Abstract:China has a large geographical area and a diverse agricultural variety. To withstand drought, cold, diseases, and pests, reduce field crop growing period, boost yield per unit area, and develop high-yield and effective agriculture, seedling growing and transplanting are crucial. Mechanized transplanting is a key technical link in crop mechanized production. Its technology does reflect the level of mechanization and modernization of agricultural production. That being so, the current state of research and development of mechanized transplanting technology and equipments for field crops in China and abroad was summarised, the structures and types of automatic seedling picking and planting equipments used in transplanters in China and abroad were analysed and summarised, and the working principle and structural features of diverse transplanting technologies and mechanisms were explained. The cornerstone of transplanting equipment innovation was transplanting mechanism innovation, and mechanism synthesis lied at the heart of mechanical transplanting mechanism innovation. The current state of transplanting mechanism configuration research and the dimension synthesis approach were discussed. Finally, the issues that risen in China’s development of mechanized transplanting technology and equipments were examined, and suggestions for future development were made. It was stated that efficient, precise, and reduced failure of seedling picking and planting technologies and mechanisms were important indications for automatic transplanting in paddy and dry fields, and that efficient automatic transplanting technologies and equipments based on information and intelligent technologies were an important direction for the future development of mechanized transplanting technology.

Abstract:The intelligentization of agricultural machinery is an inevitable trend of modern agricultural production. In order to improve the intelligence level of peanut harvester, the automatic driving operation system of peanut harvester was designed. Taking Dongtai Machinery’s 4HBL-2 self-propelled peanut combine harvester as the research platform, a manual and automatic integrated electronic control system with CAN bus communication interface was designed for the peanut harvester operating table, speed change mechanism and operating mechanism. PD control algorithm and Bang-Bang control algorithm were used to control the walking and operating system. According to the agronomic requirements of peanut harvesting operation, the joint operation strategy, automatic navigation path planning and path tracking control method of peanut harvester were designed. The automatic driving harvesting experiment was carried out on the cement road and the sandy soil peanut field at a driving speed of 0.25m/s. The results showed that the average absolute deviation of the straight line tracking of the peanut harvester on the cement road was 4.34cm and the maximum deviation was 9.30cm. The results of the field test on sandy soil showed that the average absolute deviation of the straight line tracking of the peanut harvester was 5.12cm and the maximum deviation was 12.20cm, which met the requirements of the combined peanut harvesting operation.

Abstract:In order to accurately identify the different fruit targets on apple trees, and automatically distinguish the fruit occluded by different branches, providing visual guidance for the mechanical picking end-effector to actively adjust the pose of apple picking to avoid the shelter of the branches, a real-time recognition method of apple picking pattern based on improved YOLOv5m for picking robot was proposed. Firstly, BottleneckCSP module was designed and improved to BottleneckCSP-B module which was used to replace the BottleneckCSP module in backbone architecture of original YOLOv5m network. The ability of image deep feature extraction of the original BottleneckCSP module was enhanced, and the original YOLOv5m backbone network was lightweight designed and improved. Secondly, SE module was inserted to the proposed improved backbone network, to better extract the features of different apple targets. Thirdly, the bonding fusion mode of feature maps, which were input to the target detection layer of medium size in the original YOLOv5m network, were improved, and the recognition accuracy of apple was improved. Finally, the initial anchor box sizes of the original network were improved, avoiding the misrecognition of apples in farther plant row. The experimental results indicated that the graspable, circuitous-graspable (up-graspable, down-graspable, left-graspable, right-graspable) and ungraspable apples could be identified effectively by using the proposed improved model in the study. The recognition recall, precision, mAP and F1 were 85.9%, 81.0%, 80.7% and 83.4%, respectively. The average recognition time was 0.025s per image. Contrasted with original YOLOv5m, YOLOv3 and EfficientDet-D0 model, the mAP of the proposed improved YOLOv5m model was increased by 5.4 percentage points, 22 percentage points and 20.6 percentage points, respectively on test set. The size of the improved model was 89.59% of original YOLOv5m model. The proposed method can provide technical support for the picking end-effector of robot to pick apples in different poses avoiding the shelter of branches, to reduce the loss of apple picking.

Abstract:Aiming at the problems that the irregular shape of wheat seeds leads to the high missing rate，low seed-filling qualified rate and difficult precision sowing of the traditional type hole wheel seed metering device under high-speed operation，a limited seed filling posture and combined type of positive and negative air pressure and wheat precision seed metering device was designed. The mechanical model of wheat seed was established by analyzing the movement and force of wheat seed during seed metering. The posture of seed filling was limited and the phenomenon of seed jamming was improved by adding arc-shaped auxiliary seed filling plate and seed stirring plate. The key structural parameters of the seed metering device were designed and calculated by analyzing the seed filling process and the distribution of seeds in the seed ditch，including the diameter of the type hole wheel was 100mm，the length of the hole was 8mm，the width of the hole was 5mm，and the depth of the hole was 3mm，etc. The optimum angle of arcshaped auxiliary seed filling plate was determined to be 5°through the simulation analysis with EDEM software. On the basis of the above research，the quadratic orthogonal rotation combination test of three factors and three levels was carried out with the rotary speed of type hole wheel，the vacuum degree and the rotary speed of seed stirring plate as the test factors and the seedfilling qualified rate，and the missing rate and the multiple rate as the evaluation indexes. The regression equation between the evaluation index and the test factors was constructed，and the significance of the influence of each test factor on the index was analyzed. By optimizing the test parameters，the optimal parameter combination was obtained：the rotary speed of type hole wheel was 66.27r/min，the vacuum degree was 3.52kPa，and the rotary speed of seed stirring disk was 52.00r/min. Under the conditions of above factors，the test verification was carried out，the seed-filling qualified rate was 92.70%，the missing rate was 3.47%，and the multiple rate was 3.83%. The experimental results showed that there was little difference between it and the optimization results，which showed that the seed metering device could meet the performance requirements of precision sowing for wheat.

Abstract:Aiming at the problem of poor operation effect caused by seed leakage due to the highspeed operation of the existing corn high speed air suction seed metering device, a double disturbance auxiliary filling high speed air suction seed metering device was designed by increasing the seed filling area, increasing the seed filling time of the seed metering disk, strengthening the dispersion of the population and reducing the adsorption pressure. The stress balance equation of seed layer with different heights was analyzed, and the position and structural parameters of expansion plate were calculated. The mechanical model of seed movement under the action of disturbing seed column and shaped hole was analyzed, and the key structural parameters of seed metering disc with middle shaped seed suction hole with disturbing seed column were determined. Taking the transient normal stress of particles as the evaluation index, the disturbance performance of the three seed trays was simulated by EDEM software, and the seed filling performance of the three seed trays was tested by bench test. The results showed that the designed seed trays can effectively strengthen the dispersion of the population and strengthen the seed absorption performance of the seed trays. The bench test results showed that when the operating speed of the high speed air suction maize seed metering device with disturbance assisted seed filling was 8~10km/h and the adsorption negative pressure was 3.0~4.0kPa, the leakage absorption rate was less than 5.1% and the gravity absorption rate was less than 4.2%, the qualified index of grain distance was greater than 94.6%, and the coefficient of variation of qualified grain distance was less than 15.33%. When the operating speed was 12~14km/h and the adsorption negative pressure was 3.5~4.0kPa, the leakage and absorption rate was not higher than 7.9% and the gravity rate was not higher than 1.3%, the qualified index of grain distance was greater than 92.1%, and the coefficient of variation of qualified grain distance was less than 17.67%. The operating performance was better at high speed, and all the indexes were better than that of the national standard.

Abstract:To solve the problems of seed filling difficulty and seed damage caused by the irregular shape and damaged epidermis of Pinellia ternata seeds, based on determining the physical properties of Pinellia ternate seeds, a kind of wheel-spoon type precision seed-metering device for Chinese herbal medicine Pinellia ternate with was designed, through the force analysis of seeds in the seed filling area and the seed clearing area, the working principle of the wheel-spoon type precision seedmetering device was expounded. The number of spoons, the rotation speed of seed wheel, the seed layer height, and the seed spoon hole radius of the seed-metering device were analyzed for the discrete element single factor simulation test. For the quadratic regression orthogonal rotation combined bench test, the rotation speed of seed wheel, the seed layer height, and the seed spoon hole radius were chosen as the test factors, and the qualified index, multiple index, and the missing index were taken as the test indexes. Regression models of three test indexes were established during the experiment. Experimental results showed that the primary and secondary factors affecting the qualified index was the rotation speed of seed wheel, the seed layer height, and the seed spoon hole radius. For the seed spoon hole radius of 7.5mm, the rotation speed of seed wheel of 17.0~19.0r/min, and the seed layer height of 123.0~133.0mm, the qualified index was more than 95.5%, and the missing index, the multiple index were less than 1.0% and 3.5%, respectively. The experiment results further proved that the wheel-spoon type precision seed-metering device for Chinese herbal medicine Pinellia ternata could fully meet the requirements of Pinellia ternate planting.

Abstract:In order to solve the low mechanization rate of ginseng sowing, a ginseng precision seeder with ditching and seeding monomer type was designed. Through the analysis of the seed dropping point of chain spoon type precision seed-metering device for ginseng and working performance and structural parameters of double disc trencher, the key parameters of ditching and seeding monomer were ascertained and the whole transmission system was designed to realize the adjustable plant spacing. Then through the construction of a soil trough test bench, the quadratic regression orthogonal rotation combination test was done, with the operation speed, ditching depth and relative horizontal distance between furrow opener and seed-metering device as the test factors, and the qualified index, the multiple index and the missing index as the text indexes. Experimental results showed that when the operation speed was 0.42m/s, ditching depth was 45mm and the relative horizontal distance between furrow opener and seed-metering device was 95mm, the qualified index was 94.53%, the multiple index was 4.308%, and the missing index was 1.165%. The field experiment was carried out with 2BS-10 ginseng precision seeder with ditching and seeding monomer type to testify the performance of the seeder, which showed that when the plant spacing was 4cm, the seeder’s qualified index was 92.7%, the multiple index was 5.0%, the missing index was 2.3%, the qualified rate of sowing depth was 95.1%, and no injury was found, which can meet the requirements of non-forest ginseng sowing.

Abstract:In order to solve the problem of the first season rice pile rolled by the track of the harvester in the production process of ratooning rice, a spatially planetary gear train planting mechanism with large deviation, direct harvesting and small lateral hole transplanting trajectory was proposed, starting from the perspective of mechanical planting with the condition of the effective number of strains per unit planting area and the crawler walking row distance by reasonably increasing the planting row spacing. Firstly, planning the wide and narrow rows trajectory and determining the key position posture by combining with planting agronomy, geometric equations for key 3 poses containing spatial constraints were established by constructing a simplified spatially open-chain 2R mechanism model. The parameters were solved by using homotopy algorithm. Then, the relative angular displacement relationship between the input and output axis of the open chain mechanism was obtained based on fitting and optimizing parameters. The mechanism drive ratio and the non-circular gears pitch curve were determined. The spatial trajectory that meets the transplanting requirements was recovered. Finally, a kind of wide and narrow rows transplanting mechanism of ratooning rice was designed by coupling the motion of the two rotating axes of the open chain mechanism with the non-circular gear pair, which was driven by the plane non-circular gear and helical gear. Virtual simulation and prototype test were carried out, and the results showed that the actual motion trajectory and attitude of the prototype were basically consistent with the theoretical design. The spatial track satisfying the characteristics of wide and narrow rows transplanting of regenerative rice with large offset from taking seeding to pushing seeding ΔS1 was 65.59mm, side push seedling angle γ1 was 16.13°and side hole width ΔS3 was 23.69mm, which verified the feasibility of gear train wide and narrow rows transplanting mechanism design.

Abstract:In order to solve the problem that the screen holes are prone to clogging during the operation of the drum screen type film and debris separator, resulting in poor screening performance and high debris content in the film, a screen hole clearing device was designed to disturb the flow field at the screen holes by spraying airflow from the nozzle and disrupt the balance of force of the clogged material at the screen holes. The device consisted of a centrifugal blower mounted diagonally on both sides of the drum screen type film separator, a pressure stabilizing pipe and an airflow duct on the outer edge of the drum. The critical air velocity for screen hole cleaning was determined to be 1.151m/s by theoretical analysis, and a positive correlation between fan air velocity and nozzle jet air velocity was investigated by using computational fluid dynamics simulation and curve fitting to determine a final fan air velocity of 9.2m/s. The prototype was designed and tested, and the results showed that 8.28% of the screen holes were clogged and the rate of debris in the film was 7.33%, which was 16.27 and 4.64 percentage points lower than before the installation of the screen hole clearing device. The average jet air velocity of each nozzle was 3.81m/s and the minimum value was 1.22m/s, which met the requirement of clearing the blockage, and the blockage rate of the screen pore and the rate of impurities in the film remained basically unchanged during the continuous operation of the clearing device. The research result can provide reference for the design of various forms of screen hole clearing devices.

Abstract:Existing rape combine harvester has the challenges to separate the small rapeseed from the threshing outputs whose components are mixed, highlighting the need to improve the cleaning rate to reduce the labor consumption in manual re-cleaning process. Therefore, a modularized re-cleaning cylinder sieve, which could hung on the grain tank, was designed. Based on the kinematics and dynamic analysis, the range of structure and operation parameters of the material lifting screw conveyor and screening device were analyzed. A three-factor and three-level orthogonal experiment was carried out by using EDEM. The loss rate, cleaning rate and screening efficiency were taken as the indexes. The rotating speed of cylinder sieve, the screw pitch of the spiral blade, and the diameter of the sieve hole were taken as the factors. The optimal parameter combination was determined by the orthogonal experiment, and verified by the bench verification experiment. The simulation results indicated that the optimal parameter combination of the diameter of sieve hole, the rotating speed of cylinder sieve, and the screw pitch of the spiral blade were 5mm, 105r/min and 250mm, respectively. When the feeding rate was 0.6kg/s, under this condition, the loss rate, cleaning rate, and screening efficiency of re-cleaning cylinder sieve were 0.92%, 98.96% and 95.12%, respectively. The results of bench experiment showed that the cleaning system with re-cleaning cylinder sieve could work smoothly. Under the optimal parameter combination, the loss rate, cleaning rate, and screening efficiency were 0.96%, 98.67% and 95.36%, respectively. Compared with the performance of the cyclone separation cleaning device without re-cleaning cylinder sieve, the cleaning rate was improved by 4.38 percentage points. The research could provide a reference for the structural improvement and optimization of cleaning device of rape combine harvester.

Abstract:Existing development of rape windrower, which could cut and windrow the plants, has the challenge to improve the trafficability and windrow quality to overcome the complicated field conditions, highlighting a need for exploration of the windrow equipment with simplified structure and stable trafficability. As a solution, the self-propelled middle-placement rape windrower with a high ground clearance crawler was designed. Except for the crawler, the transverse conveying device, cutting system, and hydraulic drive system were also designed and selected. Combined with the rapeseed cultivation agronomy, the laying process of plants was analyzed. The plant parameters, the technical parameters of the windrower and the cultivation agronomy requirements which directly and indirectly affected the quality of plants placement were determined. In order to verify the performance of the windrower, the trafficability test and field functional experiment were carried out. The trafficability test results showed that the average deviation degree of the windrower on hard road surface and soft soil was 0.73% and 1.28%, respectively, and the average turning radius of the windrower under unilateral braking condition was 1.91m and 2.03m, respectively. The windrower could work stable when driving uphill and downhill or overcoming the field ridge and ditch. To verify the function of the windrower, the field experiment was carried out. Under the condition of the forward speed of the machine was 0.7m/s, the reel rate was 30r/min,transverse conveying device rate was 240r/min, cutter crank rate was 320r/min, the average windrowing width was 968.7mm and the height was 389.4mm when harvesting the plants at the mature green stage. Besides, the average laying angle was 13.3°, and the difference between the upper and lower laying angles was 3.5°. When harvesting the plants at the yellow ripeness stage, the average windrowing width and height were 956.8mm and 468.3mm, respectively, while the average laying angle was 13.6°, and the difference of laying angle between upper and lower layers was 4.4°. The placement quality of harvesting rape plants at different maturity stages could both meet the basic requirement of rapeseed production. Furthermore, the soil compaction of each side of the crawler was almost the same, indicating that the mass distribution of the windrower was relatively reasonable. The research result can provide a reference for the structural design and optimization of rape windrower.

Abstract:In order to solve the problem of forage harvesting in hilly and mountainous areas, the folding mechanism of the hanging large-width lawn mower was optimized, the operation ability of the lawn mower under complex terrain conditions was enhanced, and the operation efficiency of the lawn mower was improved. In order to expand the swing angle range of the folding mechanism, the size requirements of each component were analyzed and calculated under the condition that the folding mechanism can achieve positive and negative swing of 30°. The ADAMS simulation software calculation showed that the maximum pulling force of the hydraulic cylinder of the folding mechanism was 51600N. It was calculated that the maximum stress value of the rotating arm was 216.67MPa, and according to the theoretical calculation results, a folding mechanism suitable for a 3.2m wide lawn mower was trialproduced. Folding mechanism, a matching hydraulic profiling system was developed, the profiling system adopted the method of providing auxiliary oil source for the lifting hydraulic cylinder, and was connected to the accumulator through a check valve and a stop valve at the oil inlet of the hydraulic cylinder. The initial volume of the accumulator was 2.0L, and the pre-charge pressure was 7.0MPa. The ADAMS-AMESim co-simulation technology was used to simulate and analyze the folding mechanism and the profiling system. The volume variation range of the airbag of the cutter was 0.4~0.7L, and the variation range of the ground pressure of the cutter was 1700~2500N. Finally, the trial-produced profiling system was mounted on the folding mechanism, and a field test was carried out in Sanshan District, Wuhu City, Anhui Province. The results showed that the prototype can successfully complete the folding action of the lawn mower and meet the mechanical and strength requirements under various working conditions. The mower had a swing range of ±30°;the test prototype equipped with the profiling system can smoothly pass through the 250mm high wave convex road, which improved the terrain adaptability of the lawn mower in hilly and mountainous areas, and can be used for hanging mowers. It provided a reference for the design of the lawn mower folding mechanism and the grounding profiling technology of the lawn mower.

Abstract:Compared with field corn, seed corn peeling operation has higher agronomic requirements on the bracts peeling rate, grains crushing rate, and grains falling rate. Because of the lack of efficient and low-loss peeling means for seed corn ear, TRIZ theory combined with explicit dynamics simulation and high-speed camera technology was used to study the design method of seed corn peeling mechanism. Firstly, based on TRIZ theory, the critical structural design of the seed corn peeling mechanism was solved, and the detailed design of the peeling mechanism was completed. Secondly, LS-DYNA was used to simulate the explicit dynamics of the peeling system-seed corn ears. The movement process and stress of seed corn ears were analyzed, which verified the rationality of the peeling mechanism design. In addition, a high-speed camera test-bed was set up. Through the frame-by-frame analysis of a high-speed camera in the peeling process of seed corn and compared with the simulation results, the maximum error of speed under three working conditions was 0.035m/s, 0.066m/s and 0.095m/s, respectively, which verified the rationality of the sectional design of peeling roller. Finally, the seed corn peeling experiment was carried out by selecting the net rate of bract leaf peeling, the rate of grain breaking and the rate of falling grains as performance indexes. Under three working conditions, the test results met the performance index requirements of the seed corn peeling mechanism.

Abstract:A bamboo shoots peeling machine with rolling friction feeding based on knife-cutting method was designed according to the low degree of mechanization of bamboo shoots peeling in China. Based on the physical characteristic parameters of bamboo shoots and the principle of artificial bamboo shoots peeling, the force and movement analysis in the peeling process were described in detail, and the main factors affecting peeling efficiency, damage rate, no-hell rate were identified as the blade installation angle, the rotation speed of the peeling roller, the axis height difference between the roller and the peeling roller. On this basis, the design principle of the key components and the bamboo shoots peeling machine were given. In order to obtain the best experimental materials for the prototype experiment, single factor experiment was carried out with the length and base diameter of bamboo shoots as experimental factors, and the length from 300mm to 320mm and the base diameter from 29mm to 32mm were selected as orthogonal experimental materials samples for the bamboo shoots peeling machine. The Design-Expert software was used to design the orthogonal experiment and determine the optimal parameter combination of peeling quality according to the actual work situation. The results showed that when the blade installation angle was 30.12°, the rotation speed of the peeling roller was 229.18r/min, the axis height difference between the roller and the peeling roller was 15.43mm, the damage rate of bamboo shoots was 6.81%, and the no-shell rate was 94.59%. The verification experiment was carried out under these conditions, and the results showed that the damage rate and no-shell rate were 7.10% and 93.22%, respectively, which were basically consistent with the optimized parameters. This product can meet the requirements of bamboo shoots peeling.

Abstract:Aiming at the problems of serious magnetic field interference, poor dynamic performance of magnetometer calibration and low accuracy of UAV attitude estimation, a two-stage heading and attitude estimation method based on real-time magnetometer calibration was proposed. According to the characteristics of small variation of geomagnetic field vector, the real-time calibration model of magnetometer was established by using Levenberg-Marquardt (LM) algorithm and magnetometer error model, and the error parameters of magnetometer were calculated in real time. Considering the disturbance of motion acceleration, motor magnetic field and environmental magnetic field, the unscented Kalman filter (UKF) was used to fuse gyroscope and accelerometer to realize the first-stage attitude estimation, and the attitude information of roll angle and pitch angle was accurately analyzed through quaternion. The second-stage attitude estimation combined the realtime calibration data of the magnetometer and the gyroscope to correct the heading angle, and finally realized the accurate estimation of the UAV attitude and heading. The test results showed that when the external magnetic field interference was up to 30.97μT, the real-time calibration algorithm can still quickly calculate the calibration parameters of the magnetometer, and the mode length root mean square error was 0.59μT, which reduced the noise of heading observation information. The root mean square error of the attitude angle of the attitude measurement system was no more than 0.75°, and the root mean square error of the heading angle was 1.40°. Compared with that of complementary filtering algorithm, the attitude angle accuracy was increased by 0.6°, and the heading angle estimation accuracy was improved by 1.38°. In the dynamic flight test, the attitude estimation algorithm greatly reduced the influence of magnetic interference, the attitude tracking was accurate, the heading angle converged quickly, and the steady-state accuracy was higher.

Abstract:Aiming at the leakage problem of axial tip clearance of the liquid ring pump impeller, dielectric barrier discharge plasma excitation was proposed to control the gas phase leakage flow in the axial clearance of the liquid ring pump. The interference of plasma on leakage flow field under different excitation voltages was simulated by coupling phenomenological model, RNG k-εturbulence model and VOF gas-liquid two-phase model. The control mechanism of plasma excitation on clearance leakage flow field was investigated. The results show that the direction of the wall jet induced by plasma excitation was opposite to that of the clearance leakage flow. The reverse wall jet can effectively suppress the leakage flow intensity, improve the secondary flow caused by the clearance leakage flow to a certain extent, and reduce the clearance leakage flow loss. Meanwhile, the vortex structure was induced by plasma excitation in the non-clearance near-wall region, which caused additional hydraulic loss. The excitation voltage and position had an important influence on the control effect of leakage flow. The plasma flows control effect of 15kV excitation voltage was significantly better than that of 10kV excitation voltage. When the excitation position was near the outlet of blade tip clearance, the plasma excitation had a good suppression effect on the leakage flow. The research results can provide a theoretical reference for the performance optimization of liquid ring pumps.

Abstract:It is of great significance for agricultural resources monitoring to accurately extract cultivated land map information from remote sensing images.To improve the defects of traditional models for extracting cultivated land and solve the problem that most FCN model pays more attention to accuracy but ignores the consumption of time and computing resources, a lightweight model for extracting cultivated land map spots was established based on FCN (LWIBNet), and post-processing combined with mathematical morphology algorithm were used to carry out automatic extraction of cultivated land information. LWIBNet drew on the advantages of lightweight convolutional neural network and U-Net model, and it was built with the core of Inv-Bottleneck (composed of deep separable convolution, compression-excitation block and inverse residual block) and the skeleton of efficient coding-decoding structure. Compared LWIBNet with the cultivated land extraction effect of traditional model, and the computational resources and time consumption of classical FCN model.The results showed that LWIBNet was 12.0% higher than the Kappa coefficient of the best traditional model, and compared with U-Net, LWIBNet had 96.5%, 87.1%,78.2% and 75% less parameters, calculation, training time and split time-consuming, respectively. Moreover, the segmentation accuracy of LWIBNet was similar to that of the classical FCN model.

Abstract:The mixed-cell cellular automata (MCCA) model improves the traditional cellular automata (CA) model, and introduces mixed cells based on the actual complex land structure, realizing the progress from qualitative and static simulation to quantitative and dynamic simulation. The applicability of the MCCA model in the Gan-Lin-Gao area (Ganzhou District, Linze County and Gaotai County) in the middle reaches of the Heihe River was firstly explored;after that, the multiple-objective programming (MOP) model and the ordinary linear regression model were separately used to predict the area values of different land use types in the sustainable development (SUD) scenario and the basic development (BAD) scenario in 2035, and then the area number was input into the MCCA model to visualize the land use spatial structure of different scenarios, and carry out comparative studies. The results showed that all accuracy evaluation indicators indicated that the simulation accuracy of the MCCA model was relatively high. The Kappa coefficient, mixed-cell figure of merit (mcFoM) and mean relative entropy (RE) were 0.886, 0.261 and 0.508, respectively, which was better than the patch-generating land use simulation model (PLUS) based on pure cells, so the MCCA model was suitable for the simulation of land use structure in the study area. In 2035, the scope of forest land in the SUD scenario was significantly higher than that in the BAD scenario, and its ecological benefits increased faster than that of the BAD scenario, construction land and arable land expand moderately, and the comprehensive benefits increased relatively fast. The results showed that the optimal land use allocation scheme simulated by coupling the MOP and MCCA model can better coordinate the relationship between economy and environment, which was not only conducive to rapid economic development, but also protected the ecological environment and maintains social stability.

Abstract:The traditional wheat area extraction methods mainly depend on artificial field investigation, which shows some disadvantages such as a big workload, low efficiency and high cost. Conversely, remote sensing technology has the advantages of high accuracy, rapid response and dynamic monitoring. It has become an effective measurement to extract crop areas. The Landsat-8 satellite remote sensing image of Zhengding County in Shijiazhuang was used as the training data, the image of Zengcun Town in Gaocheng District was used as test data. The GF-6 with resolution of 8m and Sentinel-2 with resolution of 10m were selected as comparative validation data. An improved U-Net was proposed to extract winter wheat planting areas. Landsat-8 was firstly preprocessed and the label set of wheat areas were marked and trained by using the U-Net network. The Squeeze and excitation (SE) attention mechanism module was introduced to better consider the information between feature channels, and the Batch normalization (BN) layer was used to suppress the over-fitting problem. The classification results were obtained through the Softmax classifier. SegNet, Deeplabv3+ and U-Net were selected as the comparison models and GF-6, Sentinel-2 and Landsat-8 data were used to construct the models, respectively. The results showed that the SE-UNet network performed best in the test data set based on Landsat-8 data prediction model, with the MPA and MIoU of 89.88% and 81.44%, respectively. This method can provide a reference for identifying large-scale winter wheat planting areas.

Abstract:With the development of unmanned aerial vehicle (UAV) and remote sensing technology, crop yield estimation through rapid acquisition of multitemporal and highresolution remote sensing images at field scale has become a research hotspot. In order to determine the optimal growth stage and sampling times for winter wheat yield estimation by UAV multispectral remote sensing, a field experiment on winter wheat in sandy soil was conducted, which was divided into four groups (36 management zones) by irrigation level and five groups (15 management zones) by nitrogen application level. Then the multi-spectral remote sensing images of eight growth stages for winter wheat from rising to late filling were collected by the UAV platform. Additionally, partial least squares (PLS), random forest (RF), least absolute shrinkage and selection operator (LASSO) were used to establish the yield prediction model of winter wheat at each growth stage. Based on the optimal model selected, five yield estimation schemes for the vegetation indices integration during specific growth periods were developed by the cubic B-spline curve and compound trapezoidal formula. The results showed that significant differences were found for estimation accuracy at different growth stages, which was increased with the growth of winter wheat. In single growth period, the optimal growth periods of PLS, RF and LASSO models were early filling, early filling and late filling, respectively. Compared with PLS and LASSO models, RF had the best precision in estimating winter wheat yield except early joint stage. The accuracy of yield estimation in the multi-growth stages was better than that in a single one. The optimal yield estimation scheme was the vegetation indices from rising to the late filling stage for eight sampling times of remote sensing (the determination coefficient R2 of 0.96 and the normalized root mean square error （NRMSE） of 5.39%). Meanwhile, the yield estimation scheme of six sampling times from rising to flowering stage also performed excellently (NRMSE of 9.16%), which meant that it can not only reduce sampling times and remote sensing cost, but also can predict the winter wheat yield in advance. The results were of great significance for the accurate prediction of winter wheat yield by UAV remote sensing.

Abstract:In order to further accurately and real-time monitor the growth of winter wheat and estimate its yield, taking Guanzhong Plain in Shaanxi Province as study area, and vegetation temperature condition index (VTCI), leaf area index (LAI), fraction of photosynthetically active radiation (FPAR) at the ten-day or growth stage scales were selected as remotely sensed characteristic parameters. The GRU model was constructed based on different input parameters and time scales to obtain the growth comprehensive monitoring index I of winter wheat. The results showed that the accuracy of the models at the ten-day scale were generally higher than those of the growth stage scales. Based on the five-fold cross-validation method, the robustness of the multi-parameter GRU model on the ten-day scale was further verified, and the winter wheat yield was estimated based on the linear regression model between the growth comprehensive monitoring index I and the official yield records. The results showed that the R2 between the estimated and official yield records of winter wheat was 0.62, the RMSE was 509.08kg/hm2, the mean relative error (MRE) was 9.01%, and the correlation reached the extremely significant level (P<0.01), indicating that the multi-parameter yield estimation model at the ten-day scale can accurately estimate the yield of winter wheat in the Guanzhong Plain. The distribution of yield presented the spatial characteristics of high yield in the west and low yield in the east, and the inter-annual change characteristics of overall stability and steady growth. In addition, based on the GRU model, the cumulative effect of winter wheat growth was captured, and the influence of inputting parameters step by step in consecutive ten days on yield estimation was analyzed. The results showed that the model had the ability to identify the key growth stages of winter wheat, and late March to late April was the critical period for the growth of winter wheat.

Abstract:In order to make up for the defects of the one-time modeling analysis and improve the operation efficiency and accuracy of wheat stripe rust remote sensing detection model, based on the characteristics of model population analysis (MPA) algorithm and the advantages of spectral interval selection algorithm and spectral point selection algorithm, a feature variable selection algorithm was proposed, combining correlation coefficient (CC) and MPA. Based on the selection of feature variables by CC algorithm for the full band spectrum, competitive adaptive reweighted sampling (CARS) and variable combination population analysis (VCPA) developed based on MPA were used to further optimize the feature variables sensitive to wheat stripe rust, and partial least squares regression (PLSR) algorithm was used to construct CC-CARS and CC-VCPA models for remote sensing monitoring of wheat stripe rust. The results showed that the accuracy of CC-CARS and CC-VCPA models constructed by combining the feature variables selected by CC-MPA algorithm was higher than that of CC, CARS and VCPA algorithm. In the three groups of validation set samples, CC-CARS model compared with CC model and CARS model, the R2V between predicted disease index (DI) and measured DI was increased by at least 6.78% and 6.66%, RMSEV was decreased by at least 15.31% and 10.98%, and RPD was increased by at least 18.08% and 12.34%, respectively. Compared the CC-VCPA model with CC model and VCPA model, the R2V between predicted DI and measured DI was increased by 9.58% and 0.73%, RMSEV was decreased by 20.78% and 3.86%, and RPD was increased by 26.22% and 4.02%, respectively. The spectral feature optimization algorithm based on CC-MPA was an effective feature selection method. In particular, the number of feature variables selected by CC-VCPA method was less and the model prediction effect was better. The research results had important reference value for spectral feature optimization and improving the accuracy of remote sensing monitoring of crop diseases.

Abstract:UAV-satellite remote sensing scale-up transformation method can effectively improve the monitoring accuracy of soil salt content. Sand trench canal irrigation area in Hetao Irrigation Area of Inner Mongolia was taken as the study area, the surface soil in bare soil period in April was taken as the research object. The dominant class variability-weighted method, local average method and nearest neighbor method were used to scale up the quadruple-band image (0.1 m) of UAV in the experimental area to the same scale as GF-1 satellite (16m). Subsequently, three combinations of variables were introduced as the input variables of the model for the UAV dataset and GF-1 satellite dataset, and the quantitative monitoring model of soil salt content was constructed by using multivariable linear regression (MLR) and back propagation neural networks (BPNN). On this basis, the GF-1 satellite data was modified by the mean band ratio method, and the scale-up inversion of soil salinity in the study area based on satellite factors was realized. The results showed that the dominant class variability-weighted method had the best monitoring effect, followed by the local average method. The nearest neighbor method had the worst monitoring effect among the three UAV-satellite remote sensing scale-up transformation methods;after comparing the four statistical evaluation indexes of mean value, standard deviation, information entropy and average gradient with the original UAV image, it was found that the quadruple-band UAV image pushed by the three methods had scale differences with the original image data to different degrees;by comparing R2 and RMSE of three variable combinations based on different data sources, it was found that the accuracy of the model constructed by the dominant class variability-weighted method was better than that of the other three data sources as a whole, and the scale-up dataset of the dominant class variability-weighted method based on mixed variable groups achieved the best monitoring effect in MLR model and BPNN model;the monitoring model with the best validation effect was multivariate linear regression model, its validation R2 was 0.420, RMSE was 0.219%. The research results can provide reference for integrated monitoring of farmland soil salt content in bare soil period by multi-spectral remote sensing of satellite and unmanned aerial vehicle.

Abstract:In order to explore the feasibility of GF-1 satellite inversion of farmland soil moisture content（SMC）under the condition of vegetation coverage, taking Shahaoqu District of Hetao Irrigation Area as study area, and GF-1 satellite remote sensing images as the data source. Simultaneously, the soil moisture content data were collected with various depths at 0～20cm, 20～40cm, 40～60cm, 0～40cm,and 0～60cm. Then a set of independent variables, including four bands and 15 spectral indices were obtained based on the GF-1 data, and the full subset selection was used to select the optimal combination of independent variables at five depths. Based on these, the combinations before and after full subset selection were used to build soil moisture content inversion models（multiple linear regression, MLR;back propagation neural network, BPNN;support vector machines, SVM）at five depths in the vegetated area, and evaluate the sensitivity of GF-1 to SMC at different depths and the inversion capability of the models. The model performance was assessed by using adjusted coefficient of determination (R2adj) and root mean square error (RMSE). The results showed that the model inversion accuracy was greatly improved after the full subset selection, and the overfitting phenomenon can be reduced. The sensitivity of GF-1 to the SMC at different depths under vegetation coverage was ordered from the largest to the smallest as follows: 0~40cm, 0~60cm, 20~40cm, 0~20cm, and 40~60cm. The SMC inversion capabilities of all the three models under vegetation coverage ordered from the largest to the smallest were as follows: BPNN, SVM, and MLR. After the full subset selection, the R2adj of the modeling set and verification set of BPNN at depth of 0~40cm can reach more than 0.50, and the RMSE was within 0.02%. The research result can provide a reference for using GF-1 satellite to monitor SMC of farmland under vegetation coverage.

Abstract:In order to extract the lodging area timely and accurately, a lodging area extraction model, namely Attention_U2-Net, was proposed. By integrating multi-scale features and based on U2-Net, Attention_U2-Net employed non-local attention mechanism to replace the hole convolution with large step size, expanded the receptive field of high-level network and improved the recognition accuracy of ground objects with different sizes, and utilized channel attention mechanism to improve the cascade mode and enhanced the accuracy. A multi-level joint weighted loss function was designed to balance the difficult and easy samples, and solve the challenge of imbalance between positive and negative samples. Patch-based pipelines were utilized to extract the lodging area. Experimental results on the self-built dataset showed effectiveness of Attention_U2-Net. The precision rate was 86.53%, the recall rate was 89.42%, and the F1 value was 87.95%, respectively. Compared with FastFCN, U-Net, U2-Net, FCN, SegNet and DeepLabv3, Attention_U2-Net achieved the highest F1 value and showed strong robustness and extraction accuracy. Compared with the labeled area, the extracted area obtained by Attention_U2-Net via cropping method was the closest one, and the accuracy rate of lodging area can reach 97.25%. Meanwhile, the false detection area of Attention_U2-Net was the smallest among all models. Experimental results showed that Attention_U2-Net had strong robustness and high segmentation accuracy, which can be utilized as a valuable reference for UAV remote sensing of wheat affected area and loss assessment.

Abstract:Cucumber plug seedlings usually grow at different speeds in the growth process. In order to make the cucumber plug seedlings in the unified growth stage before leaving the factory, it is necessary to detect late emergence seedlings. An automatic detection method for late emergence of plug seedlings was proposed based on point cloud processing.The RGB-D camera was used to build the point cloud collection platform for plug seedlings. Through conditional filtering, statistical filtering and Euclidean clustering, the point cloud of plug seedling leaves could be segmented. Adopting α-shape algorithm calculation, and then the leaf area of plug seedlings was calculated by fitting method. Average error between the fitting value and the true value was 0.75cm2, and average relative error was 8.51%. The method of locating the seedling stem top positions based on the principal curvature was used to automatically obtain the plant height. The average error between the true values and the calculated values of plant height was 0.359cm, and the average relative error was 9.32%. The product of leaf area and plant height was used as the grading coefficient, and the value of subtracting the standard deviation from the mean value of the current seedling grading was used as the threshold for the classification of late emergence seedlings, so as to realize the automatic detection of late emergence seedlings in the plug tray. Comparing the calculated grading coefficient with the total fresh weight of seedlings, the change trend of the two was basically the same (little difference). The grading coefficient of late emergence seedlings with small total fresh weight was significantly lower than that of other normal seedlings. The proposed grading coefficient can effectively describe the growth of seedlings. The results showed that the success rate of the automatic detection method for late emergence seedlings in plug trays based on point cloud processing reached 95%, which can provide technical support for the detection of seedlings in industrial.

Abstract:Rice planthopper is one of the most important pests of rice, which mainly includes white back planthopper, brown planthopper and small brown planthopper. In order to realize the rapid and accurate identification of rice planthoppers and prevent the same insect from being repeatedly identified and classified, the object detection algorithm combining image redundancy elimination and CenterNet network was proposed. Firstly, the field insect collection device independently developed by the team was used to automatically obtain insect images and make a data set. The data set was divided into four classes which included white back planthopper, brown planthopper, small brown planthopper and non-rice-planthopper. Secondly, for the live images with high similarity obtained by the field insect collection device, CenterNet with image similarity detection, image subtraction, image thresholding and bilateral filtering image redundancy elimination algorithms were combined, and a deep feature fusion network (deep layer aggregation, DLA) was selected, which was used as the backbone network to extract the characteristics of insects. Compared with the classic machine learning and deep learning models used in rice planthopper detection in the past, it had obvious advantages. The experiment results showed that for the preprocessed test set, the algorithm can not only quickly process insect images, but also can successfully solve the problem of insect repeated detection. The mean average precision was 88.1%, and the detection rate was 42.9f/s. The research effectively completed the identification and classification of the three types of rice planthoppers, and showed good generalization ability for insects collected in different time periods, which can be used for intelligent early warning and forecasting of rice pest outbreaks in the later period.

Abstract:It was difficult to take high-quality images when pests were still and in close distance in the natural environment of rice field, which led to the problem that satisfactory identification accuracy could not be achieved when using the actual environmental identification model detection. A low-resolution rice pest image recognition method based on self-calibrated convolutions and ResNeSt block for ResNet50 (SCResNeSt) was proposed. Firstly, the enhanced super-resolution generative adversarial networks (ESRGAN) super partition network was used to enhance the data of low-resolution images to solve the problem of less effective information about rice pests. In SCResNeSt network, three consecutive 3×3 convolutional layers were used to replace the first 7×7 convolutional layer to reduce the computational cost. Using self-calibrated convolution instead of the 3×3 convolution in layer 2, through internal communication, the field of view of each convolutional layer was explicitly extended to obtain part of the background information of pest images, to enrich the output features. The split-attention network block (ResNeSt block) was used in the backbone network to further improve the accuracy of obtaining pest information in the image. Finally, the optimized model was deployed on the mobile terminal, and a lightweight mobile rice pest identification system was developed. The experimental results showed that compared with the existing methods, the ESRGAN model could recover the real information about crop pests, and the SCResNeSt model could effectively improve the performance of rice pest identification, the accuracy can reach 91.20%, which showed that the depth model could accurately identify rice pest types. The research result can provide an important technical basis for the intelligent identification and control of rice pests, and it would improve the level of rice production informatization.

Abstract:Aiming at the problem of how to efficiently mine prescription big data and assist in accurate diagnosis, tomato virus disease, tomato late blight and tomato gray mold were selected as the research objects, and an intelligent diagnosis model of tomato disease based on Bayesian optimization LightGBM was constructed to explore the data mining and accurate diagnosis of crop disease prescription. The primary data (text data label and One-hot coding, etc.) were preprocessed, and the features of crop disease prescription data were further extracted by recursive feature elimination method based on Wrapper. The tomato disease diagnosis model was constructed based on LightGBM algorithm, and compared with the running results of K-nearest neighbor (KNN), decision tree (DT), support vector machine (SVM), random forest (RF), gradient boosting decision tree (GDBT), AdaBoost and XGBoost common machine learning models. An Android mobile terminal plant doctor disease diagnosis APP was designed based on LightGBM model. The experimental results showed that the comprehensive diagnosis accuracy of LightGBM model based on Bayesian optimization can reach 89.11%, which was 3.65 percentage points higher than that of other seven machine learning models on average. At the same time, the LightGBM model after feature selection reduced the difficulty of data collection in the early stage on the basis of ensuring the accuracy of the model, and the comprehensive accuracy of the model was improved to 89.34%. Among them, the diagnostic accuracy of tomato virus disease and F1-score could reach more than 96%, and the running time was reduced by 47.73%. Finally, the generalization ability of the proposed model was tested by tomato leaf mildew and tomato early blight, and the experimental results indicated that the model had strong generalization ability and practicability. The APP designed based on LightGBM model can realize user friendly interactive visualization and meet the actual diagnostic needs.

Abstract:In view of the complex links of the pufferfish supply chain, high regulatory requirements, complicated information transmission in the traceability system, and low data query efficiency, in order to improve the efficiency and security of pufferfish traceability query, a type of pufferfish supply chain traceability optimization model based on blockchain technology was established, and a corresponding system was built. Firstly, the supply chain information of the pufferfish supply chain business was analyzed, the traceability source information and product information of each business link of the supply chain was sorted out and extracted;and then based on the blockchain technology, the trusted traceability optimization model architecture of the pufferfish supply chain was established, and a multi-chain storage mode and fast query process and formulate corresponding smart contracts were designed;finally, a pufferfish supply chain information traceability system was implemented based on Hyperledger Fabric, and tested by using the Caliper performance test tool. The results showed that when the number of data records was greater than 1000, the query efficiency of the model would be higher than that of the traditional single-chain model. After 10000 data records were uploaded to the chain, the query efficiency of this model was about 92.9% higher than that of the traditional single-chain model. Taking a pufferfish enterprise in Jiangsu Province as an example, the information security transmission and rapid traceability of the pufferfish supply chain were realized. The proposed model can be applied to the pufferfish industry to improve the traceability efficiency and security, and it can provide a tamper-proof and high query efficiency model for the pufferfish industry.

Abstract:In order to explore greenhouse gas emission and soil mineral nitrogen characteristics of paddy field under different irrigation methods in northeast black soil region of China, three test treatments (controlled irrigation (KG), intermittent irrigation (JG) and wet irrigation (CI)) were set according to different irrigation methods, with local conventional transplanting and inundation (CK) as control. The processes of greenhouse gas methane (CH4) and nitrogen oxide (N2O) emissions, global warming potential, global warming potential based on yield and NH+4N and NO-3N contents in 0~60cm soil of paddy fields with different treatments were studied, as well as the correlation between soil temperature in 0~20cm soil layer and mineral nitrogen and CH4 and N2O emissions. The results showed that with the advancement of rice growth and development, the soil temperature of each layer of paddy soil in each treatment was increased first and then decreased. The emission of CH4 and N2O was increased at first and then decreased in reverse V-shaped trend. The emission peaks of CH4 and N2O appeared at jointing and booting stage and heading and flowering stage, respectively. In terms of time, inflection points of NH+4N content in paddy soils treated with CK, JG, CI and KG appeared at midtillering stage, heading and flowering stage, jointing and booting stage and milking stage respectively, while the maximum NO-3N content in all treatments occurred at early tillering stage. Spatially, the average content of NH+4N in paddy soils of different treatments was gradually decreased with the increase of soil depth, while the average content of NO-3N in CK treatment was gradually increased with the increase of soil depth, while the other treatments were decreased first and then increased. There was a significant correlation between soil temperature and CH4 emissions, but no significant correlation with N2O emissions. NH+4N in each treatment soil was positively correlated with CH4 and N2O emissions, while NO-3N in soil was negatively correlated with CH4 and N2O emissions. Cumulative emissions of CH4 from large to small in paddy fields were CK, JG, KG and CI, and accumulative emissions of N2O from large to small in turn were CI, KG, JG and CK. Cumulative emissions of CH4 and N2O from each treatment were significantly different from those of CK treatment (P<0.05). In terms of greenhouse effect per unit yield (GWPy), KG, JG and CI treatment were decreased by 24.98%, 27.69% and 24.06%, respectively compared with that of CK treatment. The research results can provide theoretical basis and technical support for reducing emission of paddy field and improving utilization rate of mineral nitrogen in soil in northeast black soil region of China.

Abstract:At present, the formulation of the irrigation quota of cotton under film drip irrigation condition is still based on the field irrigation experiments, while there are few studies on determining the irrigation quota of cotton under film drip irrigation condition by considering the soil moisture content during sowing time or by modeling. Through the continuous field test of cotton field trials in the southern Xinjiang region in 2017 and 2018, the DSSAT-CROPGRO-Cotton model was parameter calibrated and verified by using the measured biomass, soil moisture contents, leaf area index and yield data of cotton flowering and maturity in 2017 and 2018, respectively. Under film drip irrigation condition, three cotton irrigation quota levels, 24mm, 30mm, and 36mm, were designed in the experiment. The validated DSSAT-CROPGRO-Cotton model was also used to simulate the growth and yield of under-membrane drip irrigated cotton under eight different initial soil moisture content conditions of 1.2θFC, 1.1θFC, θFC, 0.9θFC, 0.8θFC, 0.7θFC, 0.6θFC, and 0.5θFC (θFC is the field water holding rate). The results showed that the simulated values of cotton phenology, soil moisture contents, leaf area index and cotton yield were well agreed with the measured values after parameter calibrated and model verification, meeting the requirements of cotton simulation accuracy under film drip irrigation condition in the field. However, the simulated biomass values deviated obviously from the measured ones. At the same time, based on the verified DSSAT-CROPGRO-Cotton model, the yield and biomass of cotton were simulated under different conditions of initial soil moisture content and irrigation quota. The results showed that the maximum simulated yield and biomass of cotton corresponded to an initial soil moisture content of 0.8θFC~θFC. Accordingly, irrigation quota should be between 330mm and 396mm during cotton-growing period. The simulation results can be used for reference in the cotton planting and irrigation management in southern Xinjiang.

Abstract:In order to improve the hydraulic performance and anti-clogging performance of Y-type screen filter, taking the angle between outlet and cylinder, the inlet reduction ratio, the height of guide vane and the center angle of guide vane as experimental factors, the filtration process inside the filter was studied by the method of CFD-DEM coupling and physical test. Through the range analysis of a series of numerical simulation orthogonal test results, the sensitivity of various factors to water head loss, area proportion of medium speed flow area, relative standard deviation of particle distribution and interception rate was explored. The results showed that the medium and high speed flow areas were concentrated on the exit side, and the exit cross-sectional area was positively correlated with it when the angle between the outlet and the cylinder was small, and the larger the corresponding medium-speed flow area was, the better the hydraulic performance was. The concentration of particles through the region corresponds to the medium-speed flow area, and the increase of the inlet reduction ratio made the more uniform particle distribution, the better the anti-clogging performance can be. The combination of structural optimization parameters was as follows: the angle between outlet and cylinder was 35°, the inlet reduction ratio was 22/26, the height of guide vane was 10mm, and the center angle of guide vane was 90°. Among them, the included angle had the most significant influence on the comprehensive index, which was the key parameter of structural design. Compared with that before improvement, the head loss was reduced by 36.6%, the relative standard deviation was reduced by 43.26%, the interception rate was increased by 3.93 percentage points, and the area proportion of medium-speed flow area was increased by 15.77 percentage points, which showed the effectiveness of the optimization scheme. The research results can provide some reference for the optimal design of screen filter.

Abstract:Greenhouse environmental regulation can effectively improve crop production conditions. But due to the cost, acceptance and other reasons, most farmers use experience-based methods or simpler methods to control the greenhouse environment. In this situation, how to simply and effectively regulate greenhouse temperature and humidity accurately is a pressing issue in current greenhouse production. Firstly, a modern greenhouse intelligent monitoring system based on Internet of things was designed. The experiments of cooling, dehumidification, humidification and fully closed equipment were designed, and the mathematical models of temperature and humidity under closed greenhouse were constructed. By comparing the simulated values without equipment operation under the same conditions with the measured values after equipment regulation, a temperature and humidity control method was proposed according to the time period of crop growth demand and the size of equipment control capacity, and the experimental verification was carried out. The results showed that the subsection and multi-interval control strategy can effectively regulate the temperature and humidity. The temperature within 59.46% of the day was in the target range, and the humidity within 66.80% of the day was in the target range. The equipment operated stably without frequent opening and closing of the equipment. The subsection and multi-interval control strategy put forward an idea for greenhouse environmental control, combined with the typical IoT structure, remote automatic control can be realized, and the research results can be directly applied to practice.

Abstract:The collector is one of the main components of the axial flow fan, which has a great influence on the fan performance. However, the problems of unreasonable structure design, high energy consumption and air volume loss are common in the collectors of agricultural axial flow fans. Effects of the inlet length (L), fillet radius (R), and the outlet diameter (D) of the collector on the fan performance and flow field were studied, taking a typical 24 inch negative pressure FRP axial flow fan as the research object, using wind tunnel tests and CFD simulations. Besides, single factor analysis and response surface analysis were used in this process. The results showed that the suitable ranges of the dimensionless parameters of the collector were 1.00≤EL≤1.46, 2.95≤ER≤3.22, 0.9898≤ED≤0.9971. The effects of each influencing factor on the performance of the fan were obtained through the response surface model. When L=149.27mm, R=321.68mm, and D=678.00mm, the fan performance can be achieved better, the optimized Q was increased by 7.58% and N was increased by 8.07% compared with the prototype fan, the Q=10.06145m3/h, N=20.24m3/(h·W). The numerical simulation indicated that the optimized Q was increased by 5.86% and N was increased by 6.79% compared with the prototype fan, the Q=9900.54m3/h, N=20.03m3/(h·W). The optimized collector can significantly increase the positive vortex value in agricultural axial fan shroud，the maximum positive vortex of the leaf top clearance was 2197.21s-1, and its distribution area was increased in agricultural axial fan hub and shroud. The research result can provide a parameter optimization method to improve the airflow and energy efficiency ratio for agricultural axial fan collectors.

Abstract:The duration of light and the temperature in winter counter-seasonal production in solar greenhouses are generally unable to meet the growth demands of crops. Based on the temperature and light requirements for crop growth, the light equilibrium point to maintain a stable greenhouse temperature, the nighttime minimum temperature prediction model, and the empirical uncovering/covering time were used as constraints. The method of uncovering decision with the constraints of crop low temperature limitation point, light compensation point (LCP), light balance point and empirical uncovering time, and the method of covering decision with the constraints of nighttime minimum temperature prediction model, crop low temperature limitation point, light compensation point and empirical covering time were proposed. A decision control system for the roller shutter was also constructed based on a wireless sensor network. The results showed that the decisionmaking method and control system of uncovering/covering curtain can effectively reduce the incidence of low temperature by 43%. Compared with the empirical uncovering/covering curtain of the control greenhouse, the daily light duration of the experimental greenhouse was extended by 1.25h on average, and the accumulated light duration was increased by 75.16h, the product of thermal effectiveness and PAR was increased by 61.41MJ/m2, and the effective accumulated temperature was increased by 22.28℃, which effectively improved the heat and light duration of the greenhouse. The tomato plants in the test greenhouse were lower in height, which had thicker stems and performed more robustly, with significantly higher growth rates of leaf length and width than that in the control greenhouse. Statistical analysis of the first fruit yield, which was more influential during the trial, showed that the yield of the trial greenhouse was increased by 30.74%, and the first harvest was earlier. It was proved that the proposed method and control system for decision-making of curtain uncovering/covering can promote crop growth and effectively increase the material accumulation of tomato, which provided an idea for decision making of rolling shutter uncovering in heliostat.

Abstract:Sheep house humidity has the characteristics of large time delay, nonlinearity and spatial distribution difference, and the interaction mechanism with a variety of environmental parameters is complex and highly coupled. The humidity prediction model constructed by traditional prediction methods is difficult to meet the needs of largescale accurate breeding of mutton sheep. Too high or too low humidity of sheep house will directly threaten the healthy growth of sheep. Timely control of the trend of humidity and early regulation is the key to ensure the welfare of sheep. A nonlinear combined prediction model of sheep house humidity based on singular spectrum analysis (SSA), particle swarm optimization (PSO) and optimized long short-term memory network (LSTM) was proposed for accuracy humidity prediction. Firstly, the normal sequence and noise sequence were separated by SSA, and the original sequence was transformed into smooth sequence. Secondly, the optimal parameter combination of LSTM was determined through PSO iterative optimization to reduce the training cost of LSTM. Finally, a combined prediction model was established according to the optimized parameters to predict the two sequences respectively, and the sum of the model results was the final prediction result. The model was used to predict the air humidity in sheep houses in Xinjiang Uygur Autonomous Region from March 17, 2021 to March 27, 2021. The results showed that the combined prediction model had good generalization, stability and convergence. Compared with the standard ELM, SVR, LSTM, PSO-LSTM,EMD-PSO-LSTM and other models, the proposed SSA-PS-LSTM combined model had higher prediction accuracy. Its mean square error (MSE), mean absolute error (MAE) and determination coefficient (R2) were 1.127%2, 0.803% and 0.988, respectively. The experimental results showed that the established model had better prediction performance, which can provide important decisions for formulating optimized sheep house environmental control strategy, solving the lag problem of environmental control effect, and it made a strong support for the healthy growth of sheep.

Abstract:Meat and bone meal is the main food transmission chain of Bovine Spongiform Encephalopathy. The research on the detection method of meat and bone meal in feed has been one of the key tasks in the field of feed safety in the world. In view of the differences in the microscopic lacunae of bone powders from aquatic and terrestrial animals, a method of Micro-CT in-situ three-dimensional visualization of the lacunae of bone particles of different species was proposed. Taking fish, bovine, porcine and chicken bone particles as research objects, Bruker Skyscan 1275 Micro-CT was used to optimize scanning and reconstruction conditions of each sample on the basis of scanning condition optimization. Global, Otsu, K-means clustering, Adaptive mean-C and Adaptive median-C were compared for the segmentation effect of bone particle microscopic lacunar imaging. The optimized Micro-CT scanning conditions were as follows: voltage 80kV, current 125μA, image resolution 8μm, rotation step 0.20°, exposure time 46 ms and scanning 360°. Smoothing 1, Post-Alignment 0, Ring-Artifacts 3, and Beam-hardening 30% were used for three-dimensional reconstruction. The local adaptive mean method had the best imaging effect of bone particles, and its optimal parameters WS was 5, C was 0. The results of Micro-CT in-situ three-dimensional visualization of the lacunae structure of bone particles from different animal sources were in good agreement with the results of microscope standard method. The results provided a method for rapid and nondestructive discriminant analysis of aquatic and terrestrial animals feed.

Abstract:The corn production is high in China, the high efficiency, portable and low cost corn component detection technology and its devices are very important for the detection of corn quality. A portable-corn quality detection device was designed based on visible/near infrared spectroscopy technology. In order to explore the feasibility of the designed solution, a visible/near infrared spectrum acquisition system was built, and the spectra of 72 corn samples of different varieties were collected. The partial least squares prediction model of protein, fat and starch contents in corn grains and the CARS-PLS prediction model combined with competitive adaptive reweighted sampling were established respectively. The results showed that CARS method could effectively screen out the correlation variables of each component and improve the model effect. The root mean square error of prediction set (RMSEP) was decreased, and the RMSEP of protein was from 0.4866% to 0.4068%. The RMSEP of fat was decreased from 0.1549% to 0.0989%;and the RMSEP of starch was decreased from 0.4714% to 0.4675%. The correlation coefficient Rp of prediction set was improved. The Rp of protein was increased from 0.9309 to 0.9603. The Rp of fat was increased from 0.9497 to 0.9770. The Rp of starch was increased from 0.9520 to 0.9605. According to the characteristic variables of each component screened by CARS method, a suitable near infrared spectroscopy sensor was selected. On this basis, the spectral acquisition unit, control unit, display unit, power supply unit and heat dissipation unit of the detection device were designed. Based on NodeMCU development board and Arduino IDE development tool, the device control program was developed with Arduino language to achieve “one-click” rapid detection. The detection accuracy and stability of the device were verified by experiments. The results showed that the correlation coefficients of protein, fat and starch contents were 0.8431, 0.8243 and 0.8154, respectively, and the root mean square error of prediction were 0.3576%, 0.2318% and 0.2333%, respectively, and the relative analysis errors were 1.8577, 1.7761 and 1.5735, respectively. When the same sample was repeatedly predicted, the coefficient of variation of each component was 0.235%, 0.241% and 0.028%, respectively.

Abstract:The changes of four edible quality indicators, including moisture content, fatty acid content, eating value and straight-chain amylose content, were measured under three storage conditions: low temperature (8℃), quasi-low temperature (15℃) and room temperature (20~25℃), using South Japonica 9108 produced in Dongtai, Jiangsu Province as the object of study. The results showed that before storage, the rice and brown rice were more stable than the others, and the high moisture content samples had better eating quality than the normal and low moisture content samples. The initial moisture content, storage temperature and time had a significant effect on the fatty acid content, flavour value and straight chain amylose content (P<0.01). At different storage temperatures, the eating quality of both rice and brown rice at different moisture contents was gradually decreased with increasing storage time. The higher moisture content samples showed a higher reduction in quality at higher storage temperatures, and the brown rice showed a higher reduction in eating quality than the rice. The results also showed that at low moisture content (12.1%), there was little change in the eating quality of both rice and brown rice at different storage temperatures, suggesting that storing rice in the form of paddy at a lower moisture content is clearly beneficial to the maintenance of eating quality.

Abstract:Compared with the artificial sensory evaluation method, the tea bud grading based on deep learning and computer technology can reduce the time cost and greatly improve the accuracy, but the commonly used recognition model has the problem of large redundant calculation and large model specifications. For this reason, the tea buds picked from the Hongfeng Mountain Yun Tea Farm in Guizhou were used as the research object, and the tea samples were divided into three grades based on the workers’ experience. The multiscale convolutional block attention module (MCBAM) and multiscale depth shortcut (MDS) were embedded in the ShuffleNet-V2 0.5x basic unit, a tea bud grading model (ShuffleNet-V2 0.5x-SMAU) was proposed, which focused on the feature information in tea samples that was conducive to grading. The models pre-trained on two different source domains was taken as the teacher and student model. A tea bud grading method was proposed which combined dual migration and knowledge distillation. With the help of dark knowledge, the classification performance of the grading model and the ability to resist over-fitting were further enhanced. The results showed that the classification accuracy of the method can achieve 100%, 92.70% and 89.89% respectively for the three different grade samples in the test set under the condition of ensuring the lightweight of the model, which was better than the comprehensive performance of the complex network model. The application was more advantageous in production scenarios with limited storage resources and low hardware levels.

Abstract:Due to the complex and changeable operating environment of the highpower tractor, the densely arranged gears of the power shift transmission (PST) or field random load fluctuations could easily lead to frequent and random shifts, destroying the stability of the working conditions and affecting the performance and operating quality. Therefore, to solve the problem caused by random load fluctuations during tractor operation, a shift strategy by the adaptive correction of random loads was proposed. Firstly, the throttle opening, slip rate and vehicle speed were chosen as the control parameters to calculate the theoretical shift schedule of PST. Then, three correction parameters such as random load variation coefficient, throttle opening change and steadystate load change were introduced to calculate the shift correction under random load fluctuations through fuzzy rules. Combined with the convergent shift delay strategy, the light load or the transportation condition adopted the theoretical downshift law, and the heavy-duty operation adopted the adaptive upshift delay, the larger the load fluctuation was, the larger the upshift delay was. In addition, a high-power tractor hardware-in-the-loop simulation platform for the test of the gearbox control system was designed and developed. The tractor transmission system model based on AMESim and Simulink was used to test the performance of the transmission control system under different operating conditions. The simulation results showed that under road transportation and plowing conditions, the shift times of the proposed shift strategy were 63.16% and 45% lower than that of the traditional shift strategy, respectively, and the fuel consumption of the proposed shift strategy were 0.76kg and 0.47kg, which were 0.51% and 1.03% lower than that of the traditional control strategy, respectively. The strategy could effectively suppress frequent random shifts caused by random load fluctuations and balance the power and fuel economy of the tractor.

Abstract:The 3-PUS-PU compliant parallel mechanism with high precision and high stiffness was proposed. Its kinematic analysis and optimization design were carried out. According to the structure of the flexible hinges and the geometric relations of mechanisms, the motion principle of 3-PUS-PU compliant parallel mechanism was described, and the motion characteristics of the mechanism were analyzed. The kinematics model of the mechanism was established by the closedloop vector method, and the input-output relationship and velocity characteristics of the mechanism were analyzed. Then, the limit angle of flexure hinges was analyzed based on the theory of elasticity. According to constraint conditions and geometric parameters, the workspace of the mechanism was obtained by the numerical search method. According to the work requirements of the mechanism, the performances of positioning accuracy and compact were proposed. Then, the influence of structure parameters of the mechanism on the performances were analyzed. Finally, the particle swarm optimization algorithm was adopted to optimize the size parameters of the mechanism. Compared with the original mechanism, the optimal mechanism was improved by 25% and 81% for positioning accuracy and compactness, respectively. The results showed that the optimal mechanism performance of the kinematics was good. It can provide guidance for the subsequent prototype development.

Abstract:According to the topology design method of parallel mechanism based on sub-kinematic-chain generating sub-workspace and its superposition principle, a pure three-translation parallel mechanism with symbolic positive solution and motion decoupling was designed. Through the topology structural analysis of the mechanism, it was found that the mechanism was composed of two sub-kinematic-chains, and the coupling degree of the mechanism was 0. Then, using the mechanism position analysis method based on topological characteristics, the forward and inverse position solutions of the mechanism were solved and verified. And it was found that the mechanism had the characteristics of partial motion decoupling and symbolic positive solution. Finally, the singular configuration of the parallel mechanism was analyzed by using the singular analysis method based on the sub-kinematic-chain element, and the regular and non-singular rectangular workspace of the parallel mechanism was solved. Using this singularity analysis method, all singular configurations in the sub-kinematic-chain can be found, and it was conducive to solving the regular workspace without singularity. The research result can provide a theoretical basis and specific method for the design of a class of parallel mechanisms with regular workspace, and it played a guiding role in the design of three-translation parallel mechanism based on the superposition principle of sub-workspace generated by sub-kinematic-chain, and provided an idea and method for singularity and workspace analysis of complex parallel mechanism with multiple sub-kinematic-chains.

Abstract:The soft manipulator has its unique advantages in picking some fragile objects，which has been the focus of research field of robot. In order to evaluate the overall performance of the soft manipulator and further realize the precise grasping control of manipulator, it is necessary to carry out modeling analysis and experimental research on the soft manipulator. A four-finger soft manipulator integrated with flexible strain sensor that can illustrate the bending angle of soft actuator was developed. Moreover, the mathematical model for predicting the bending angle and end output force of the soft actuator upon various pressure was established, and the influence of variable stiffness for constraint layer on output performance of soft manipulator was analyzed as well. Furthermore, hardware control system of soft manipulator was developed and experimented on the bending angle, and output force of the soft pneumatic actuator were performed. Experimental results showed that the force output of soft manipulator can be improved by changing the stuffiness of constraint layer. In addition, the experimental data agreed well with the theoretical analysis results, which validated the correctness of the proposed mathematical model. Grasping tests on several commonly fruits of various shapes and length such as strawberry, orange, pear and apple demonstrated that the soft manipulator can achieve nondestructive grasping of fragile and fragile objects easily. The envelope grasping force of the soft manipulator was up to 11.89N, and the fingertip grasping force was 2.81N. The research results can provide theoretical guidance and reference for the widely application of soft manipulator.

Abstract:The decline of motion accuracy of CNC machine tools is a dynamic evolution process. To detect the potential failure risk of CNC machine tools as early as possible, the motion accuracy’s deterioration information contained in various monitoring data sequences was mined. Based on the difference and complementarity of multisource monitoring big data, a prediction method for motion accuracy’s deterioration of CNC machine tools was proposed by combining the deep gated recurrent unit (GRU) and attention mechanism. In order to overcome the defect that the traditional deep convolution neural network cannot learn the time series feature, the deep learning modeling method of motion accuracy based on deep GRU was proposed by using deep encoder-decoder structure. By datadriven, the temporal and spatial characteristics of motion accuracy and state signal time series were automatically mined to predict the change curve of motion accuracy and the deterioration trend of accuracy. At the same time, in order to enhance the information expression of main state signals and key time points, and improve the accuracy of accuracy deterioration prediction, a method of integrating attention mechanism in deep learning network was proposed. The method can establish the attention network of state parameter, calculate the correlation degree between vibration, temperature and other status signals and machine tools’ accuracy, and automatically adjust the weight of each signal. Furthermore, through establishing timeseries attention network to select the key time points of historical information of accuracy deterioration, the accuracy of longterm prediction was improved. The experimental results showed that the prediction model based on deep learning network and attention mechanism can well track the deterioration trend and law of CNC machine tools’ motion accuracy, and it had high prediction accuracy than traditional methods.

Abstract:To reduce the energy consumption and improve the energy utilization efficiency of the hydraulic power source system under the condition of changing loads, it can realize flexible and changeable output flow, pressure and power control. A threephase asynchronous motor was controlled under the variable frequency V/F control mode, a constant pressure pump was driven as a hydraulic power source. It was a high-energy electro-hydraulic power source that realized the compound control of pressure, flow and power.In view of the slow start of frequency conversion,a controllable accumulator was connected in parallel with the main circuit to solve the problem of slow start-up characteristics of an AC asynchronous motor. Firstly, the mathematical models of constant pressure pump, frequency converter and motor were established by using mechanical equations. Secondly, the system simulation model was established by AMESim software on this basis. And a test system platform was built, through the test, the constant pressure characteristic simulation was consistent with the test, confirming that the simulation model was accurate. The pressure and flow characteristics of the power source of the constant pressure pump driven by the variable frequency asynchronous motor were further simulated and tested. The accuracy of the simulation model was verified by comparing the pressure characteristic test with the simulation model. Finally, the simulation model was used to find the parameters of the auxiliary control switch accumulator, and three sets of data were determined for simulation analysis and P-Q and energy consumption tests. The P-Q test showed that the dynamic response time of the pressure controlled by the pilot pressure valve was not more than 0.2s, and the overshoot was not more than 15%;the dynamic characteristics of the flow were poor, and the accumulator was used to assist the start. The time did not exceed 0.2s. The energy consumption test showed that under the condition of high pressure and small flow and non-duty cycle pressure unloading, the motor speed was reduced from 1500r/min to 450r/min, and the motor power was reduced by 70.3% and 64.8%, respectively. The low speed can reduce the energy consumption of the test system by about 3.8kW.