Abstract:Crop growth models have evolved from initial crop development models to agricultural decision support models, playing an increasingly important role in scientific research, agricultural management, and policy-making. In the paper the development process of crop growth models was firstly reviewed. Based on the main driving factors, the models were categorized into three types: soil factors, photosynthetic factors, and human factors, and comprehensive introductions to each category were provided. Then a comparative analysis of typical models was presented from ten aspects, including model modules, spatiotemporal scales, and range of crop types that can be simulated. Furthermore, the applications of crop growth models in climate change assessment, production management decision support, and resource management optimization were discussed. The challenges faced by these models were also highlighted, such as extreme conditions, complex agricultural landscapes, and model complexity. Based on the comprehensive discussions, two promising directions for the future development of crop growth models were identified: remote sensing data assimilation and twin farming. Remote sensing data assimilation techniques have the potential to significantly enhance the spatial range and accuracy of the simulations, providing more precise information for agriculture. Twin farming, on the other hand, offers virtual replicas of actual farming systems, enabling comprehensive analysis and optimization of crop growth. These research findings provide valuable insights for selecting and improving crop growth models, driving advancements in this field.

Abstract:In response to the issues of accuracy and speed being difficult to balance simultaneously, as well as the weak generalization ability in the current navigation path recognition methods of fruit ridges, an optimization approach was proposed based on the U-Net model. The optimization involved integrating MobileNet-v3 Large as the backbone feature extraction network for U-Net and introducing coordinate attention at the skip connections to construct a lightweight path recognition model. Based on the drivable area segmented by this model in the inter-ridge, the edge points of the area were reshaped by using the least squares method, and further the inter-ridge navigation lines were extracted. Firstly, the model was trained on the augmented strawberry interrow dataset, and further migrated to the grape and blueberry datasets for weight fine-tuning to improve the model’s adaptability. Finally, the navigation path was identified on the corresponding verification set, and the recognition results of different models were compared visually to verify the accuracy of the model. Experimental results demonstrated that the model achieved an average intersection over union of 98.06%, 97.36%, and 98.50% for strawberry, blueberry, and grape interridge navigation path segmentation accuracy respectively, and the average pixel accuracy reached 99.13%, 98.75%, and 99.29%. The theoretical reasoning speed of the model for segmenting of RGB images was up to 19.23f/s, and the average time from image input to completed path extraction was 0.211s, meeting the requirements of real-time navigation and accuracy. A method of path extraction based on semantic segmentation was proposed, which provided a general method for the navigation of agricultural machinery equipment in interridge operation.

Abstract:Aiming at the problems of large plowing load and low measurement accuracy, a six dimensional force sensor of radial beam type was designed on the basis of classical cross beam structure, which could measure force and moment at the same time. The sensor structure was optimized by simulation method, and the dimension length, width and height of strain beam were determined to be 9mm,10mm and 6mm, respectively. The strain capacity of the sensor structure under load was analyzed, and the position of the strain gauge patch was determined. Based on the calibration data, the improved XGBoost (extreme gradient boosting) machine learning network was used to decouple the force signal. The improved XGBoost model achieved R2P (determination coefficient of test set) of 0.9804, 0.9418, 0.9434, 0.9868, 0.9969, and 0.9822 in six loading modes of force and torque in X, Y and Z directions, respectively. The prediction performance was good, avoiding getting stuck in local optimal solutions. And then compared with the conventional network, the R2P and MAEP (average absolute error of test set) of the improved XGBoost model in the six dimensional force loading direction were significantly better than that of the random forest model and the traditional multiple linear regression. Compared with the traditional multiple linear regression method, the R2P of the six dimensional loading force/moment was increased by 22.57%, 20.99%, 23.32%, 26.27%, 26.05% and 18.72%, respectively. Machine learning based decoupling algorithms could significantly reduce the impact of coupling errors and improve the measurement accuracy of sensors and provide technical support for optimizing agricultural machinery.

Abstract:With a large banana growing area in the tropics of China, about 35 million tons of waste straw will be produced each year, and the traditional way of dealing with banana straw is time consuming and polluting. Crushing banana straw back to the field is a green way of dealing with banana straw, and varieties of banana straw crushers have been developed in China. However, the diameter of banana straw in the mature stage can reach 250mm, banana straw crushing and return machine in the operation of uneven crushing, banana straw fiber entanglement serious, return machine operation quality and other problems still exist. In response to the problem that banana straw is not thoroughly crushed during the operation of banana straw crushing and returning to the field in tropical areas of China, and the straw block is entangled and blocked the key parts, which affects the normal operation of the machine, a double fixed knife slip-cutting type anti-tangle banana straw crushing and returning machine was designed. Based on the slip-cutting theorem, the dynamic slip-cutting angle during the rotation of the crushing knife with the shaft and the slip-cutting angle during the crushing process of the crushing fixed knife were analyzed, the L-shaped crushing fixed knife blade curve was designed according to the equal velocity spiral, and the structural parameters of the crushing knife were determined;the anti-winding plate was designed, the force analysis of the banana straw winding crushing knife roll was performed, and the number of anti-winding plate assembly and structural parameters were determined. The anti-winding plate was designed, the force on banana straw winding crushing knife roller was analyzed, the number of anti-winding plate assembly and structural parameters were determined;the anti-winding cover was used to improve the anti-winding performance at the knife roller connection;the device forward speed, crushing knife roller speed and anti-winding plate height were taken as test factors for banana straw crushing pass rate, throwing unevenness and the number of banana straw winding as three-level and three-factor orthogonal test, the establishment of factors and indicators of the response surface mathematical model was done, the relationship between factors and evaluation indicators was analyzed, and optimization of the influencing factors was carried out. The experiment results showed that the optimal combination of parameters for each test factor was as follows: the forward speed of the machine was 1.5m/s, the height of the anti-tangle plate was 41.6mm, and the speed of the crushing knife roller was 1800r/min, and then the banana straw crushing qualified rate was 93.8%, the number of banana straw winding was 26, and the unevenness of banana straw throwing was 12.1%. And the optimal combination of field test verification, with the optimization of the solution results of the error was small;the design of double fixed knife slide cut anti-tangle banana straw crushing machine whole machine anti-tangle performance was superior to satisfy the design requirements.

Abstract:In view of the production requirements of ridge cultivation of rice and lack of related machinery and equipment in cold waterlogged paddy fields, low water and soil temperature and poor drainage, a ridging and leveling device suitable for cold waterlogged paddy field was developed. The device was mainly composed of a ridging roller, a micro-ridge opener and a leveling component to achieve large ridge drainage and micro-ridge water storage operations, which was conducive to cold waterlogged paddy field drainage and rice seed growth and development. According to the technical regulations of mechanical ridging rice cultivation in cold waterlogged paddy field, the key parameters of ridging roller, micro-ridge opener and leveling parts were designed. The radius of rotation of ridging roller was 560mm, the chamfer of conical surface was 28°, the width of micro-ridge opener was 50mm, the height was 40mm and the adjustable angle of leveling parts was 5°. Based on the DEM-FEM coupling simulation analysis, the optimal operating parameters of the device were determined. When the forward speed of the machine was 0.6m/s, the rotation speed of the rotary blade roller was 230r/min, and the rotation speed of the ridge roller was 120r/min, the soil backflow rate was 3.51%, and the stress and strain of the ridge roller were analyzed. The results of the field experiment showed that the average furrow depth was 160.03mm, the width of top of furrow was 174.84mm, the average width of top of ridge was 1888.89mm, the average flatness of box was 11.26mm, the average width of micro-ridge was 60.16mm, and the depth of furrow was 36.48mm. All indicators met the requirements of mechanized ridge operation in cold-waterlogged fields. The research can provide a reference for the design of ridge leveling machine in cold waterlogged paddy field.

Abstract:Focusing on addressing the issues commonly found in existing seed metering devices, such as uneven seeding and inadequate hole forming when dealing with large sowing rates, a double-track ejection type rice precision direct seeding metering device was developed. Drawing from theoretical analysis, the key components were designed. Using the DEM-MBD coupling simulation technology, a range of spring force parameters and factors were obtained and the accuracy of the adjusting mechanism to meet the requirements of seeding was determined. It was evident that the performance of the seeder was decreased significantly after the rotation speed exceeded 35r/min. To investigate the effects of rotation speed, adjustment depth, and sphericity on the performance of the device, a bench orthogonal test was conducted. A regression prediction model of the device performance evaluation index was established. The test results showed that when the rotation speed of the seeding wheel was 23.06r/min, the depth of the hole was 8.99mm, and the sphericity of the rice seed was 52.7%, the qualified rate was 88.58%, the missed sowing rate was 4.43%, and the reseeding rate was 6.99%. To verify the working performance of the device and the accuracy of the optimized parameters, a field experiment was conducted. The results showed that the verification test results were consistent with the optimized results, and the error between them and the predicted results of the regression equation was less than 2%, thus confirming the feasibility of the experiment and the accuracy of the parameters. Under the optimum parameters, the qualified rate of hole diameter was 100%, the average hole diameter was 3.62cm, the coefficient of variation of hole diameter was 18.45%, the average hole distance was 22.98cm, and the coefficient of variation of hole distance was 8.43%, average number of holes was 11.08, and the coefficient of variation was 17.56%. The designed seed metering device had good seeding performance, high qualified rate of hole diameter and low coefficient of variation, indicating that the seed metering device had good hole forming performance.

Abstract:For corn and soybean belt compound planting conditions under the traditional mechanical seed metering device is not easy to realize both precision seed discharge requirements, the existing pneumatic seed metering device seed discharge speed due to leakage of holes leakage charging exists leakage sowing broken stripes and other issues, a cavity plate combined hole structure of the seeding plate was designed and analyzed to determine the key structural parameters of the seeding plate, and the adsorption process and suction and transportation process mechanics model was constructed. EDEM discrete element simulation and bench test were combined to carry out the seeding plate type selection test, and the results concluded that the cavity plate combined hole seeding plate had the effect of improving the average speed of directional transport of seed populations in the seed filling chamber and increasing the dragging angle of seed filling, which effectively suppressed the leakage rate of the hole. With the installation of the preferred seed plate of the jade and soybean hybrid seed metering device as the object, with the unit forward speed and negative pressure as the test factors, with the leakage rate and the seed filling rate as the test indexes, the two-factor full factorial experimental design was done to carry out the seed filling performance test, and the results showed that when the unit forward speed was 4.0~7.0km/h, the work of the negative pressure was 3.0~4.0kPa, the leakage rate of maize and soybean seeds was less than 3.6%, and the qualified rate of seed filling was not less than 96%. Field verification test showed that under the conditions of unit forward speed of 4.0~7.0km/h and working negative pressure of 3.0~4.0kPa, the leakage rate of corn and soybean seeds sown by the seed metering device of cavity plate combination hole-type seed metering device was no more than 3.8% and 4.2%, respectively. When the working negative pressure was 3.0kPa and the forward speed of the unit was 7.0km/h, the leakage rate of sowing corn and soybeans was decreased by 14.8 and 12.6 percentage points, respectively, compared with that of the undisturbed flat seeding plate with the combination of holes in the self-disturbing cavity plate.

Abstract:In order to solve the problem that it is difficult to test the performance of seeding unit profiling mechanism，a pitching seeding unit profiling performance testing bench was designed.The composition and working principle of the test bench were described，and the high-speed transmission system，hydraulic lifting system，monitoring system and key components were designed. ANSYS Workbench was applied to carry out static analysis and modal analysis on the whole test stand and key components to verify the reasonableness of the structural design.In order to test the actual testing effect of the pitching seeding unit profiling performance testing bench，taking the seeding unit of Debont 1205 tractor-type no-till precision planter as the research object，the first test was carried out with the hydraulic rod extension and conveyor belt speed as the test factors，and with the error of the monitoring system as the evaluation indexes.After confirming the accuracy of the monitoring system，the maximum error of angle sensor of the monitoring system was 0.69mm within the range of hydraulic rod extension of 0~200mm，and the maximum error of the photoelectric encoder was 0.18km/h within the range of conveyor belt speed of 8~19km/h. After confirming the accuracy of the monitoring system，taking the speed of the monocoque as the test factor，the data of the land’s undulation curve was collected at speeds of 8km/h, 10km/h and 12km/h for the objective of the test curve，the absolute error average of the terrain undulation simulation curve as an indicator of the single-factor test，the test concluded that the designed test bench can effectively simulate the field ground undulation frequency and the amount of undulation，the absolute error average value was 1.86mm, which can meet the needs of seeding monomer profiling performance testing.

Abstract:In order to solve the technical difficulties of lateral displacement and lateral deflection angle in the process of mechanically seedling picking process in wide and narrow rows of regenerated rice using existing space planetary gear transplanting mechanisms, a spatial trajectory unequal velocity planetary gear train mechanism was proposed with local planar trajectory characteristics, and a comprehensive research on the wide and narrow row transplanting mechanism was carried out based on the constraints of the position of rice gate and mechanism’s rotation center. A spatial gear train mechanism motion model based on key pose points (take seedling starting point, take seedling end point, push seedling point) was constructed, and the mechanism rod length parameters and spatial interleaved axis parameters was solved by using key pose points. The transmission ratio distribution of the mechanism was realized by optimizing the relative angular displacement parameters of the two rods. The incomplete non-circular gear pair was introduced into the space planetary gear system mechanism, and the intermittent mechanism was used to restrain the planetary axis to realize the plane trajectory segment with zero lateral offset and zero lateral angle during process of taking seedlings. Through simulation analysis and prototype test of the mechanism, the actual operation performance of the mechanism was consistent with the theoretical design, and the operation performance parameters of the mechanism were verified as follows: the transplanting mechanism had zero lateral offset during the process of picking seedlings, and zero lateral offset angle during the process of picking seedlings, the total lateral offset of pushing seedlings was 50.24mm, the picking angle was 5.18°, the pushing angle was 71.56°, the lateral angle of pushing seedlings was 16.26°, the width of the transplanting hole was 22.43mm, and the trajectory height was 289.76mm, which met the expected design requirements. Finally, the field experiment verified that the rice transplanting mechanism could achieve wide row (40cm) and narrow row (20cm) spacing plant under the equal distance taking position and the established rotation center of the mechanism, which met the requirements of wide row and narrow row planting.

Abstract:Aiming at the problems of the existing plant stalk clamping device, such as clamping injury, clamping stroke and unadjustable clamping force, a plant stalk flexible clamping device was designed. The built-in spring of the clamping hand can realize the flexible clamping of plant stalk, and double adjust the clamping force and clamping stroke, so as to better meet the demands of plant stalk clamping in the fields of mechanics experiment, grafting and cutting test. Based on the rigid-flexible coupling simulation model of the clamping device, the multi-factor dynamic comparison simulation and test analysis were carried out, including clamping force, clamping stroke and spring stiffness. The sensitivity and stability analysis were carried out with the spring stiffness as the design variable to study the influence of different spring stiffnesses on the clamping performance of the flexible clamping device. The simulation and experiment results showed that the maximum clamping diameter of the device was 68.8mm, and the clamping stroke, speed and clamping force can be controlled smoothly based on the accurate clamping force model of the clamping device. With the increase of spring stiffness, the clamping force was gradually increased, and the clamping stability would be increased. Comprehensive comparison showed that when the spring stiffness was 10N/mm, it can meet the stable clamping demand of plant stems when the device applied 400N cutting force, and the flexible clamping force had little fluctuation, which can effectively avoid damage to plant stems.

Abstract:Aiming at the problems of large size, cumbersome operational procedures, and poor maneuverability of agricultural chassis currently for hilly orchard, as well as the practical requirements of cultivation management tasks such as trenching, weeding, and pruning in hilly orchards, a hydraulic remote control crawler power chassis was designed. Firstly, the overall structure and working principle of the power chassis were described. Secondly, key components such as the front-mounted mechanism, walking system, variable-width chassis, hydraulic system, and remote control system were designed and appropriately matched. Finally, performance tests were conducted on the entire machine. The tests revealed that the deviation rates of the power chassis during straight-line travel at the minimum width (1220mm) and maximum width (1620mm) were 2.24% and 2.2% respectively, both satisfying the corresponding national standards (≤6%) requirements. The chassis exhibited good steering maneuverability, with a minimum width turning radius of 905mm, enabling it to adapt to the narrow slope working environment of hilly orchards. The remote control operation demonstrated smooth processes such as ascending and descending slopes, crossing field ridges, and traversing furrows, meeting the requirements of walking on unstructured terrain in hilly orchards. When equipped with a mounted chain-type ditcher for trenching operations, the stability coefficient of trench depth was 885%, and the stability coefficient of trench width was 92.5%, both meeting the national standards (≥85%) requirements. The overall performance of the machine met the demands of managing complex sloped terrains in hilly orchards, providing a comprehensive application platform and technical support for the effective implementation of field management operations in hilly mountainous orchards.

Abstract:Aiming at the problems of serious differentiation of residual film in the tillage layer, poor mechanical properties, low pick-up rate of deep residual film, and high soil content caused by the mixing of residual film and soil, an active recovery method of rotary tillage and throwing soil mixture, picking up spring tooth to forward rotate and pick up the film, and reversely rotate and unload film was proposed. The overall solution of the roll-type tillage layer residual film recovery machine was designed, and its structural composition and working principle were expounded. Design and parameter calculation of key operating components such as film picking device, film picking device, forward and reverse mechanism and film unloading device were carried out to obtain the critical conditions for the effective hook and picking out of the residual film in the mixture under the mechanical force of the spring-tooth. By using the coupling method of ANSYS and SPH (Smoothed particle hydrodynamics), the numerical simulation model of the spring tooth picking up the residual film process was constructed, and the maximum stress and deformation of the residual film in the process of picking up the residual film were obtained. The effectiveness of picking up the residual film with the spring-tooth hook was analyzed. The verification test of the prototype showed that when the rotation speed of the taking-film was 213.75r/min, the rotation speed of the picking drum was 43.75r/min, the rotation speed of the roll forward rotation was 131.27r/min, the rotation speed of the roll reverse unloading film was 167.86r/min, and the rotation speed of the unloading film wheel was 43r/min. The surface pick-up rate of the roll-type topsoil residual film recycling machine was 826%, and the deep pick-up rate was 71.1%. The test results met the design requirements and realized the mechanized recovery of topsoil residual film from the film-soil mixture. It provided a method and theoretical basis for the research of tillage layer residual film recovery mechanism.

Abstract:In order to solve the problems of existing grapevine soil clearing machines in Xinjiang region, such as high injury rate and a large amount of soil over the color-striped cloth affects the grapevine soil clearing on the shelves and recovery of the color-striped cloth, a kind of brush roll grapevine soil clearing machine was designed by using a flexible brush. Its core component of the machine was the soil clearing device, which removed the soil above the color-striped cloth through the rotating motion of the soil clearing brush and the reciprocating motion of the transverse telescoping mechanism. Firstly, the structure and material of the soil clearing brush was designed, then the kinematic theory analysis of the working process of the soil clearing brush was analyzed, and the main parameters affecting the operating performance of the soil clearing brush was obtained. The reasonable range of rotation speed of soil clearing brush, reciprocating speed of transverse telescopic mechanism and clearing brush rotation diameter were determined by single-factor test, taking the rotation speed of soil clearing brush, reciprocating speed of transverse telescopic mechanism and clearing brush rotation diameter as the influencing factors, and taking the soil clearing rate and grapevine injury rate as the evaluation indexes, the three-factor and three-level orthogonal test was carried out. The test results showed that the order of significance of each factor on the soil clearing rate and grapevine injury rate was rotation speed of soil clearing brush, reciprocating speed of transverse telescoping mechanism and clearing brush rotation diameter. The optimal combination of parameters was the rotation speed of soil clearing brush of 250r/min, the reciprocating speed of the transverse telescoping mechanism of 0.14m/s and the clearing brush rotation diameter of 600mm, and the average value of three validation tests under this parameter combination, the soil clearing rate was 90.98% and the grapevine injury rate was 3.27%. The research result can provide reference for the mechanization technology of grapevine clearing in Xinjiang.

Abstract:In response to the problems of low efficiency, poor quality, and inapplicability to high moisture content grass in existing grass kneading machines, a collaborative cutting and crushing grass kneading machine was designed for alfalfa processing. Theoretical analysis was conducted on the cutting and crushing processes, and the overall structure and key component structural dimensions of the wire kneading machine were determined through design calculations. A single factor experiment was conducted with motor output speed, mesh diameter, feed rate, and moisture content as experimental factors, using productivity and spinning rate as performance evaluation indicators. A quadratic orthogonal rotation combination experiment was conducted by using alfalfa with moisture content of 65% as the processing object, with motor output speed, sieve diameter, and feed rate as experimental factors. The single factor experiment determined the range of experimental factors and explored the silk kneading effect of the silk kneading machine on alfalfa with different moisture contents. Response surface analysis, regression analysis, and objective optimization were conducted on the experimental results of the quadratic orthogonal rotation combination experiment by using the Design-Expert 12.0 software. The regression equation between the experimental factors and evaluation indicators was obtained. With the goal of maximizing both productivity and silk rate, multi-objective optimization was carried out on the output speed, mesh diameter, and feed rate of the motor. The optimal parameter combination was determined as follows: the output speed of the motor was 443.77r/min, the sieve diameter was 14mm, the feeding amount was 1.27kg/s, and the validation test of wire rolling showed that the production rate was 5065.98kg/h, and the silk rate as 94.87%. The device had high silk kneading efficiency, good quality, which can knead grass with high moisture content, meeting the design requirements of the grass kneading machine.

Abstract:The existing field spraying system mainly control spray volume by the variable-rate technology, which is lack of pesticide spraying data remote monitoring and traceability management. A precision variable-rate monitoring and control system was designed based on pesticide spraying traceability. The system can realize precision variable-rate spraying, on-line monitor and real-time display different spray information, including spray plot, spray time, spray area, type and ratio of pesticide, spray volume, spray pressure, real-time spray flow and spray speed, which can be used to pesticide spraying traceability management. Based on the system, the tests of spray volume calculation accuracy, spray area calculation accuracy, data transmission stability of Internet of things, dynamic response of variable-rate control system, variable control precision and pesticide spraying uniformity were carried out respectively. The test results showed that the maximum error of Beidou positioning speed measurement was 1.33% and the mean error was 0.82%, the calculation error of spray volume was 1.73%, the calculation error of spray area was 2.61%, the data loss rate was 3.51%, the system stability adjustment time under continuous speed change was 4~5s, the variable-rate regulation accuracy was 245%, the density of droplet deposition was greater than 20 drops/cm2, and the variation coefficient of spray coverage in sprayer walking and spray direction was less than 10%, which met the operation requirements of precision variable-rate spray. The research can realize the traceability management of the pesticide spraying data under the precision variable-rate adjustment of spray volume, and provide support for the pesticide residue risk assessment in field crops.

Abstract:In response to the lack of connected greenhouse plant protection spraying machines, low precision in mechanical straight-line positioning and rail switching, a segmented and variable distance spraying robot was designed for multi-span greenhouses to achieve unmanned spraying while improving operational precision. To meet the requirements of combining road and track operation and precise switching for mechanical operations in multi-span greenhouses, a universal mobile chassis was proposed with its key design parameters determined. To reduce deviations during chassis movement along the upper and lower rails, a rail correction device was designed. Through analysis, calculations, and experimental validation, an installation clearance of 4mm was established as suitable. Considering the significant chassis tracking errors, a road surface key point positioning and steering control method combining QR codes, gyroscopes, and photoelectric sensors was proposed. The design of the segmented and variable distance spraying device involved proposing a screw slide table-driven spray boom with variable-distance capabilities. The driving parameters were analyzed and validated to meet the operational requirements. Additionally, an auxiliary anti-vibration device based on roller bearings was developed to reduce damage caused by severe vibration of the spray rod. The chassis motion and segmented and variable distance spraying control system were developed to enable full automation of the spraying robot within multi-span greenhouses. Finally, performance and spray effect tests were conducted on the prototype, yielding the following results: the average straight-line travel error and tracking error of the chassis were 4.8mm and 5.8mm, respectively, meeting the requirements for control precision. The obstacle avoidance distance was 34cm, ensuring safety. The installation of the anti-vibration device reduced the vibration in the travel direction of the spray rod from -1° to 1.3° to within ±0.4° and limited the vibration in the nozzle direction from ±0.5° to within ±0.3°, demonstrating significant improvement in anti-vibration effectiveness. Following segmented and variable distance spraying, the deposition of mist droplets on the front surface of tomato leaves during the fruiting stage was approximately 1.76μL/cm2, while the back surface achieved approximately 0.2μL/cm2 of deposition. The mist droplet volume median diameter ranged from 100μm to 180μm, meeting the operational requirements.

Abstract:Frost damage often occurs in early spring night, causing huge economic losses to the famous tea industry. Sprinkler irrigation is an effective method of frost protection, which works by maintaining leaf and bud temperatures above the critical temperature through spraying water onto the tea plant and releasing latent heat as it freezes. This method has lower economic input and labor intensity compared with other methods, and sprinkler systems can also be used for irrigation, and fertilization. However, research is still lacking on the storage of water and ice on the surface of tea trees during irrigation to prevent freezing. A tea tree was planted in pot and placed on a high-precision electronic scale to enable weighing of water mass on tea plant surface during sprinkler irrigation. An umbrella cloth was used to keep out sprinkler water from falling into the pot or onto the electronic scale. The patterns of water and ice storage on the tea tree surface under non-freezing (4.0~8.0℃) and freezing conditions (-5.0~0℃) were explored and the effects of sprinkler irrigation duration and nozzle type were analyzed. The results showed that the water storage in tea plants under non-freezing conditions went through three stages of accumulation, dynamic equilibrium and drying respectively;while ice storage under freezing conditions went through four stages of accumulation, retention, melting loss and drying. The tea tree surface was constantly renewed through icing, allowing the sprinkler water to be retained. Under the same irrigation pressure and duration, the water storage and ice storage capacity of micro-sprinkler were 1.2 times and 2.0 times of that of impact sprinkler, respectively. The maximum water storage capacity was about 0.22kg, but the maximum ice storage was increased significantly with the increase of irrigation duration. The ice storage of sprinkling 1.5h was 2.9 times of that of 0.5h. The stem water volume of the micro-sprinkler was higher under non-freezing and freezing conditions, which was 3.0 times and 2.7 times that of the impact sprinkler, respectively. The water and ice storage on the tea tree surface under different conditions was investigated, a method for measurement of shrub canopy interception was proposed, which laid the foundation for further determining the water required for sprinkler irrigation to prevent frost and optimizing the sprinkler irrigation method.

Abstract:Droplets will cause the blade to bend and deform when impacting the blade, and this deformation will affect the droplet impact behavior. Based on the previous theoretical model of droplet impact on rigid surfaces, the influence of blade elasticity coefficient on droplet impact behavior when droplet impacted on flexible chili blades was investigated, and the mathematical prediction model of the maximum diffusion factor of droplet impact on flexible chili leaves was established by taking into account the elastic potential energy of blades, and the gravitational potential energy of droplets and blades. The reasonableness of the mathematical model was verified with droplet particle size, impact velocity, and the percentage of the distance from the impact point to the leaf tip as experimental variables. The results showed that all three factors had a significant effect on droplet adhesion and splashing (P<0.05). The prediction error of the mathematical model for the maximum diffusion factor of droplet impact on flexible chili leaves was within 10%. When the droplet impacted the flexible chili blade without rebound consistent with theoretical model predictions and the droplet adheres, the size of the blade’s elastic coefficient was negatively correlated with the time for the droplet to reach the maximum diffusion, and compared with the impact on the rigid fixed blade, the increase in time to reach the maximum diffusion was within 0.5ms, and the reduction rate of the maximum diffusion factor was within 25%, and the ratio of elastic potential energy to the total initial energy transferred to the blade showed a trend of decreasing and then increasing with the increase in the blade elasticity coefficient, which showed a tendency to decrease first and then increase. When the droplet splashed at 20%, 40%, 60%, 80% of the distance to the leaf tip, the splash critical value Kcrit was increased by 16.202%, 10.515%, 6.508%, 4.467%, respectively, compared with the splash value of the rigid blade,which indicated that the closer the leaf was to the region where the blade elasticity coefficient was small, the droplet would be less likely to splash. The research result can provide a method for the study of droplet impact behavior on flexible plant leaves and spray parameter selection.

Abstract:Sediment erosion of centrifugal pump impeller is a major issue of pumping stations on the Yellow River. The sediment erosion at the outlet of the impeller blade suction side of a double-suction centrifugal pump was mainly analyzed by experiments and numerical simulation. In the experiment, the blade sediment erosion was visualized by using the multi-layer coating method and the near-wall flow pattern were observed by the tuft visualization method coupled with the endoscopic imaging technique respectively. The experimental results showed that there were two triangular severe erosion regions symmetrically distributed at the outlet of the suction side of the blade. In the same regions, the flow separation can also be observed clearly. Furthermore, the Euler-Lagrangian method was used to analyze the mechanism for the sediment erosion on the two triangular severe erosion regions. The inter-blade vortex in the pump impeller and the recirculation vortex at the blade outlet were considered responsible for the formation of severe erosion. The inlet-blade vortex can guide sediment particles to aggregate and hit the outlet of the blade suction side. In addition, the recirculation vortex near the outlet of the suction side of the blade made the particles hit the blade repeatedly. Their combined effects lead to severe sediment erosion on the outlet of the blade suction side. The research results can supply theoretical reference for anti-erosion design of centrifugal pump.

Abstract:To address the current problems that a single optical satellite is easily affected by clouds and SAR satellite is easily affected by vegetation and soil roughness when being applied into soil moisture content inversion, taking Shahaoqu of Hetao Irrigation Area as study area, and taking soil moisture content of four depths in April 2019 as study object, PCA and GS were used to fuse Landsat8 and Sentinel-1 images and the quality of the fused images was evaluated. Then a total of 1134 remote sensing indices were constructed with the gray value of the fused images, and soil moisture content inversion models were constructed based on three variable screening methods (correlation coefficient analysis, variable projection importance analysis and gray correlation analysis) and four machine learning algorithms (BP, ELM, RF, and SVM). The study results showed that the fused images of PCA and GS fusion could successfully maintain the advantages of both Sentinel-1 and Landsat8 images in quantitatively inversion of soil moisture content. The three-dimension indices constructed based on the fused images were generally more sensitive to soil moisture content than two-dimension indices constructed based on fused images. The VIP-ELM model based on GS fusion had the highest accuracy in the surface soil moisture content inversion (R2=0.66, RMSE was 1.35%). When VIP-ELM model based on GS fusion was applied to the soil moisture content inversion at all depths, 20~40cm achieved the best performance (R2=0.79, RMSE was 0.94%), followed by 0~10cm, 40~60cm and 10~20cm. This finding can provide a strong reference for using multi-source satellite image fusion to monitor soil moisture content.

Abstract:Leaf water content and leaf water potential reflect the state of water in plant tissues and are important indicators of plant water availability and water use efficiency. To investigate the differences in leaf water content and leaf water potential modelling based on UAV multispectral image inversion at different altitudes, multispectral image data were collected at three flight altitude treatments F30, F60, and F100 (30m, 60m, and 100m). By using six combinations of spectral reflectance + empirical vegetation index (EVI) and ground data for correlation analysis, the inversion models and their decision coefficients of the combinations of spectral reflectance + EVI with leaf water content and leaf water potential at different flight altitudes were obtained. Support vector machine (SVM), random forest (RF) and radial basis neural network (RBFNN) models were constructed based on the determination coefficients to analyze the accuracy of UAV multispectral inversion models for leaf water content and leaf water potential of aromatic camphor at different flight altitudes. It was found that the inversion accuracy of the RF-based model was higher than that of the SVM model and the RBFNN model at all three flight altitudes. The F30 treatment was better than the F60 and F100 treatments for leaf water content and leaf water potential inversion. The sensitive spectral reflectance+vegetation index combinations for leaf water content inversion in the F30 treatment were reflectance in the red band (R), reflectance in the red-edge 1 band (RE1), reflectance in the red-edge 2 band (RE2), near-infrared reflectance (NIR), and enhanced vegetation index (EVI), soil adjusted vegetation index (SAVI). The R2, RMSE, and MRE for the training set of the RF model were 0.845, 0.548% and 0.712%, respectively;and for the test set, the R2, RMSE, and MRE were 0.832, 0.683% and 0.897%, respectively. The sensitive spectral reflectance + vegetation index combinations for leaf water potential inversion were R, RE2, NIR, EVI, SAVI, anthocyanin reflectance index (ARI). The R2, RMSE, and MRE for the training set of the RF model were 0.814, 0.073MPa and 3.550%, respectively;and for the test set, R2, RMSE, and MRE were 0.806, 0.095MPa and 4.250%. The results showed that the 30m flight altitude and RF method were the optimal spectral acquisition altitude and optimal model construction method for inverting leaf water content and leaf water potential, respectively. The research result can provide technical support for the moisture monitoring of Cinnamomum camphora based on UAV platform, and can provide application reference for screening UAV multispectral bands and empirical vegetation indices, and realising rapid estimation of plant growth parameters.

Abstract:In order to improve the accuracy of winter wheat identification, the difference between radar and optical remote sensing data on winter wheat area extraction was compared and analyzed based on Google Earth Engine (GEE) platform and random forest algorithm. The importance analysis of multiple feature variables was performed to study the influence of feature optimization on the accuracy of winter wheat extraction. The Sentinel-1 and Sentinel-2 images during the main growth period of winter wheat (from March 1 to May 31, 2019) were chosen as the data sources. The polarization and texture features of Sentinel-1 data as well as the spectral, vegetation index and vegetation index change rate features of Sentinel-2 data were constructed. Six winter wheat identification schemes were constructed based on different remote sensing data sources and feature combinations, and the accuracies of the schemes were compared and analyzed. Then the feature variables were optimized and the optimal feature combination was obtained to extract the planting area of winter wheat in Zhumadian City, Henan Province. The results showed that regardless of feature optimization, the results of winter wheat area extraction based on multi-source remote sensing data were superior to those by using only optical or radar data. After feature optimization, the classification accuracy of each scheme was further improved, indicating that both the combination of multi-source feature variables and feature optimization can improve the winter wheat identification accuracy. In addition, the feature variables of different months and types had different contribution rates to classification accuracy, and the months with contribution rates from high to low were April, March and May. The feature types with contribution rates from high to low were polarization, vegetation index change rate, vegetation index, spectral features and texture. The accuracy of winter wheat extraction in Zhumadian based on both multi-source satellite data and feature optimization were the best, with the overall accuracy of 95.60% and Kappa coefficient of 0.93. The relative error between the extracted area of winter wheat and official statistical data was 2.23%. The research result can provide an important theoretical reference for crop planting area extraction based on multi-source optical and radar remote sensing images.

Abstract:Image edge detection is a technique that extracts mutation information from images and is widely used in the fields of image processing and computer vision. The effectiveness of image edge detection directly affects the accuracy of subsequent region information extraction, target recognition, and pose measurement. Taking into account two factors: local extremum and gradient direction, and combining with the trend of image edge direction, a single-pixel edge tracking strategy was proposed for the edge detection problem of low contrast and edge blurred images. Compared with the widely used Canny algorithm, this tracking strategy did not require setting a global threshold, and its implementation was more concise and efficient. The extracted image edges were continuous, smooth, and complete, effectively reducing redundant pixels at the image edges, thereby improving the efficiency of subsequent image processing. Edge tracking direction had strong anti-interference ability and robustness. In order to reduce the deviation between the detected image edge and the real image edge, and improve the accuracy of image edge detection, the adjacent gray values of edge pixels were referred to, and the gradient distribution of edge pixels was used as the basis for sub-pixel localization of that pixel. Through experimental verification, the sub-pixel optimized image edge detection strategy can be used to detect images with blurred edges and low contrast. The detected image edges were complete, continuous, and smooth. This strategy effectively eliminated truncation errors introduced in program operations, improved the accuracy of image edge detection, which was suitable for high dynamic imaging scenes with a brightness range of 5~100000lx.

Abstract:In order to solve the recognition and detection of branches at the young leaves of mulberry trees in complex natural environments, overcome the current situation of relying on manual assisted positioning in the operation process of mulberry leaf harvesting equipment, and improve the problem of low recognition rate caused by diverse target postures and complex environments, a mulberry branch and trunk recognition model was proposed based on the improved YOLO v5 model (YOLO v5-mulberry) and combined it with the depth camera to construct a location system. Firstly, convolutional block attention module (CBAM) attention mechanism was added to the backbone network of YOLO v5 to improve the neural network’s attention to the mulberry branches;and a small target layer was added to enable the model to detect 4pixels×4pixels targets, which improved the model’s performance in detecting small targets. At the same time, the GIoU loss function was used to replace the IoU loss function in the original network, which effectively prevented the position relationship between the prediction box and the real box from being correctly reflected when the size of the prediction box and the real box was small. Subsequently, the pixel alignment of the depth map and the color map was completed, and the 3D coordinates of the mulberry tree trunk were obtained through the conversion of the coordinate system. The test results showed that the average accuracy of YOLO v5-mulberry detection model was 94.2%, which was 16.9 percentage points higher than that of the original model, and the confidence level was also 12.1% higher;the number of targets that should be detected by the model outdoor detection was 53, and the number of actually detected targets was 48, and the detection efficiency was 90.57%;the positioning error of the three-dimensional coordinate recognition and location system of the mulberry branch and trunk at the tender leaves was (9.4985mm,11.285mm,19.11mm), which met the requirements for use. The research result can achieve the recognition and positioning of branches and trunks at the tender leaves of mulberry trees, which can help to further promote the research, development and application of intelligent mulberry leaf picking robots.

Abstract:Chrysanthemums have a wide variety of flower types with subtle differences in flower phenotypes, which are difficult to label accurately, and this poses a great challenge for intelligent classification and recognition of chrysanthemums. Based on deep active learning and hybrid attention mechanism module, i.e. convolutional block attention module (CBAM), a method and framework for intelligent recognition of chrysanthemum phenotypes under insufficient labeling data was proposed. Firstly, the more informative samples among the unlabeled chrysanthemum samples were selected for labeling by an active learning strategy based on the optimal labeling and second-optimal labeling method BvSB (Best vs second-best), and the labeled samples were put into the training samples;secondly, a deep convolutional neural network ResNet50 was used as the backbone network to train the labeled samples, and the hybrid attention mechanism module CBAM was introducted, so that the model can more accurately extract the high-level semantic information in fine-grained images;finally, the classification model continued to be trained with the updated training samples until the model reached the number of iterations and then stopped. The experimental results showed that the method can achieve 93.66%, 93.15% and 93.41% of precision, recall and F1 value respectively with a small number of chrysanthemum labeled samples. The method can provide technical support for intelligent identification of chrysanthemums and other flowers under the situation of insufficient labeling data.

Abstract:To address the problems of low efficiency of traditional manual identification and inconsistent identification standards, a ripening identification method for Hemerocallis citrina baroni based on lightweight and efficient layer aggregation network LSEB YOLO v7 was proposed. Firstly, lightweight convolution was introduced to lighten the efficient layer aggregation network and transition module to reduce the model computation. Secondly, the channel attention mechanism was added between the feature extraction and feature fusion networks to improve the model detection performance. Finally, in the feature fusion network, the channel information fusion method was optimized, and the bi-directional feature pyramid network was used to replace concatenate to increase the information fusion channels and continuously improve the model performance. The experimental results showed that compared with the original algorithm, in the Hemerocallis citrina baroni maturity detection, the number of parameters and floating-point operations of the improved LSEB YOLO v7 algorithm were reduced by about 2.0×106 and 7.7×109, respectively, and the training time was reduced from 8.025h to 7.746h, and the model volume was compressed by about 4MB. Meanwhile, the training precision and recall were improved by about 0.64 percentage and 0.14 percentage, respectively. The mAP@0.5 and mAP@0.5：0.95 were improved by about 1.84 percentages and 1.02 percentages, respectively. In addition, the harmonized mean remained unchanged at 84.00%. It was evident that the proposed LSEB YOLO v7 algorithm solved the problem of the paradox between model complexity and performance, and provided technical support for intelligent ripening and harvesting inspection equipment for Hemerocallis citrina baroni.

Abstract:Apple tree thinning is an important step in orchard production management. Accurate and efficient recognition of apple king flowers and side flowers is the premise of the development of intelligent flower thinning robot. According to the actual demand of apple flower thinning, a method for recognizing king flowers and side flowers of apple based on CRV-YOLO was proposed. Based on YOLO v5s model, the following improvements were made: firstly, C-CoTCSP structure was integrated into Backbone to better learn contextual information and improve the detection performance of the model for king flowers and side flowers that were similar and the position relationship was not obvious. Then an improved RFB structure was added to the Backbone, with which the receptive field of feature extraction was expanded and the branch contribution degree was weighted to make better use of different scale features. Finally, VariFocal Loss loss function was used to improve the detection ability of the model for samples in occlusion and other scenes. Experiments were conducted on a dataset of 1837 images from three varieties. The results showed that the precision, recall and mAP of the proposed model were 95.6%, 92.9% and 96.9%, respectively, which were 3.7 percentage points, 4.3 percentage points and 3.9 percentage points higher than those of the original model. The model was less affected by light changes and apple varieties. Compared with that of Faster R-CNN, SSD, YOLOX, and YOLO v7, precision, the mAP and model size and complexity performance of CRV-YOLO were optimal, and recall was close to optimal. The research results can provide technical support for apple intelligent flower thinning.

Abstract:The diagnosis, prevention and control of livestock and poultry diseases is of great significance to ensure the healthy development of animal husbandry in China. Based on natural language processing, the word segmentation effect of livestock and poultry disease texts was improved to improve the diagnosis level of livestock and poultry diseases. In order to deal with the problems of lacking text corpus in livestock and poultry diseases, and a large number of out of vocabulary words contained in the texts, such as epidemic names and phrases, a word segmentation model was proposed based on BERT-BiLSTM-CRF combined with dictionary matching. Taking sheep diseases as the research object, the text datasets of common diseases were constructed combined with the general corpus PKU, and the text vectorizations were processed by BERT pre-trained language model. Then the context semantic features were obtained through the bidirectional long short-term memory network (BiLSTM), and globally optimal label sequences were outputted by conditional random field (CRF). Based on this, dictionary matching was refined by adding a dictionary in the field of livestock and poultry diseases after the CRF layer, which reduced the ambiguity segmentation caused by the epidemic names and phrases in the process of word segmentation, and the accuracy of word segmentation was further improved. Results showed that the F1 value of the BERT-BiLSTM-CRF model combined with dictionary matching on the text datasets of sheep common diseases was 96.38%, which was increased by 11.01, 10.62, 8.3 and 0.72 percentage points, compared with that of jieba word segmentation, BiLSTM-Softmax model, BiLSTM-CRF model, and BERT-BiLSTM-CRF model that did not combine with dictionary matching, respectively, which verified the effectiveness of BERT-BiLSTM-CRF. Compared with a single corpus, the mixed corpus combined with the general corpus PKU and the text datasets of sheep common diseases could accurately divide the professional terms of livestock and poultry diseases and common words in the texts of diseases at the same time, the F1 values of the general corpus and the text datasets of diseases were more than 95%, which illustrated its better generalization ability. BERT-BiLSTM-CRF model can be effectively used for word segmentation of texts on livestock and poultry diseases.

Abstract:Multiple-object tracking of mice is a fundamental task in behavioral analysis and an important method for studying social behavior. In response to the limitations of traditional mouse tracking methods, such as the ability to track only a single mouse and the need for mouse labeling to track multiple-object mice, which affects mouse behavior, an unlabeled multiple-object mice tracking method was proposed based on the combination of instance segmentation network YOLO v8n-seg and improved Strongsort. RGB cameras were used to capture daily behavior videos of multiple-object mice, and a dataset for segmenting mouse body parts was annotated. After augmenting the dataset, the YOLO v8n-seg instance segmentation network was trained. The model achieved a precision of 97.7%, recall of 98.2%, mAP50 of 99.2%, and single-image detection time of 3.5ms. It accurately and quickly segmented mouse body parts, meeting the detection requirements of the Strongsort multi-object tracking algorithm. To address tracking errors in the Strongsort algorithm for multiple-object mice tracking, two improvements were made. Firstly, the matching process was improved by re-matching trajectories that did not match objects and unmatched objects based on Euclidean distance. Secondly, the Kalman filter was improved by replacing the rectangular bounding box representing the mouse position and motion state in the Kalman filter with a square box centered on the centroid of the mouse body contour and with a diagonal equal to the mouse body width. After testing, the improved Strongsort algorithm showed an ID switches of 14, MOTA of 97.698%, IDF1 of 85.435%, and MOTP of 75.858%. Compared with the original Strongsort, the ID switches count was reduced by 88%, MOTA was improved by 3.266 percentage points, and IDF1 was improved by 27.778 percentage points. Compared with Deepsort, ByteTrack, and Ocsort, there was a significant improvement in MOTA and IDF1, and the ID switches was greatly reduced. These results indicated that the improved Strongsort algorithm can enhance the stability and accuracy of unlabeled multiple-object mouse tracking, providing a technical approach for analyzing social behavior in mice.

Abstract:Aiming at the shortcomings of existing water quality prediction models in data noise reduction, initial setting and optimization of network parameters, and accuracy improvement, an optimized three-dimensional water quality prediction model was constructed. The key parameters of water quality were screened by using the principal component analysis algorithm, the three-dimensional water quality parameters and meteorological data were de-noised based on the fully set empirical mode decomposition algorithm based on adaptive noise combined with the wavelet threshold model, the feature data set was extracted by using the three-dimensional convolutional neural network (3D CNN) and the dynamic initial values of hyperparameters in radial basis function (RBF) neural networks were optimized by an improved cuckoo search algorithm (ICS) based on autoencoder (AE). The comparison and verification results of the measured data from twenty-two typical online monitoring stations and six handheld monitoring stations in the Dashuiqiao Reservoir area of Xuwen County, Zhanjiang City, Guangdong Province showed that turbidity and algae density were positively correlated with total nitrogen, and chlorophyll was positively correlated with temperature, the proposed water quality prediction model was superior to the existing literature methods in five typical accuracy evaluation indicators. The research results can provide reference for management departments and researchers to monitor water quality. Introducing inertial weight and adjusting position parameters to improve CS to speed up the convergence of RBF network. The autoencoder was used to initialize the initial values of network parameters to avoid the defects of artificial random setting. Adding WT algorithm can effectively reduce the white noise in the decomposition and reconstruction process of the fully set empirical mode decomposition algorithm based on adaptive noise.

Abstract:Global warming has direct impacts on agricultural production and food security. The North China Plain (NCP) has adopted the double later-cropping system (the later sowing of winter wheat and the later harvest of summer maize) as a strategy to cope with climate change and increase yield. Determining the latest sowing period for winter wheat in this region under the backdrop of climate change is crucial for ensuring high yield and efficiency, thereby safeguarding China’s food security. Based on the WMAIP integrated model, the suitable sowing period and the minimum pre-winter accumulated temperature were analyzed for winter wheat at six representative sites in NCP, including Beijing and Tianjin in the north, Xingtai and Ji’nan in the middle, and Xihua and Nanyang in the south. The results showed that the yield reduction rate due to late sowing was increased from south to north in NCP, with the southernmost part of NCP having the smallest reduction rate (<19.5%), the central part having a reduction rate within 26.4%, and the northernmost part having the largest reduction rate, reaching up to 32.0%. The suitable sowing periods for the northern, central, and southern parts of NCP were between September 25 and October 5, September 30 and October 20, and October 10 and November 5, respectively. All sites showed a very significant quadratic relationship between the high stability coefficient and the pre-winter accumulated temperature. Based on this quadratic relationship, the lower limit of suitable pre-winter accumulated temperature can be determined. The suitable pre-winter accumulated temperature for winter wheat varied at different sites, showing a gradually decreasing trend from north to south in NCP. The suitable pre-winter accumulated temperature limits for winter wheat in NCP, based on high yield and high water use efficiency, were between 497℃·d and 629℃·d and 344℃d and 581℃·d, increasing from south to north. Under conditions of high yield, the latest sowing dates for the majority of regions in Hebei, Shandong, and Henan were between October 1 and October 13, October 10 and October 16, October 22 and October 28, respectively, and under high water use efficiency conditions, they were extended to October 7 and October 19, October 16 and October 22, and October 31 and November 12. The insights can provide a reference for the promotion and application of the double later-cropping system in NCP under the context of global warming.

Abstract:To investigate the changes in soil organic carbon (SOC), bacterial communities and metabolic functions, as well as the dynamic response relationships with each other in facility agricultural land with different cropping periods, organic carbon, bacterial community structure and metabolic functions in soil were systematically characterized in unplanted (CK), 0 to 5a (0~5a), 5a to 10a (5~10a), 10a to 20a (10~20a) and more than 20a (20a+) of continuous cropping by using temporal-spatial substitution method in a centralized area of facility agriculture in the southern suburb of Beijing as the study area. The dynamic response of SOC and its active components to the bacterial dominant communities and their interrelationships with bacterial metabolic functions were explored with redundancy analysis (RDA), typical correlation analysis (CCA), prediction of PICRUSt2 function, and Mantel test. Based on the results, continuous cropping caused the contents of microbial biomass carbon (MBC), easily oxidizes organic carbon (EOC), SOC and soil organic carbon density (SOCD) all showed an increasing and then decreasing trend with the increase of cropping period. Dissolved organic carbon (DOC) content was the highest at 20a+. Microbial entropy (Q) was decreased with the increase of cropping period. Continuous cropping decreased the diversity of soil bacteria and increased the differences among bacterial species. PICRUSt2 prediction showed that the functions of soil bacteria in different continuous cropping period were dominated by metabolism, and the abundance of secondary metabolism functions of bacteria was significantly higher in the soil from 5~20a than that in the soil from 0~5a, 20a+, and CK. Among the bacterial groups in the top 10 relative abundance, the Acidobacteriota were negatively correlated with both SOC and its active components. The key bacterial groups dominating and participating in SOC accumulation and cycling showed a shift from eutrophic to oligotrophic and pathogenic taxa after 20a+ of continuous cropping. Mantel analysis showed that a total of 53 classes of functions in the third tier were significantly correlated (p<0.05) with SOC and its components. Among them, totally 23 classes were affiliated to metabolic functions in the first tier, and as many as 51 classes were significantly correlated with EOC. The finding can provide scientific references for optimizing carbon cycle-related bacterial functional groups and enhancing the carbon sink function of small-scale facility agricultural soils.

Abstract:Ammonia recovery from aerobic composting of livestock and poultry manure, or recovery from biogas slurry, can not only reduce pollutant and greenhouse gas emissions, but also obtain nitrogen fertilizer products. Aiming to use hollow fiber membrane absorption process for ammonia gas capture to solve the problems of large equipment volume and poor flexibility in the existing ammonia capture process, air stripping of diluted aqueous ammonia solution was used to prepare ammonia contained air. Then, the simulated ammonia contained air was flowed into hollow fiber membrane contactor, in which ammonia was captured by acid solution flowing in the tube side of the hollow fiber. By analyzing the total ammonium nitrogen concentration in the aqueous ammonia solution and acid solution, the ammonia mass transfer characteristics can be acquired. The results indicated that the mass transfer resistance of ammonia capture was mainly influenced by the gas phase mass transfer resistance and the membrane mass transfer resistance. The gas phase mass transfer resistance dominated at low air flow rates. However, when the air flow rate was increased to 5L/min, the gas phase mass transfer resistance was decreased by 53.6% compared with that of 0.5L/min, and the membrane mass transfer resistance dominants. The ammonia capture flux was increased with the elevation of ammonia concentration. Ammonia recovery ratio was above 95% at an air flow rate below 1L/min, while it was increased to above 99% at air flow rate of 0.5L/min. The ammonia recovery ratio was only related to the acid absorption capacity if sufficient ammonia retention time was provided. Under the dual influence of temperature and concentration differences, water vapor in the air would transfer into the membrane thus decreasing the product concentration. When the mass fraction of absorbent was 26%, the water transfer flux was 13.3 times higher than that of using 1% sulfuric acid solution. Besides, ammonia separation factor was decreased from 41.6 to 3.06. Acid concentration variation had no significant impact on the mass transfer of ammonia. Literature review showed that the ammonia release concentration in typical manure treatment was basically consistent with the concentration range. The hollow fiber membrane proposed had a wide applicability for capturing ammonia.

Abstract:In order to investigate the effects of ultrafine-grinding/NaOH synchronous treatment on the separation of wheat straw cellulose, wheat straw was mixed with 6% NaOH solution at a ratio of 0.1g/mL and subjected to ultrafine-grinding/NaOH synchronous treatment with different time. Then, 1.4% acidified sodium chlorite solution and water bath ultrasound was used for cellulose fibers separation. The effects of different treatment times on wheat straw lignocellulosic components and microstructure, crystal structure and thermal stability of separated cellulose were systematically characterized. The results showed that within 0~60min treatment, mechanical force significantly reduced the particle size of wheat straw samples, which effectively promoted the separation of lignocellulosic components. In the separated wheat straw cellulose obtained by ultrafine-grinding/NaOH synchronous treatment, a large amount of micron and nanoscale cellulose fibers was intertwined together. As the treatment time prolonged, the crystallinity of cellulose was firstly decreased and then tended to stabilize after 30~60min. The results of Person correlation analysis indicated that the thermal stability of cellulose was significantly correlated with its crystallinity and treatment time (P<0.05). The research provided important experimental data for the innovative research and development of straw based cellulose materials prepared by mechanochemical synchronous treatment.

Abstract:The oxidation-reduction potential (ORP) reflects the oxidation reduction state of the fermentation system, which is closely related to yeast metabolism and aroma compound synthesis, and the ORP level varies during each phase of alcohol fermentation. In order to realize the effective control of ORP in different fermentation phases, different levels of ORP were controlled during different phases of wine alcohol fermentation, and the effects on yeast growth, aroma compound aroma compound synthesis and sensory quality were observed by gas chromatography-mass spectrometry (GC-MS) and sensory quantitative analysis. Results showed that during 0～48h, the number of viable yeast and the content of aroma compounds in yeast were greatly affected by the change of ORP than other fermentation phases. Increasing ORP at this phase was beneficial to yeast growth and sugar consumption, while reducing ORP was beneficial to the synthesis of aroma compounds. Compared with the control, at the level of 0mV (0～48h)-natural conditions (48～96h)-natural conditions (96h to the end of fermentation), the contents of ethyl hexanoate and ethyl decanoate can be significantly increased. The content of acetates can be significantly increased at the level of 0mV-0mV-60mV, and the content of higher alcohols can be significantly increased at the level of 60mV-0mV-60mV. The treatment of 0mV-natural conditions-natural conditions enhanced the aroma of red fruits in Cabernet Sauvignon wine and the fragrance of flowers and fruits in Ecolly wine, while weakened the odor of animals in Ecolly wine. The phased control of ORP provided a basis for the precise regulation of wine fermentation, and laid a foundation for improving the aroma quality of wine by increasing the content of aroma compounds in wine.

Abstract:Soybean protein isolate (SPI) exhibits remarkable sensitivity to environmental changes, particularly alterations in pH value during the neutralization process. Even minor fluctuations can have a profound impact on the structure and functional properties of the protein. Revolving around the addition of NaOH to regulate the pH value of the curd, and analysis of SPI’s structure and functional attributes was done by using infrared and endogenous fluorescence spectra. It was revealed that SPI displayed excellent foaming properties under neutral conditions, while its emulsification capabilities excelled under slightly more alkaline conditions. Specifically, pH value of 7 yielded the best foaming type, while pH value of 8.5 led to superior foaming properties alongside optimal emulsification. In order to fine-tune the neutralization of a 25L soybean SPI section, a pH value control system was established, and the entire production process was simulated by using Matlab. Dynamic linear and static nonlinear fitting techniques were employed, with a combination of fuzzy adaptive control and the Wiener model utilized to regulate the amount of alkali added to the neutralization tank. The adjustment time of 37.4s achieved when the pH value in the neutralization tank was set to 7. Likewise, at pH value of 8.5, the adjustment time remained efficient at 33.4s, while the SPI emulsion activity index reached 69.35m2/g. Notably, the system successfully avoided any overshooting of alkali during the process, ensuring precise control and stability throughout. To summarize, the research emphasized the significance of pH value control during the neutralization of soybean protein isolate. The findings can provide important insights into optimizing SPI’s structure and functional properties, which can have significant implications for its applications in the food and industrial sectors.

Abstract:Phenolic compounds play a crucial role in assessing the internal quality of grapes and hold significant importance in this regard. The capability of visible-near-infrared (Vis-NIR) spectroscopy combined with multivariate regression models was explored to detect the contents of total phenolics and tannins in grape skins and seeds. Reflectance spectra data of Muscat Kyoho grapes were collected within the wavelength range of 400nm to 1029nm, and the samples were divided into correction set and prediction set by SPXY algorithm. Six commonly used preprocessing methods such as standard normal variate (SNV), multiplicative scatter correction (MSC), first derivative (1D), second derivative (2D), Savitzky-Golay smoothing (SG) and SG+1D were applied to the spectral data, and the competitive adaptive reweighted sampling algorithm (CARS) was utilized to select informative wavelengths. The quantitative models for comprehensive analysis of total phenolics and tannins in grape skins and seeds based on full spectra and effective wavelengths were established by partial least squares regression (PLSR), support vector machine regression (SVR), and convolutional neural network (CNN). The results showed that for the total phenolics in grape skins, total phenolics and tannins in grape seeds, the models on the basis of effective wavelengths performed better than those with full spectra data. While for the tannins in grape skins, the models constructed with full spectra yielded better results than the feature wavelength-selected models. The optimal models for the total phenolics and tannins in grape skins and seeds were RAW-CARS-SVR, 1D-CARS-SVR, RAW-CNN and RAW-CARS-PLSR, respectively. The correlation coefficent of calibration set (Rc) were 0.96, 0.99, 0.96 and 0.91, the correlation coefficent of prediction set (Rp) were 0.95, 0.99, 0.83 and 0.89, the residual predictive deviation (RPD) were 3.56, 7.30, 1.92 and 2.25, respectively. Therefore, the developed method could realize the non-destructive detection of the contents of total phenolics and tannins in grape skins and seeds.

Abstract:With the aim to address the issues of low efficiency and high labor costs in crack detection and sorting of preserved eggs, a method for online crack detection based on an improved version of YOLO v5 was proposed. The backbone feature extraction network was replaced with the EfficientViT network, and the network was trained by using transfer learning, resulting in two models: YOLO v5n_EfficientViTb0 and YOLO v5s_EfficientViTb1. YOLO v5n_EfficientViTb0 served as a lightweight model, reducing the parameter size by 148% and the floating point operations by 268% compared with that of the original model. YOLO v5s_EfficientViTb1, on the other hand, was a high-precision detection model with an average precision mean of 878%. Through the utilization of GradCAM++ for model visualization and analysis, it was discovered that the improved model demonstrated a decreased focus on the background region. This finding served as evidence supporting the effectiveness of the enhancements implemented in the model. Moreover, a target box matching algorithm was designed for video frames to enable object tracking of preserved eggs in videos. Based on the detection sequence of preserved eggs, the algorithm achieved localization of the eggs and discrimination between cracked and intact ones. The lightweight model achieved a discrimination accuracy of 92.0%, while the high-precision model achieved an accuracy of 94.3%. These research findings indicated that the improved lightweight model provided a solution for preserved egg crack detection equipment with lower computational capabilities, while the improved high-precision model offered technical support for preserved egg crack detection equipment with higher production requirements.

Abstract:The load impact of high-horsepower tractors during operation will cause a wide range of fluctuations in the output torque of the engine. In order to reduce the impact of load impact on the tractor power unit, a tractor coupled-split power system configuration with the engine and dual-motor as the power source was proposed to reduce the shift frequency of the power transmission system caused by load impact. A torque allocation strategy was proposed based on Haar wavelet decomposition algorithm and power prediction. Firstly, a priori prediction of power demand for tractor rotary tillage based on radial basis function neural network was researched in which the working load parameters were collected;and then the comprehensive dynamics of tractor loads were mathematically formulated. Then the torque requirements for high and low frequencies were determined by Haar wavelet transform and provided by motor and engine respectively. Finally, the effectiveness and feasibility of the proposed strategy were validated with the hardware-in-loop test. The result indicated that the prediction model of power requirement based on neural network can accurately predict the power demand of driving and power take-off (PTO), and driving end and PTO end root mean square error of predicted values accounted for 7.6% and 7.9% of the maximum power, respectively. The proposed model predictive controller can following tractor torque demand in operation. The torque ripple of the engine was reduced by 35.0% compared with the traditional configuration. And the strategy effectively reduced the torque variation range of the engine and alleviated the adverse effects of excessive shock of the operating load.

Abstract:A method utilizing the adaptive strong tracking unscented Kalman filter (ASTUKF) was proposed to address the issue of divergent identification results caused by state model errors and time-varying noise resulting from changes in road environments during the terrain parameters identification of distributed drive agricultural vehicles. Compared with the traditional internal combustion engine agricultural vehicles, distributed drive agricultural vehicles can directly obtain state information of the driving wheel. And combining the μ-s model which contained adhesion coefficient and limit slip ratio, a state function and a measurement function of unscented Kalman filter (UKF) identification algorithm were established. At the same time, strong tracking filter (STF) and adaptive filter (AF) were introduced into the identification algorithm to improve identification accuracy and robustness against the changing environment, and singular value decomposition (SVD) was used to solve the problem of non-positive definite matrix in iterative process. The simulation test showed that under the condition of abrupt noise environment, the identification result of ASTUKF can quickly converge to target value, which was not affected by abrupt noise. Mean absolute errors (MAE) of the adhesion coefficient estimation results of each driving wheel were 0.0144, 0.0267, 0.0144 and 0.0267, respectively, and MAE of the limit slip ratio estimation results were 0.0025, 0.0028, 0.0025 and 0.0028, respectively. The real vehicle test showed that the 95% confidence interval of average identification result of ASTUKF can match the measured value on test road of cultivated and uncultivated road. The identification results of adhesion coefficient of the whole vehicle were 0.4061 (uncultivated road) and 0.3991 (cultivated road), and the identification results of limit slip ratio were 0.1484 (uncultivated road) and 0.3600 (cultivated road), which can provide a theoretical reference for the operation parameter perception of distributed electric agricultural vehicles.

Abstract:In order to address the challenges posed by low linear tracking accuracy, slow on-line speed and poor anti-disturbance rejection ability in unmanned tractor-trailer agricultural vehicles operating in unstructured environments, a straight-line path tracking control method based on fuzzy fast power sliding mode control was designed. Firstly, a straight-line path tracking error model was developed for agricultural tractor-trailer vehicles, grounded in the vehicle’s kinematic model and the reference path. A straight-line path tracking method based on fast power sliding mode theory and fuzzy parameter adjustments was proposed, which solved the sliding mode controller chattering problem and also improved the on-line speed. Rigorous Lyapunov stability analysis was presented to affirm the stability of the closed-loop system. The designed control method ensured that the tractor-trailer effectively tracked the reference path, while the articulation angle converged to zero at the same time.The effectiveness and superiority of the proposed control method were evaluated through Simulink simulations and real vehicle tests. Field experiments revealed that when employing this control method, the maximum path tracking lateral errors for the tractor and trailer were 0.11m and 0.12m, respectively, with tracking error variance values of 0.0013m2 and 0.0015m2. Compared with the traditional sliding mode controller designed based on constant reaching law, the online time was improved approximately by 58%, and the maximum tracking error was reduced roughly by 66%. The designed method can effectively improve the stability and speed of the tractor-trailer path tracking.

Abstract:To solve the problems of ant colony algorithm in complex grid environment, such as local optimization, many turning points and slow convergence, dynamic extended neighbourhoods ant colony optimization (DENACO) algorithm was proposed. Firstly, the method of dynamic extended neighborhoods was applied in the ant search mode to obtain the optimal convergence path length and reduce the number of inflection points and the number of path nodes. Meanwhile, a computational method and increment rule of pheromone were defined to reduce space costs, and the upper and lower limits of pheromone were set to avoid premature convergence of the algorithm to local optimality. Secondly, the adaptive adjustment factor and target point factor were introduced into the heuristic function, and a weight coefficient was set to improve the global search ability of the algorithm. Moreover, an iteration threshold of the algorithm was set. When the iteration exceeded the threshold, the pheromone concentration factor and heuristic factor values were updated to improve the convergence speed of the algorithm. Finally, a double optimal strategy of nodes of path was proposed. Two optimization methods were used to further optimize the planned path, and the best was taken as the final optimization result to improve the comprehensive quality of the path. Simulation experiments on raster maps of different complexities and scales showed that compared with the traditional ant colony algorithm and other comparison algorithms, the path planned by DENACO algorithm was superior. It had a shorter path length, reduced number of inflection points, accelerated convergence speed, and significantly fewer path nodes. These results indicated that the DENACO algorithm was highly feasible and applicable.

Abstract:Agricultural environments have many ups and downs, fuzzy boundaries, and mostly unstructured distributions. Quadrupedal robots operating in complex agricultural environments are very susceptible to tipping over and thus losing locomotion, and therefore, quadrupedal robots need to have the ability to recover after tipping over. Traditional quadrupedal robots rely on leg movement to recover from tipping in most cases, while reconfigurable quadrupedal robots can realize self-recovery after tipping through the coordinated movement of the torso and legs. Based on the variable configuration of the reconfigurable torso, bionic forms of reconfigurable quadruped robots were obtained, and the recovery mechanism after tipping based on the reconfigurable theory was planned. Then comparing the two recovery modes of the reconfigurable robots after tipping by using the torso arching and folding and unilateral flipping and folding, the kinematic characteristics of the recovery of the R1type and R2type reconfigurable quadrupedal robots were analyzed and got. Then simulation was carried out by using the software ADAMS, and the simulation data were analyzed to prove that the reconfigurable torso was more effective than the rigid torso in reducing the impact during the recovery process. Finally, a prototype was designed and experiments were conducted to verify the feasibility and stability of the mechanism implementation, and the experimental results showed that the post-collapse recovery mechanism can reduce the difficulty of realizing static self-recovery.

Abstract:Linear conjugated internal gear pump plays a key role in hydraulic system, and its efficient pressure transmission characteristics make it widely used in engineering field. In this paper, the effects of axial clearance and radial clearance on leakage and flow field of linear conjugate internal gear pump are analyzed by means of computational fluid dynamics simulation. The results show that the variation of mating clearance has a wide and significant influence on the flow field characteristics of gear pump, in which the axial clearance is the main factor causing leakage, accounting for 80% of the total leakage. Specifically, when the axial clearance increases from 0.03mm to 0.07mm, the output flow rate decreases by 20.81%, the average pressure decreases by 33.15%, and the volume fraction of the gas generated by cavitation increases by 0.021. In contrast, the setting of the same radial clearance only resulted in a 0.69% decrease in output flow, a 2.76% decrease in average pressure, and a 0.005 increase in the volume fraction of the gas produced by cavitation. In addition, the study also found that the main matching clearance leading to the change of flow rate in the pump is the axial clearance. A modest reduction in the axial clearance helps to increase the fluid speed in the pump, thereby enhancing the overall efficiency of the pump. These results provide useful theoretical support for the design and optimization of linear conjugate internal gear pump, which is helpful to improve its performance and reliability in hydraulic system.

Abstract:The limitation of measurement noise to the bandwidth of the extended state observer is a key problem that affects the performance of active disturbance rejection position controller of electro-hydraulic servo system. Therefore, an improved active disturbance rejection control method based on noise suppression extended state observer was proposed. The nonlinear model of the electro-hydraulic servo system was established, the chain integrator structure was constructed by coordinate transformation, and the “total disturbance” of the electro-hydraulic servo system was defined. A low-pass filter was introduced to suppress the high-frequency measurement noise, and the improved extended state observer was constructed by using the filtered position signal to compensate for the phase lag caused by the filter, separate the state feedback from the disturbance estimation, add new disturbance estimation adjustment parameters, and reconcile the contradiction between the high bandwidth and high estimation performance of the observer and the noise amplification. The controller was designed for a typical valve-controlled symmetrical cylinder electro-hydraulic servo system and the closed-loop system stability was analyzed by using Lyapunov stability theory. The control method required few tuning parameters which was easy for engineering practice. Simulation and test results showed that compared with traditional LADRC, the proposed control method had stronger disturbance suppression ability and higher position tracking accuracy under the disturbed state, which provided a reference for the engineering application of ADRC.