the rolling friction coefficient between feed was 0.150, and the static friction coefficient between feed and 304 stainless steel was 0.607. Finally, the reliability of the parameters was verified through stacking tests and spiral blanking tests. The findings revealed that the simulated stacking test results were statistically indistinguishable from the physical counterparts, as confirmed by a double-sample t-test. Furthermore, the correlation coefficient between the simulated and physical tests for the discharge amount line under varying speeds was an impressive 0997, signifying a highly consistent trend with a discrepancy in discharge amount of less than 3% across all speeds. These results underscored the reliability of the calibrated DEM parameters for shrimp pellet feed, thereby establishing a solid data foundation and scientific rationale for investigating the motion mechanisms of shrimp feed pellets and optimizing equipment designs within DEM simulations.

Abstract:The traditional winemaking process of grading and storage is a crucial link that affects the quality of Chinese baijiu blending and finished wine, but the grading and storage process of Chinese baijiu are mainly relied on workers experience in traditional winemaking factories, causing difficulties in many characteristic quantification and form scientific basis for grading storage. Based on deep learning technology, a gas chromatograph was used to quantitatively detect and analyse the content of flavour components in the base wine of small-curve clear-flavoured Chinese baijiu, and a regression model for the quality grading of the base wine was established using PLSR based on the content of the components detected in the base wine,screening the first eight components that have the most significant impact on the quality grading of the base wine: ethyl acetate, ethyl lactate, ethyl valerate, isobutyl alcohol, ethyl caprylate, ethylene acetal, ethyl acetate, sec-butanol, n-propanol, and then the PSO-RBF neural network-based base wine quality classification model was explored. Finally, according to the predicted outputs of the test samples, our algorithm achieved 98.6% accuracy in base wine classification, which was significantly better than the 943% of the traditional RBF neural network. As a result, the base wine classification grading model was successfully constructed.

Abstract:Milk tea beverage is a drink widely loved by consumers. The astringency caused by the rich tea polyphenols in the tea soup is one of the key factors affecting the sensory quality of milk tea beverage. Existing research has shown that milk components, such as milk proteins, can reduce astringency. Milk fat, an important component in cow’s milk, is composed of milk fat globules. However, the effect of its content and structure on astringency in milk tea systems remained unclear. It was aimed to investigate the impact of milk fat content and fat globule structure on astringency in milk tea systems. Firstly, it used descriptive sensory evaluation to explore whether milk fat and fat content affected the astringency of milk tea system. Next, three types of milk fat droplets with the same fat content but different structural characteristics were constructed by adjusting the heating intensity and homogenization pressure: milk fat droplet C, milk fat droplet T and milk fat droplet H, and the particle size, ζ-potential, microscopic structure, milk fat globule membrane protein content and composition, and milk fat globule membrane phospholipid content were measured, which characterized the characteristics of milk fat globules. Then different milk fat droplets were added to the simulated milk tea system to conduct sensory evaluation of astringency intensity, and quantify astringency through fluorescence spectroscopy analysis. The results showed that when the milk fat content reached above 4g/(100mL), the addition of milk fat can significantly reduce the astringency level in milk tea. In addition, changing the particle size and membrane structure of milk fat globules also had a significant impact on the astringency of the milk tea system. Compared with the other two milk fat droplets, the fat globule size in milk fat droplet H was the smallest (0.49μm), the protein content per gram fat globule membrane was the highest (12.60mg), and the astringency was the weakest. Milk fat droplet T had the largest particle size (4.20μm), the lowest protein content per gram fat globule membrane (4.12mg), and a strong astringency. Therefore, milk fat content, particle size and membrane structure would affect the astringency intensity of milk tea system. Smaller milk fat globule size and more milk fat globule membrane interface proteins had the greatest effect on reducing the astringency intensity of milk tea system. It was confirmed the impact of milk fat on the perception of astringency in the milk tea system, laying a theoretical foundation for the development of green and healthy milk tea drinks with a smoother taste.

Abstract:In response to the problems of steep slopes and narrow roads, complex terrain and landforms, poor applicability of existing models on sloping land, and lack of design theory for mountain power machinery in hilly and mountainous areas, this paper proposes a “lateral swing longitudinal shift” center of gravity adjustment crawler tractor scheme suitable for hilly and mountainous areas. In order to reduce research and development costs and shorten the research and development cycle, a “lateral swing longitudinal shift” center of gravity adjustable remote control crawler proportional prototype is specially created for feasibility verification and slope stability research of the scheme. A mathematical model and force analysis were conducted for slope driving before and after adjusting the center of gravity, and a stability index that considers both lateral and longitudinal directions-the uniform distribution coefficient of track contact pressure-was proposed. Grounding pressure and passability tests were conducted, and the results showed that the “lateral swing longitudinal shift” center of gravity adjustable scheme proposed in this paper can greatly improve the uniformity of grounding pressure distribution for two tracks traveling on slopes;Under the slope conditions of 5°, 10°, 15°, and 20°, after adjusting the center of gravity of the whole machine longitudinally and laterally, the maximum traction force of the vehicle when inclined 45° uphill is 398.70N, 339.41N, 265.67N, and 222.32N, respectively. The maximum traction force has increased by 13.9%, 18.1%, 27.3%, and 50.5% compared to before regulation. A maximum rollover test was conducted, and the results showed that when the longitudinal distance of the upper body extended to a maximum of 150mm, the static longitudinal maximum rollover angle increased from 39° to 46°, an increase of 17.9% compared to before the center of gravity adjustment;When the upper body sway angle is adjusted to the maximum value of 25°, the static lateral limit tilt angle is increased from 40.0° to 43.5°, an increase of 8.8% compared to before regulation;When passing obstacles on a slope, adjusting the horizontal or vertical center of gravity will increase the dynamic tipping coefficient of the vehicle on the slope, and the lifting effect will weaken with the increase of slope. Prove that the proposed solution is feasible and can significantly improve the driving stability and anti rollover ability of tracked vehicles on slopes, providing theoretical and data support for performance analysis and structural optimization design of tracked vehicles.

Abstract:The design parameters of the transmission directly affect the dynamics and fuel economy of the vehicle, due to the complexity of the tractor’s working conditions and a variety of operating modes, so in the optimization of the design parameters of the tractor transmission cannot be directly borrowed from the automobile based on the optimization of the standard cycle of working conditions. An independently designed hydro-mechanical continuously variable transmission (HMCVT) was taken as the research object, and a transmission design parameter optimization method based on the whole life cycle speed usage rate of the tractor was proposed, and a multi-objective genetic algorithm (MOGA) was used for optimization and solution. Fuel consumption rate and climbing ability were used as the optimization objective functions, and parameters significantly affecting the optimization objectives were selected as design variables. Constraints were determined, and a multi-objective optimization model for the transmission device was established by using modeFrontier, combining MOGA based on Pareto optimal principles and experimental design, to perform a global search for optimization, resulting in Pareto optimal solutions. The research results obtained the optimal solutions under the constraints, reflecting the contradictory characteristics between dynamic performance and fuel economy. The design variables of the Pareto optimal solutions obtained through optimization iterations well met the matching requirements of the transmission device.

Abstract:A multi-mode mobile parallel mechanism with over constrained branch chains was designed, and its degrees of freedom and kinematics were deeply analyzed. When exploring the motion of the parallel mechanism, the influence of the over constrained branch chain on its motion stability was studied, and the motion simulation and physical production were carried out to verify the motion of the mechanism and the realization of each mode. Firstly, according to the analysis of the connection between the constrained branch chain and the platform branch chain, a multi-mode mobile parallel mechanism with constrained branch chain was obtained according to its configuration design. The degree of freedom of the two modes of the mechanism, quadrilateral rolling mode and spherical rolling mode, was analyzed by means of the bar group model. The comparative analysis of the mechanism with or without over constrained branch chains was carried out, the centroid coordinates were solved according to ZMP theory, the relationship between the mechanism with or without over constrained branch chains and the angle was compared and analyzed, and Matlab was used to carry out comparative analysis of the centroid situation of the mechanism with motion. Finally, the kinematic simulation of the mechanism was carried out by ADAMS, and the prototype was made to verify the movement of the mechanism and the realization of each mode.

Abstract:A novel structure model of 2-RRC/1-RRS parallel manipulator mechanism was proposed for the screening of agricultural materials, and its kinematics and dynamic characteristics were studied. Firstly, based on the motion characteristics of each branch chain in the 2-RRC/1-RRS parallel manipulator mechanism, the manipulator structure characteristics and the coupling relationship between the motion variables were analyzed. Secondly, the kinematics of the 2-RRC/1-RRS parallel mechanism was deduced, and the analytical expression of the inverse kinematics of the robot was given. Then based on the discussion of kinetic energy and potential energy of the 2-RRC/1-RRS parallel manipulator, the dynamic equation of the manipulator was derived by Lagrange equation method of classical mechanics. Finally, the dynamic characteristics of the 2-RRC/1-RRS parallel manipulator mechanism, such as motion parameters, driving torque and energy consumption, were discussed by an example. The results showed that the parameters of moment of inertia in the dynamics equation of the 2-RRC/1-RRS parallel manipulator system were closely related to the structure size of the manipulator, the mass size and mass distribution of each component, and the motion configuration of the manipulator, so as to determine the kinematics and dynamics characteristics of the 2-RRC/1-RRS parallel manipulator. The research content was of great significance to deeply understand the kinematics and dynamic characteristics of 2-RRC/1-RRS parallel manipulator, as well as the structural optimization design and the high-precision control of the manipulator, which can provide an idea for the innovative design of new equipment for agricultural material screening.