Abstract:Strawberry is one of the world’s most widely distributed fruit plants, but its cultivation management relies heavily on a large number of labors and high-intensity operations, so the mechanization of strawberry cultivation has become one of the most important fields all over the world. The research status of strawberry production mechanization technology and equipment all over the world was introduced. In view of the differences of strawberry planting pattern, planting scale and industrial characteristics of China, United States, Europe and Japan, their differences in the development of strawberry production machinery were compared, and their respective weakness of strawberry mechanization was pointed out. And then, aiming at the realistic foundation of facilities strawberry cultivation in China, such as smaller planting scale, larger proportion of fresh-eating, diversified cultivation modes and technical levels, the development goals of strawberry planting equipment and technology in China, including small-type, low-cost, multi-functional, all-electrified, easy-operated and intelligent, were put forward. Finally, the development trend of deep integration of agronomy-machinery, informationization and leisurization of it was prospected.

Abstract:Winter jujube fruit recognition is the key technology to realize automatic picking, fruit trees precision management and yield forecast in winter jujube orchard. The rapid and accurate recognition of winter jujube fruits in natural scene affects real-time operability of automatic picking and reliability of monitoring and prediction directly. According to the complex recognition conditions, such as dark light, backlighting, occlusion, and dense fruits in winter jujube orchards, YOLO v3-SE model embedded in SE Net was proposed based on YOLO v3. SE Net adaptively recalibrated channel-wise feature responses by explicitly modelling interdependencies between channels. It strengthened important and valid features, and weakened unimportant and invalid features to improve the performance of feature maps. The deep convolutional neural network built in the article was TensorFlow. After the YOLO v3-SE model was trained and its recognition effect was tested on test samples, and 0.55 was selected as the optimal confidence threshold for the final detection. The P, R, F and mAP were used to assess the differences between YOLO v3-SE and YOLO v3 models. Test results showed that the model proposed got significantly good results. The detection results had the P value of 88.71%, R value of 83.80%, F value of 86.19%, and mAP value of 82.01%. Compared with the results of YOLO v3, the F value and mAP value had an increase of 2.38 percentage points and 4.78 percentage points. Meanwhile, there was no significant difference in detection speed. The further experiments compared the test results of the proposed model and YOLO v3 in complex conditions. In the data sets of backlight and dark-light fruit, the F value and mAP value of the proposed model reached 83.10% and 76.58%. In the data sets of occlusion and dense fruit, the F value and mAP value of the proposed model were 85.02 % and 74.78%. In the data sets of white-ripe, crisp-ripe and full-ripe stage fruit, the F value and mAP value of the proposed model were 86.37%, 89.91%, 91.49%, and 81.18%, 85.15%, 87.49%, respectively. Compared with the original YOLO v3, the F value was increased by 1.75~2.77 percentage points, and the mAP value was increased by 2.38~4.81 percentage points. The detection performance was significantly improved. The above content verified the effectivity of the YOLO v3-SE model. The model proposed can provide a method for winter jujube automatic picking and orchard yield forecast.

Abstract:In order to solve the problems of low recognition rate and poor robustness of cotton seedlings and cross-growth of various weeds distribution status, seven kinds of common weeds in the field were taken as the research object under natural conditions of Xinjiang cotton seedling period. A Faster R-CNN method of growing cotton seedling weed identification with data augmentation was proposed. A total of 4694 images of weeds in cotton seedling stage under different growing backgrounds and different weather conditions were collected, then the objects of images were annotated and the data sets were augmented. The suitable anchor scale of the model was designed, and four feature extractors involving VGG16, VGG19, ResNet50 and ResNet101 were compared. VGG16 was selected as the optimal feature extractor to train cotton seedling and weeds images and optimized Faster R-CNN network detection model was obtained for weeds of different weather conditions and the variety growth status, which can effectively identify and localize seven types of weeds and cotton seedlings. The average identification time for single picture was 0.261s and the average precision of optimized Faster R-CNN was 94.21%. With the same sample, characteristic extractor network, computer condition, the proposed method was compared with the state-of-the-art methods YOLO and SSD algorithms. The results showed that the proposed Faster R-CNN model had obvious advantages in the identification of various weeds in the seedling stage of cotton field. The trained model was placed in field environment for verification test. During the recognition process, totally 150 valid images were verified, and the average recognition rate reached 88.67%. The average recognition time for each image was 0.385s. The result indicated that the proposed method had certain applicability and generalization in precise control of weeds.

Abstract:The current seedling picking device used for vegetable transplanter was restricted by the working principle and mechanical structure, the frequency of picking seedling was mostly 40~90 plants/min per row, seedling picking had become a bottleneck restricting the development of technology and equipment for high-speed vegetable transplanting (≥90 plants/min). In order to realize the high-speed seedling picking operation, a pneumatic punching high-speed seedling picking device and supporting combined plug tray were designed. Through the operation process of “supplying tray orderly, conveying tray continuously, punching seedlings pneumatically and feeding seedlings freely”, the high-speed seedling picking operation at the frequency of 120 plants/min was realized. The mechanics model of seedlings in the process of punching was constructed, the layout of the pneumatic punching seedling monomer, the working pressure of the seedling picking cylinder, and the movement trajectory of the seedling ejector were analyzed and calculated. The structure and size of the U-shaped end of the ejector were optimized, the high-speed seedling picking sequence control system was designed and built. Taking 60d seedling-age pepper seedlings as the experimental object, under the conditions of cylinder working pressure of 0.26MPa and frequency of picking seedlings at 120 plants/min, seedling picking experiments were carried out by taking seedling picking success rate, matrix damage rate and stem and leaf damage rate as evaluation indicators of seedling picking effect. The result of experiments showed that the average success rate of seedling picking was 100%, the average substrate breaking rate was 22.46%, and the average stem and leaf damage rate was 3.54%, which met the requirements of high-speed seedling picking operation.

Abstract:In view of the problems of high labor intensity and high operation cost caused by the manually laying trays one by one in the field, a symmetrical arrangement of combination double cams controlled automatic tray laying mechanism was presented. The working principle of this mechanism was analyzed, to realize four actions of tray, chuck, separation and reset of stacked rice trays in order, and finally achieve the laying effect one by one in order. The parameter calculation model of the mechanism was established, and the auxiliary analysis and design software was written, and a set of optimal mechanism parameters were optimized as follows: base circle radius of the two cams were 15mm and 16mm, key length parameters were as follows: the extended length of the claw was 15mm, the width of the plate was 40mm, and the cam rotation center distance from the sheet metal shell wall was 30mm. The prototype was developed and the orthogonal experiment design method was used for experimental research. It was found that when the motor speed of the laying mechanism was 600r/min, totally 10 trays were put into the rice tray box in each pile, the trays were laid at a height of about 100mm from the ground, the passing rate of the total trays reached 97.16%, and the efficiency of the trays laying was 360 trays per hour. The mechanism can be loaded into the field automatic tray-laying equipment to replace manual plate-laying one by one and reduce the intensity of manual tray-laying. It can also be integrated into the seeding line as an automatic tray feed device.

Abstract:In the middle and lower reaches of the Yangtze River, the rice and rapeseed are the most important crops. They are planted in turn, and after the rice is harvested, rapeseed will be planted in time. High-efficiency combined precision seeder is often used to plant the rapeseed, which could complete seedbed preparation, ditching, fertilizing and sowing at once. However, here is a difficulty left by the previous rice harvest that is the harvest machine threshes the grain and leaves the rice straw in the field. The straw is easy to entangle the soil-touching parts of the combined precision seeder to block the machine, as well as the seeds are easy to fall on the straw, and difficult to emerge. Considering the kind of agronomic practice of straws covering for rapeseed cultivation, an innovation mechanized operation scheme both suitable for rapeseed seeding and returning rice straw mulch to the seedbed was proposed. A straw gather-mulch device matching with combined precision rapeseed seeder was designed. Through theoretical analysis, the structural parameters, installation position and angle, and working speed range of the key components of the straw mulching device were determined. Carrying the performance tests, controlling the straw feed rate of 0.9kg/s, 1.1kg/s and 1.3kg/s respectively, the feasibility of the working speed of each component was determined by theoretical analysis and the stability of straw transportation were verified. The experiment showed when the seeding straw mulching operation unit was equipped with a tractor of 69.9kW, at the operation speed of 0.7m/s, the roller speed of the pickup device was 80r/min, the speed of auger and chain lifting device of straw collecting device was 270r/min, the planter and components worked steadily, and the straw picking rate could reach 90%. Controlling the speed of the uniform placement device to be 210r/min, 240r/min, 270r/min, 300r/min and 330r/min, the experiment showed when the rotational speed was 300r/min, the uniformity rate of straw covering was over 92%. The field experiment showed that the straw had gathered clearly and mulched appropriately, the working parts of each link were stable, and the design indexes all met the technical requirements. The research showed that the designed straw gather-mulch device was integrated with the combined precision rapeseed seeder, which could complete the rice straw picking, stacking, transporting, rotating tillage, furrow opening, fertilizing, rapeseed seeding and mulching of rice straw in one operation, and it was suitable for the operation of rice stubble field without any treatment of straw after harvest.

Abstract:To solve the problems of poor land leveling and nonuniform axial distribution of soil in rotary tillage after opening drainage ditching in the middle and lower reaches of the Yangtze River, a blade roller for rotary tillage and axial land leveling with gradual spiral angle was designed. The volume parameter equation of soil disturbance of rotary tillage blade roller was established, the mechanical conditions of axial soil transport of rotary tillage blade were analyzed, and the spiral line equation of rotary tillage blade arrangement with gradient spiral angle was established. The key factors affecting the axial uniformity of soil layer were determined as blade roller rotary speed, cutting pitch of rotary tillage and initial spiral angle. The discrete element simulation model of rotary tillage blade roller with gradual spiral angle was established by using distinct element method. The regression equation of soil layer levelness was established by orthogonal test with the blade roller rotary speed, cutting pitch of rotary tillage and initial spiral angle as the test parameters. The Design-Expert analysis software was used, when the blade roller rotary speed, cutting pitch of rotary tillage and initial spiral angle were 260r/min, 8.3cm, and 71°, respectively, the optimized soil layer levelness was 17.35mm. The field test under the optimized parameters showed that the soil layer levelness, soil distribution uniformity, tillage depth stability and soil broken rate were 14.5mm, 8.82%, 92.34% and 81.66%, respectively, and the comprehensive tillage performance was better than that of the traditional rotary tillage blade roller. The research result could provide a reference for the equipment optimization and upgrading of the combined operation of wheat planting, drainage ditching and rotary tillage in agricultural areas of the middle and lower reaches of the Yangtze River.

Abstract:In order to solve the problems of poor soil throwing performance, high power consumption and no special blade, a cutting and throwing combined ditching blade was designed for tea garden ditching. The dynamic analysis of the soil throwing process of the soil throwing lump determined that the key parameters of the soil throwing lump were 8cm wide and the tilt angle was 30°, and the soil throwing blade could meet the requirements of tea garden ditching that the soil could be evenly distributed to the edge of the ditch instead of falling into the ditch, which was convenient for filling soil after fertilization and the ditching energy consumption could not be too high. The results of experiment showed that the width of single side throwing soil by the combined blade was 22.7cm, the uniformity of throwing soil was 90.3%, the thickness of covering soil was 2.1cm, the stability of ditching depth was 87.8%, and the thickness of floating soil at the bottom of ditch was 1.2cm, while that of the control group was 13.4cm, 84.3%, 2.4cm, 82.3% and 2.5cm, respectively, which indicated that the throwing distance of the combined blade was better than that of the single blade, the same as uniformity of throwing soil and planting stability, and the thrown soil by the combined blade was not easy to fall into the ditch. In addition, the combined blade did not significantly increase the power consumption. The results of power consumption analysis showed that compared with the single cutter, the impact force of the front and rear area of the combined blade was reduced, and the impact force of the upper area of the throwing soil was reduced, so the total impact force of the throwing soil was reduced.

Abstract:Clarifying the impact law of pesticide droplets on the surface of plants is of great significance for improving the efficiency of pesticide spraying and pest control. A numerical model of the horizontal liquid moving liquid film with single droplet impinging was established by the method of coupled level set and volume of fluid. The motion characteristics and dynamic mechanism of the liquid droplet impacting the moving liquid film was studied, by analyzing the pressure distribution, velocity distribution and vortices cloud map inside the liquid after impact, the mechanism of kinematic discontinuity, jet flow formation and jet top-end splash were verified, and the asymmetric crown formation mechanism was revealed, it had guiding significance to explore the effective deposition and loss law of pesticide spraying. It can be seen that the flow of liquid film caused crown showed obvious asymmetric characteristics, such as the development of the upstream and downstream jets flow were inconsistent, uneven crown base thickness and different splashing degree at the end of jet flow on both sides, the crown base migrated with the flow of the liquid film, under the influence of the flow inertia of the liquid film, when the dimensionless velocity U was 0.8 and the dimensionless time T was 3.47, the crown base completely migrated to the right of the impact point;the neck pressure difference mechanism led to the formation of jet flow, and the development of the jet flow was determined by the radial movement of the droplet and the vortex at the end of the jet flow, with the increase of liquid film velocity(U is 0～0.8), the jet flow in the upstream grown rapidly along the horizontal direction, while the jet flow in the downstream tended to extend vertically upward, and the end velocity of the jet flow on both sides was increased;the velocity difference between the radial movement velocity and spreading velocity determined the splashing state at the end of the jet flow, the flow direction of the upstream liquid film was opposite to the spreading direction of the droplet, therefore the splash behavior at the upstream end was more significant than that at downstream.

Abstract:Aiming at the problems of long development cycle and low design efficiency in the process of agricultural machinery product design and development, an intelligent design and optimization system for cleaning device of rice and wheat combine harvester was constructed. The system was composed of user demand module, knowledge base and inference engine module, parametric modeling module and intelligent optimization module, which can realize the design and optimization of cleaning device. Firstly, the design knowledge base of cleaning device in SQL Server 2012 was established, and the inference engine of cleaning device design was studied. According to the user’s design requirements, combined with the established knowledge base and inference engine module of cleaning device design, the relevant design knowledge in the knowledge base was called, and the reference method based on case and rule was used for designing, so as to output the parameters of key parts of cleaning device. Secondly, NX was redeveloped in Visual Studio programming software, which combined the two development languages, C++ and KF (knowledge fusion). In this way, the parametric model library of cleaning device was built, and then some key parts of cleaning device could be built rapidly in this library. Thirdly, the orthogonal tests of the CFD-EDEM coupling simulation of the cleaning device were designed with the air inlet velocity of the cleaning device, the inclination angle of the upper air guide plate, the inclination angle of the lower air guide plate and the vibration screen frequency as optimization variables. The cleaning impurity rate and loss rate in the test process were calculated to evaluate the cleaning effect. Finally, based on the simulation data, PSO-SVR algorithm was used to construct the regression model of the optimization variables, the cleaning impurity rate and the cleaning loss rate. After that, the SPEA2 algorithm was used to realize the multi-objective optimization of the cleaning impurity rate and loss rate, and to obtain a set of Pareto non-inferior solution. The results showed that when the wind speed at the air inlet of the cleaning device was 6m/s, the frequency of the vibrating screen was 4.5Hz, the inclination angle of the upper air guide plate was 32° and the inclination angle of the lower air guide plate was 18°, the impurity content and loss rate of the corresponding cleaning device model were 1.077% and 0.97%, respectively. As a reference, the model design parameters of key parts of the cleaning device can be optimized, which provided an optimization scheme for the design process.

Abstract:In order to solve the problems of potato and stone separation in main potato production areas in northern China, especially in the area of combined harvest operation, which problem is the large workload for farmers to pick up and low cleaning efficiency. The airflow suspension technology was used and combined with shaking and layered separation, potato-stone separator based on airflow suspension technology was developed, and the cleaning characteristics under different parameters based on the separator was studied. The key parts of the potato-stone separator based on airflow suspension technology mainly included centrifugal fan, air pipe, material feeding device, impurity discharge device, shaking device and material falling position switching device. According to the different aerodynamic characteristics of different materials, the potato and stone could be separated automatically in the process of movement with the aid of high-speed airflow suspension and swing action of shaking sieve. The device give full play to the double advantages of airflow suspension technology and shaking separation. Field tests showed that, the impurity removal rate was 98.34%, the potato selection rate was 96.71%，and each performance index reached the industry standard when the wind speed was 35m/s, sieve angle was 18°and crank angular velocity was 30rad/s. This device could ensure the potato selection rate and the impurity removal rate, and the cleaning efficiency was improved, the production cost was reduced. It could provide a technical reference for further exploring the technology of potato harvester and potato-stone separator.

Abstract:The research on mechanized root cutting of garlic at home and abroad is still in the exploratory and experimental stage. Two key issues that remain unresolved in mechanized root cutting of garlic are low root cutting rate and high bulb bruise rate. In order to further study the operation mechanism and optimization design of garlic root floating cutting device, the theoretical study on the mechanical properties of root floating cutting process was carried out. The dynamics equations of bulb collision during root cutting were derived, and the relative velocity of bulb collision was the key motion parameter that affected the collision damage. The principle of root sliding cutting operation was analyzed, and the cause of root group cutting resistance was analyzed by establishing mechanical models of blade cutting resistance. The process of bulb collision, root group disturbance and fracture was analyzed by high-speed photography. The mathematical models were established to predict bruise rate and root cutting rate. The effects of various factors on bruise rate and root cutting rate were analyzed and the comprehensive optimization of various factors was carried out. The bruise rate was 2.78% and the root cutting rate was 93.17% under the optimal parameters of root floating cutting device with feed speed of 1m/s, rotational speed of the cutter of 2600r/min, cutter angle of 33°, and spiral grating pitch of 28mm. The research results can provide references for the research and mechanism design of garlic root cutting technique used in garlic combine harvester.

Abstract:In order to provide a special spreader for cow bedding materials, a cow bedding material hammer spreader was designed. Taking sand as the research object, an operating theoretical model of the hammer spreading mechanism was established. The gyration radius of the hammer, spreading speed, spreading angle of the material, number and size of hammer were designed and calculated. The simulation model of sand spreading was established by EDEM, and the key factors such as spreading speed, tangent angle of spreading, opening angle of upper baffle, and length of upper baffle were simulated and analyzed. The influence of each factor on spreading width, average thickness of materials accumulated on the ground, and spreading uniformity was analyzed. The simulation results showed that by increasing the spreading speed and the opening angle of the upper baffle, reducing the tangent angle of spreading and the length of upper baffle, the spreading width and spreading uniformity of the sand accumulated on the ground can be improved, and the average thickness of sand accumulated on the ground can be reduced. Under the condition that the spreading speed was 1200r/min, the spreading tangent angle was 35°, the opening angle of the upper baffle was 70°, the length of the upper baffle was 180mm, and the driving speed was 4km/h, simulations and tests were carried out. Simulation results were in good agreement with experimental data, and the reliability of the simulation analysis was confirmed. The effective spreading width was 5.1m, and the average thickness was 22mm, and the coefficient of variation of the uniformity of spreading was 31.96%, the requirement of farms was met.

Abstract:Biomass and crop water use efficiency (CWUE) are important indicators to reflect plant growth productivity and quality, and their accurate real-time acquisition is the guarantee to achieve accurate agricultural management. To assess the feasibility of unmanned aerial vehicle (UAV) remote sensing platform combined with water use efficiency growth models to estimate crop biomass and CWUE, the silage maize was employed as the research object. The key crop parameter transpiration coefficient (kt) estimated based on the multispectral image of the high-resolution space-time UAV was firstly inputted into two simple water efficiency models to fit the WUE and WP* of the silage maize under different water stress conditions, and then the biomass of silage maize under the same and different water conditions was estimated by the fitted WUE and WP* values. The results showed that the correlation between the biomass and ∑Tc,adj and ∑ktkswkst based on the multispectral UAV platform combined with meteorological and soil water content data reached extremely significant level (P<0.001). Under the different stress conditions, the lowest determinant coefficients of fitted WUE and WP* were 0.92 and 0.93, respectively. Under the same water stress condition, the accuracy of biomass estimation by using the fitted WUE and WP* values was almost the same, which was shown in the following aspects: in the V-R4 growth period of maize, the accuracy of biomass estimation based on the fitted WUE indicating with RMSE was 126g/m2, d was 0.98, the accuracy of biomass estimation based on the fitted WP* indicating with RMSE was 91.7g/m2, d was 0.98, but the accuracy was not high in the R5-R6 growth period. When WUE and WP* values were used to estimate biomass under different water stress conditions, WUE was susceptible to water stress with low accuracy (RMSE was 306g/m2, d was 0.93), while WP* had higher accuracy (RMSE was 195g/m2,d was 0.97). At the same time, the spatial distribution maps of WUE, WP* and biomass on the field scale were obtained based on the multispectral remote sensing image of UAV. Overall, the combination of UAV remote sensing platform and crop growth model can well estimate the field silage maize biomass and water use efficiency.

Abstract:Aiming at the problems of difficult cooperative control and low working efficiency of agricultural Agent groups, the task assignment of agricultural heterogeneous Agent groups was researched based on improved stimulus response model. A layered hybrid multi-Agent architecture based on acquaintance net and the cloud platform-edge server collaborative computing system was established. The stimulus response model of ant colony algorithm was applied to the traditional contract network algorithm, and the adaptive bidding strategy was established to limit the number of bidding Agents and reduce the communication burden of the system. Based on the heterogeneity of agricultural Agents, the efficiency model of task assignment was established, by constructing time-varying coefficient and time matrix, the dynamic trust function and response threshold design method based on direct trust and recommendation-based trust were established to optimize the overall efficiency of agricultural Agent groups. Through increment PID algorithm and integral separated threshold, the adaptive stimulus update function was established to reduce the number of iterations, which reduced the workload of the Agent team overshoot, traffic and the number of iterations when the deviance was converged. The simulation results showed that when the Agent team size was 40 and 100 respectively, the overall efficiency of the improved contract network algorithm was 41.1% and 83.1% higher than that of the traditional contract network algorithm. When the Agent team size was 40, three sets of stimulus update functions were set in addition. The workload overshoot of the stimulus update function based on PID algorithm was reduced by 24.5% and 9.5% respectively compared with the second group and the third group. In terms of iteration times, it was reduced by 84.2% and 84.8% compared with the first group and the third group. When the Agent team size was 20, 40 and 100 respectively, the traffic of the improved contract network algorithm was reduced by 49.1%, 63.7% and 72.4% compared with the traditional contract network algorithm. Experimental verification showed that the traffic and workload overshoot of task allocation by the improved contract net algorithm was reduced by 70.0% and 20.2% compared with the traditional contract net algorithm, the overall efficiency was increased by 14.1% compared with the traditional contract net algorithm, and improved task allocation algorithm could guarantee that the Agent groups at work could achieve full coverage of the work area within the prescribed time limits.

Abstract:Rice phenotype has an important guiding role in rice research by analyzing genetic information of various phenotype data. Knowledge graph technology has been widely used in knowledge storage and search engines by structurally describing the information, concepts, entities and relationships in data. As a key task in the knowledge graph, the relation extraction task can extract the connection between two entity words in the text. Within this research, rice phenotypic data was collected from the National Rice Data Center, and the data were preprocessed and annotated. The rice phenotype relationship was proposed based on the plant ontology, and then method of bidirectional encoder representation from transformers (BERT) was applied for classifying relation between rice genomics, environment, and phenotype data based on plant ontology. Then the word vector, position vector and sentence vector were extracted in the relation dataset, and rice phenotype relation extraction model was realized based on BERT. Finally, the results of BERT model was compared with the convolutional neural network and the piecewise convolutional network model. In the comparison of the three relationship extraction models, BERT achieved the best performance, and reached an accuracy of 95.11% and F1 value of 95.85%. Deep learning methods were used to improve the performance of relation extraction of knowledge graphs, which can provide technical support for the efficient construction of a rice phenotype knowledge graph system.

Abstract:Aiming at the problem that GPS, lidar, vision and other positioning and navigation methods for robot sometimes do not work well or even fail in the complex environment of banana plantation, a positioning and navigation method for robot based on laser and ultrasonic sensors at banana plantation was proposed. However, when there were small shrubs and other obstacles, or at rugged tractor road in the complex environment of banana plantation, it was the premise and the key for positioning and navigation to accurately measure the shortest distance between the robot and banana tree. Therefore, a distance measurement method for banana tree using laser and ultrasonic sensors based on fitting and filtering techniques was proposed. Firstly, the laser and ultrasonic sensors simultaneously measured the distance between the robot and banana tree at each sampling time and reciprocally checked the distance data ranged by laser and ultrasonic sensor in order to generate a set of distance data for the target banana tree;after that, the set of distance data was fitted using a quadratic polynomial by the least squares method;and then the distance data was filtered out with larger error from the set of distance data based on the fitted quadratic polynomial and the suitable threshold;and finally, the shortest distance from the robot to banana tree was got by averaging the three minimum distance data at the filtered set of distance data. The experimental results showed that the ranging method proposed had an maximum ranging error rate of 1.0% for banana tree in ideal environment, and the maximum ranging error rate was 2.0%, which corresponding maximum ranging error was 1.0cm, in an environment with small shrubs and other obstacles or rugged road, even in outside natural environment. Therefore, the stability of the ranging method proposed was good, which can provide accurate distance data for the robot to realize positioning and navigation in the banana plantation.

Abstract:The topography of the orchard is variable. The planting density of the fruit trees is large, and the shape of the crown is different. Therefore, it is difficult to recognize the crown of an individual fruit tree in a complex orchard background. A novel method of crown recognition and segmentation based on Mask R-CNN neural network model was studied. The image data of the citrus orchard was obtained through the camera, and the Mask R-CNN neural network was used to realize the recognition and segmentation of the crown of an individual citrus plant. The research results showed that the recognition accuracy of the individual tree crown of the orchard participating in the modeling was 97%, and the recognition time was 0.26s, which can basically meet the requirements of tree crown recognition in the process of precise orchard operation. The recognition accuracy of the single tree crown of the orchard not participating in the modeling was 89%, which showed that the model was suitable for different kinds and environments of orchards. Compared with the SegNet model, the accuracy of the used model was about 5 percentage points higher, indicating that it had a better recognition and segmentation effect in complex orchard images with more non-target tree crowns. Therefore, the recognition and segmentation method can achieve rapid and accurate recognition and segmentation of single tree crown, which provided an important basis for accurate orchard operations such as target spraying, pest protection, growth recognition and prediction.

Abstract:Canopy volume is an important tree measurement factor for trees, and the accurate measurement of its volume value plays an important role in the in-depth study of three-dimensional green quantity and regional carbon cycle. Aiming at the problems of overestimation and underestimation in the existing methods of measuring tree canopy volume from point cloud data, a α-shape boundary extraction method that took into account the point cloud boundary density and variable threshold was proposed. The length of the optimal linear-iterative step and the interval between layers of the method were determined through experimental analysis, so as to realize the accurate calculation of canopy volume. Firstly, the canopy point cloud data was sliced at equal intervals;then, an improved α-shape algorithm was used to extract the more realistic and natural boundary polygons of the point cloud slices;finally, the section area and the volume of the platform between each layer point cloud were calculated, and the canopy volume was obtained by adding the platform volumes. The experimental results showed that the accuracy of obtaining the canopy volume was related to the structure of branches and leaves inside the canopy and the point cloud density. Regardless of high-density or low-density canopies, the canopy volume value calculated by the improved α-shape algorithm not only had good stability, but also was more accurate than that of existing methods, which avoided the overestimation of Graham convex hull algorithm, and was more conducive to the calculation of the overall volume of the canopy compared with the volume element accumulation method.

Abstract:The leaf rolling index and thickness are important indexes of plant drought resistance. However, existing measuring methods of these two indicators are wasteful, inefficient and weak universality. To solve this problem,the minimum external rectangle method based on Graham algorithm was proposed to extract the value of leaf rolling index and the concave and convex point detection algorithm combined the corner detection was proposed to measure the leaf thickness. The algorithm used in the article had the following steps: firstly,the permanent sides were obtained by paraffin sectioning technique, and then connected the microscope and computer to obtain slice images. Secondly, a red grayscale method was proposed for the Cleistogenes songorica leaf anatomical structure image to enhance contrast and details, according to the color difference of red and blue between the foreground and the background. Then through subjective judgment (observing the change of processed image, and comparing the changes to find out which method was the best one), it was found out that images processed by grayscale methods had the shortcoming of details blurred and proposed method showed better performance in improving the quality of image and the details of the images than the other methods. Also by evaluation functions: average gradient(AG),contrast(C) and information entropy(E) were used for objectively evaluated the method used and the traditional ones. The average value of AG,C and E was got by processing 30 test images, and it turned out that the values of the method proposed was better than the other methods. On this basis, eliminated noise using morphological operation and using linear filtering to eliminate serrated boundaries and eventually segmented background and objective by the maximum interclass variance (Otsu). Through subjective judgment, it was found out that the segmented target area basically coincided with the original boundary of the target. Also by evaluation functions: false positive rate(FPR), false negative rate(FNR),and global segmentation accuracy(GSA) were used for objectively evaluating the method. The average value of FPR,FNR and GSA was got by processing 30 test images were 0.75%, 3.49% and 98.14%, respectively. Finally,according to the actual measurement mode of leaf rolling index, the longest distance between the two points on the Cleistogenes songorica leaf anatomical structure image was measured, the minimum external rectangle method based on Graham algorithm was used to extract the value of leaf rolling index of the objective. The measurement was compared with its mean value of interactive measurement by ToupTek Toupview software,the average relative error of 30 test images was 0.96% and the average time consumed was 4.87s by the measurement proposed, the speed was improved by 11 times. According to the actual measurement of leaf thickness was to measure the distance between the concave points and the concave points on the left and right boundary of the Cleistogenes songorica leaf anatomical structure image, it was also the distance between the convex points and the convex points,the proposed concave and convex point detection algorithm combined the corner detection algorithm and vector product method to eliminate the useless apexes to obtain the concave and convex points in accordance with the actual measurement. Then, the leaf thickness value was obtained by concave points matching and convex points matching.The measurement was compared with its mean value of interactive measurement by ToupTek Toupview software,the average relative error of 30 test images was 3.69% and the average time consumed was 4.92s by the measurement proposed,the speed was improved by 38 times. In conclusion, the algorithm was more suitable for background segmentation and object measurement of Cleistogenes songorica leaf slices and can also provide a reference for other plant leaf slice images.

Abstract:Depth images are increasingly used to detect the body condition of dairy cows to make breeding management decisions. The scoring method of individual dairy cows’ body condition based on deep learning can further improve the degree of automation of dairy cow body condition image analysis. In order to realize the accurate recognition of the individual body condition of dairy cows without contact, high accuracy and strong applicability based on depth images in the actual breeding environment, a body condition scoring method was proposed based on deep learning and point cloud convex hulling features. Firstly, the acquired back depth image of the cow was preprocessed, included target extraction, target rotation, and acquired hindquarters images to obtain a back point cloud of the cow, containing the main body condition information. And then the hindquarters point cloud was voxelized and the convex hull feature image was obtained. In order to represent the fat and thin degree of different cows, and finally build a variety of convolutional neural network classification models, accuracy rate and average F1 value was used to optimize the model to further improve the accuracy of individual body condition recognition of dairy cows. The test results showed that the EfficientNet network can effectively identify the body condition of cows with a BCS value in the range of 2.25~4.00. The image account of recognition accuracy errors of 0.25 and 0.50 were 98.6% and 99.31%, respectively. The average F1 value was 98% and 99%, and the average recognition rate was 3.441s/f. Compared with the MobileNet-V2, XceptionNet, and LeNet-5 network models, the above indicators of the proposed method were better. The method can realize the non-contact assessment of the individual body condition of dairy cows in the breeding farm, and had the characteristics of high accuracy, strong applicability, and low cost.

Abstract:Aiming at the problems of complex business entities, long information flow, low data utilization and low regulatory coverage in rice supply chain, a information supervision model of rice supply chain driven by blockchain was built and implemented systematically. Firstly, the information flow characteristics of rice supply chain were analyzed, and the main body and key information of each business link in the supply chain were sorted out and refined. On this basis, the supervision model of rice supply chain information driven by blockchain was constructed, the hierarchical encryption storage mode of private data and the customized business logic supervision intelligent contract were designed. Finally, based on the Hyperledger Fabric open source framework, the rice supply chain information supervision prototype system was constructed and implemented. In addition, taking a grain and oil enterprise in Changde City of Hunan Province as an example, the application case analysis was carried out. The results showed that the rice supply chain information supervision model and prototype system constructed can solve the problems of rice supply chain data privacy encryption, secure storage and authority management, realize supply chain information interconnection and effective supervision, which can provide a way for grain and oil quality safety supervision feasible and practical application solutions.

Abstract:In order to reasonably improve the efficiency of cultivated land use, strengthen the construction of ecological civilization and the sustainable use of resources on the basis of cultivated landsuitability, taking Yuanyang County as an example, the cultivated land use efficiency was divided into different regions by comprehensively considering the evaluation results of cultivated land use efficiency and suitability, and the measures for improving the cultivated land use efficiency in each region was proposed. The results showed that from the perspective of cultivated land use efficiency, the average value of cultivated land use comprehensive efficiency in Yuanyang County was 0.837, the overall level was not high, and there was a large space for improvement of pure technical efficiency and scale efficiency. According to the evaluation results of cultivated land suitability, the suitable arable area, basic suitable arable area, low suitable arable area and basin protection area accounted for 37.84%, 41.36%, 19.09% and 1.71% of the total cultivated land in Yuanyang County, respectively, and the arable ability was strong. The cultivated land use efficiency in Yuanyang County was divided into conservation area, promotion area, adjustment area and protection area by superposition of the evaluation results of cultivated land use efficiency and arable land suitability.From the perspective of sustainable utilization and ecological protection, combining with differentiated remediation measures, in the conservation area, maintaining high arable land, arable land use efficiency and extending the ecological and economic value of arable land were the key points.The promotion area had strong arable and low cultivated land use efficiency. It can maximize its effective use efficiency through large-scale operation.In the adjustment area, the suitable tillage was poor, and the utilization efficiency of cultivated land should not be improved. Ecological agriculture or facility hydroponic agriculture should be developed to improve the economic benefits of cultivated land and alleviate the pressure of cultivated land and ecological protection. The cultivated land in the protected area would be withdrawn gradually, and its wetland and ecological functions would be brought into full play. The research results can provide reference for the sustainable utilization of cultivated land resources.

Abstract:The ecosystem evaluation index system and evaluation model were constructed based on GIS and remote sensing, and the temporal and spatial changes of ecosystem service values in Changting County from 2005 to 2016 were evaluated under the impacts of different soil and water conservation measures. The results showed that the ecosystem service values of Changting County spatially presented a trend of high around and low in the middle from 2005 to 2016, and the ecosystem service values in the four stages were 10.108 billion yuan, 8.839 billion yuan, 10.914 billion yuan and 13.411 billion yuan, respectively. Atmospheric balance value accounted for an average of 35%, water conservation value accounted for 15%, soil conservation value accounted for 21%, nutrient recycling value accounted for 4%, and organic matter production value accounted for 25%. The values of various ecosystem service showed a trend of first decreasing and then increasing. Under different soil and water conservation measures, there were significant differences in temporal and spatial changes of ecosystem service values in Changting County. In the towns of Zhuotian, Sanzhou, Hetian and Cewu, the governance effect was good, and the ecosystem service values were increased significantly, but the vegetation coverage of mountainous areas tended to be decreased. It was suggested to implement the modes of grass-shrub-arbor mixed management and old pine rehabilitation in mountainous and hilly areas, and to build drainage ditches, retaining walls and sand basins in the broken bare area of collapsed hills. Besides, interplanting suitable plants under the forest, speeding up the transformation of sloping fields into terraces and implementing the system of division of labor and cooperation were also needed to be considered. In the process of daily production and construction, it was important to pay attention to afforestation, prevent soil erosion, and strictly implement the three-simultaneous system of design, construction and operation of soil and water conservation projects.

Abstract:The carbon balance of the farmland ecosystem depends on the fixed carbon content of crops and carbon emissions from soil heterotrophic respiration. In order to reveal the comprehensive effects of water and fertilizer levels on carbon balance in farmland ecosystems, a field experiment of summer maize/winter wheat rotation cycle in Yangling District of Shaanxi Province was conducted during 2015—2016. Three irrigation levels: high, medium and low (W1, W0.85 and W0.7 meant 90mm, 76.5mm and 63mm in summer maize season and 140mm, 119mm and 98mm in winter wheat season), four nitrogen fertilizer levels: high, medium, low and no nitrogen (N1, N0.85, N0.7 and N0 meant N 300kg/hm2, 255kg/hm2, 210kg/hm2 and 0kg/hm2 in summer maize season and 210kg/hm2, 178.5kg/hm2, 147kg/hm2 and 0kg/hm2 in winter wheat season) and four phosphate fertilizer levels: high, medium, low and no phosphate (P1, P0.85, P0.7 and P0 meant P2O5 90kg/hm2, 76.5kg/hm2, 63kg/hm2 and 0kg/hm2 in summer maize season and 150kg/hm2, 127.5kg/hm2, 105kg/hm2 and 0kg/hm2 in winter wheat season) were designed. CO2 flux measurements were performed, using the static opaque chamber and chromatography method, and aboveground biomass and yield were measured. The net primary productivity carbon fixation (CNEP) and soil microorganisms heterotrophic respiratory carbon release (CRm) were also calculated to evaluate net ecosystem productivity carbon fixation (CNEP). The results showed that the summer maize/winter wheat farmland ecosystem under different water and fertilizer treatments performed as carbon sink. The range of CNEP in summer maize and winter wheat seasons were 6805~7233kg/hm2 and 5842~6434kg/hm2, respectively. At high, medium and low fertilizer levels, with the increase of irrigation amount, the CNPP of summer maize/winter wheat rotation cycle was increased by 2.48%~ 5.96%, the CRm of soil was increased by 2.15%~15.20%, and the CNEP was increased by 1.16%~6.47%, respectively. Under the high, medium and low water supply levels, with the increase of fertilization amount, the CNPP of summer maize/winter wheat rotation cycle was increased by 2.95%~3.43%, the CRm of soil was increased by 5.23%~18.67%, and the CNEP was increased by 0.93%~2.79%, respectively. The increase of CNEP was negatively correlated with amount of irrigation. Under low water condition, compared with single application of nitrogen or phosphate fertilizer, combined application of nitrogen and phosphorus fertilizers, the CNEP of summer maize/winter wheat rotation cycle was increased by 4.86% and 7.34%, respectively. Variance analysis of the interaction of water and fertilizer was significant for the CNEP of summer maize/winter wheat rotation cycle. While water and fertilizer supply differed by 15% had a significant positive interaction with CNEP in winter wheat farmland. Combined application of nitrogen and phosphorus fertilizers and the coordinated supply of water and fertilizer were helpful to promote CNEP of summer maize/winter wheat ecosystem. Based on the principle of water and fertilizer saving, summer maize and winter wheat were beneficial to increase CNEP accumulation under W0.85N0.85P0.85 and W0.7N0.85P0.85 treatments, respectively.

Abstract:In order to clarify the spatial differentiation of cultivated land planting structure and its influencing factors at the micro-geographical scale, the typical grain-producing counties in Heilongjiang Province were selected as the research area, the 1km2 geographic grid was used as the research unit, and the regional statistical method, spatial association analysis method and geographic detector model were used to analyze the spatial differentiation characteristics of the three main crops of rice, corn and soybean, and reveal the influence mechanism of the spatial differentiation of regional cultivated land planting structure. The results showed that the planting structure in the region was mainly single soybean type, single corn type and corn-soybean type, and single rice type and soybean-corn-rice type were the secondary types. There were some differences in the spatial agglomeration characteristics of different crop planting structures. Rice was mainly HH type, and corn and soybean were mainly HH and LL types. Terrain were the dominant factor in the spatial differentiation of rice planting structures, and that of corn and soybean were climatic and soil conditions. Natural conditions were still the key influencing factors in the spatial differentiation of cultivated land planting structures. The spatial differentiation of cultivated land planting structure was the result of multi-factor interaction. The interaction of market location and traffic location with other factors greatly improved the spatial differentiation of corn planting structure. For rice, irrigation potential had the strongest impact on factor interaction. However, the effect of factor interaction on the spatial differentiation of soybean planting structure was not much improved. The research results were helpful to understand the formation mechanism and differences of the spatial differentiation of different food crops, and provide a scientific basis for the optimization and adjustment of cultivated land planting structure.

Abstract:In view of the extremely arid climate with high temperature and low humidity, it is easy to cause “burnt flower” and low fruit setting rate and photosynthetic rate decline in jujube trees in Southern Xinjiang. Through field experiments, a set of mist system for jujube orchard canopy was set up with variable mist time and mist times (M1: 3 times a day for 20min each time;M2: 2 times a day for 20min each time;M3: 2 times a day for 40min each time;CK as control, without mist). The effects of different treatments on the flower drop rate and fruit setting rate of jujube trees and the effect of improving the quality of red jujubes were compared, and the best treatment was selected. The result showed that compared with the control, the misting treatment can effectively increase the canopy humidity and lower the temperature, it can increase the air humidity by up to 158.7%, and reduce the temperature by up to 38.9%. The soil moisture content under the misting treatment was generally higher, the soil moisture content of M3 was the highest. Compared with CK, the soil moisture content was increased by 8.44%. Changes in the canopy environment would affect the photosynthetic rate of the jujube leaves and the fixed amount of CO2, the fixed amount of CO2 on the leaves of the jujube under the mist treatment was greater than that of the without mist treatment. Therefore, the accumulation of crop organic matter and the storage of its own energy after the mist was better, which was more conducive to the growth of the crop and the growth of fruits. Through mist, a suitable environment can be created for jujube trees, which had significant effects in reducing flower drop rate, increasing fruit setting rate and fruit quality. The flower drop rate of M1 and M2 treatments was 18.39% and 15.09% lower than that of CK. Mist treatment can indeed regulate the canopy environment, a suitable canopy environment can increase the photosynthetic rate of jujube leaves, promote flowering and fruit setting, and improve the quality of red jujubes. Through this experiment, it was found that three times of mist a day, 20min each time, was the best treatment to obtain high-yield and high-quality red dates.

Abstract:Since the end of the 1990s, large-scale ecological restoration projects have been implemented in the Loess Plateau, which significantly changed land use types and land cover. These projects enhanced ecosystem service functions and improved soil and water loss to a certain extent. More and more attention has been paid to the impact of these projects on the hydrological processes. The Yanhe River Basin, with a semihumid climate in the middle of the Loess Plateau, was taken as an example. The community land model (CLM) version 5.0 was used to simulate the impact of land cover changes on key hydrological elements and their relationships in this basin over 1980 through 2018. The results showed that the cultivated area was decreased from 3277km2 in 1980 to 2400km2 in 2018, a decrease of 26.76%. Woodland and grassland area was increased by 244km2 and 528km2, respectively. The annual runoff and annual runoff coefficient was decreased significantly due to the change of land cover, which were 12.34% and 11.46%, respectively. The change of land cover also led to the decrease of monthly mean runoff and the coefficient, which had a less influence in the wet season and a larger influence in the dry season, making the annual runoff distribution to be centralized. To guide the rational use of the water resource and build a harmonious ecological environment in the Loess Plateau, understanding the hydrological process and its changing had a great theoretical value and practical significance.

Abstract:Corn straw was used as raw material, and different concentrations (10% and 30%) of HNO3 were used to modify the samples before/after hydrothermal carbonization. Pb2+ adsorption mechanism of hydrochars modified by HNO3 was studied by the results of the structure characterization, isotherm adsorption model and adsorption kinetics model. The results showed that hydrochars modified by HNO3 could form rich oxygen-containing groups. Straw carbon modified by HNO3 before hydrothermal treatment (10%N-JG and 30%N-JG) presented rough porous surface morphology and developed mesoporous structure, forming three-dimensional disordered large-scale microcrystalline structure. Straw carbon modified by HNO3 after hydrothermal treatment (JG-10%N and JG-30%N) produced a large number of micropores with uniform distribution and similar size, forming three-dimensional ordered smallsize microcrystalline structure. 10%N-JG and 30%N-JG had the best Pb2+ adsorption efficiency in all samples, they could reach adsorption equilibrium at 3.5h and 3h, respectively, and their theoretical maximum adsorption capacities were 247.51mg/g and 280.09mg/g. Pb2+ adsorption of 10%N-JG and 30%N-JG was in accordance with pseudo first and second-order kinetic model and Freundlich isotherm adsorption model, indicating that physical diffusion and chemical adsorption were equally important in Pb2+ adsorption. It was found that hydrochars mainly depended on the chemical adsorption of oxygen-containing functional groups to remove Pb2+ from water. The well-developed mesoporous structure was more conducive to Pb2+ entering into the particles, which increased the capture probability of oxygen groups on the internal channels to Pb2+, and ensured the efficient adsorption of Pb2+ by hydrochars.

Abstract:In order to explore the feasibility of the straw multi-stage continuous cold roll forming method, a prototype was designed and tested. Orthogonal test method was used to test the four-factor three-level orthogonal test with the moisture content, crushing method, crushing length and feeding method of corn straw as the test factors, and the rebound rate, density and firmness of the formed blocks as the test indicators. The influence of experimental factors on experimental indicators was researched. The test results showed that the cross-riveted feeding method can most restrict the rebound of the shaped block;the density of the shaped block was mainly affected by the crushing method, followed by the length of the crushing;the solidity of the shaped block was mainly affected by the crushing method, and the moisture content was the second of it. Comprehensive analysis determined that the preferred forming parameters of the molding machine were: moisture content of 20%, crushing method by rubbing and crushing, crushing length of 80mm, feeding method by cross riveting. The experiment was carried out under the conditions of better molding parameters, and the performance indicators of the molding machine were tested. The results showed that the rebound rate of corn straw molding block produced by the molding machine was 7.26%, the molding density was 363.28kg/m3, and the firmness was 90.23%. The research results provided a reference for the research and development of biomass compression molding technology and equipment at normal temperature.

Abstract:Indoor farming system has played an important role in precision agriculture. In recent years, many innovative greenhouses have been built in China. In order to enhance the thermal insulation of multi-span greenhouses in northern region, external thermal screen has been installed on multi-span greenhouses (Shouguang-type intelligent greenhouses). The external thermal screen and its power mechanism were installed above a large gutter. However, the larger gutter caused the lower light intensity in the greenhouse. In order to analyze the indoor light environment and optimize the gutter and greenhouse design, a dynamic model of light distribution inside the greenhouse was constructed. Based on the model, the uniformity and sensitivity of indoor light environment were analyzed. The results showed that the order of the influence on the average value of daily accumulated radiation in the cultivation area was the gutter distance, width, depth and height. Among different design combinations, there were multiple optimal ones of light distribution in the cultivation area. And the current gutter design of Shouguang type greenhouse was reasonable for uniform indoor light environment on local winter solstice day, including the distance (12.00m), and width (1.60m), depth (0.86m) and height (6.30m) of single gutter. According to the simulation results under different scenarios, the suitable ratios of the sum of gutter height and depth over the sum of gutter distance and width among all the suitable designs (the minimum variation coefficient of radiation intensity) were between 0.49 and 0.54. The dynamic model could provide a theoretical basis for the design and application of Shouguang-type intelligent greenhouses in different regions.

Abstract:Diseases and insect pests can seriously affect the yield and quality of greenhouse vegetables. Spraying chemical pesticides is currently the most effective method to control diseases and insect pests. However, the extensive use of chemical pesticides, on the one hand, leads to pathogenic bacteria and pests becoming resistant to chemical pesticides, which eventually makes prevention and control more difficult and, on the other hand, causes serious environmental pollution and leaves pesticide residues that are a threat to people’s health. Developing green control technology for diseases and insect pests is an important way to reduce the amounts of chemical pesticides used. A new type of ozone sterilizer device, named multi-functional plant protection machine, and an information management system was designed. The device included mainly ozone generation, circuit control, high-speed fan, trap lamp, environmental data acquisition, and other modules. All the modules were controlled by a STM32F103 microprocessor. The machine used a high-voltage discharge method to produce ozone that could be rapidly and evenly diffused into the whole facility space through the high-speed and large-volume fan and special air duct. At the required concentration, the ozone could oxidize and decompose eggs and larvae of insects, bacteria, and fungi, and inhibit virus reproduction. The device had yellow and blue lights to attract adult insects, which were then sucked into the machine by the fan and killed. The information management system had IoT-based architecture and included a mobile app. This web-based information management system interacted mainly with the data center to realize intelligent management of plant protection data. The device could be controlled remotely using the mobile APP for real-time adjustment of the fan’s wind speed, adjustment of ozone release, collection of environmental data, and opening of the trap lamp and other functions. The multi-functional plant protection machine was tested for its ability to control cucumber powdery mildew and Bemisia tabaci in several greenhouse facilities. The test results showed that after using the equipment during the whole growth period of the cucumber, the disease index of cucumber powdery mildew was controlled at 5.9, and the control effect on Bemisia tabaci reached 89.5%. The multifunctional plant protection machine effectively controlled cucumber powdery mildew and Bemisia tabaci in the greenhouses. The research result provided a reference for green control of diseases and insect pests in facilities that grow vegetables.

Abstract:Classification of rice processing precision is an important link in rice processing. In order to accurately identify the grade of rice processing precision, by combining the hyper column technology (HCT), max-relevance and min-redundancy (MRMR) feature selection algorithm and extreme learning machine (ELM) technique, an improved VGG16 convolutional neural network was proposed. First of all, the OneHot format in machine learning was used for coding and normalization of data;then, combining HCT, the VGG16 convolutional neural network was used as the feature extractor, which can extract local differentiating features from deep structure at different levels. Totally 5248 rice features were extracted, the MRMR feature selection algorithm was employed to eliminate massive redundant rice image features, and 500 most effective features were selected. Finally, the ELM technique was used to classify the processing grade of rice. The 5848 sample images were randomly divided into the training set, test set and verification set according to the ratio of 6∶3∶1 for training and test of model. The results showed that when the rice processing grade classification model built based on the improved VGG16 convolutional neural network was used to classify the 1755 rice samples in the test set, the overall accuracy can reach 97.32%, and the classification prediction speed of rice processing precision can reach approximately 85t/h, which basically satisfied the requirement of rice production line.

Abstract:In order to quickly and accurately detect the oil content of Camellia oleifera seed and solve the time-consuming and laborious problems of traditional detection methods, a method for detecting the oil content of Camellia oleifera seed based on hyperspectral imagery (HSI) was proposed. Two sets of hyperspectral imaging systems, spectral setⅠ (400~1000nm) and spectral setⅡ (900~1700nm), were used to collect diffuse reflectance hyperspectral images of Camellia oleifera seed, and the regression prediction model of oil content of Camellia oleifera seed was established in combination with chemometrics. The results showed that the partial least squares regression model (PLSR) established by the two sets of spectral data without pretreatment had the highest accuracy: the determination coefficient of prediction (R2p) of the spectral setⅠwas 0.681, and the root mean square error of prediction set (RMSEP) was 2.89%;R2p of spectral setⅡwas 0.740, and RMSEP was 2.92%. Comparing seven different variable selection methods, it was found that the two sets of spectral sets used genetic algorithm (GA) to filter the characteristic wavelength to establish the PLSR model with the highest accuracy: the spectral setⅠhad R2p of 0.694 and RMSEP of 2.82%;the spectral setⅡhad R2p of 0.779 and RMSEP of 2.54%. Comparing the modeling effects of spectral setⅠand spectral setⅡ, it was found that the performance of the PLSR model established by spectral setⅡwas better than that of the spectral setⅠ, so the band of 900~1700nm was more suitable for the oil content detection of Camellia oleifera seed than the band of 400~1000nm. Besides, the feasibility of using HSI to visualize the distribution of the predicted value of the oil content ofCamellia oleifera seed was further verified. This result can provide a method for the rapid detection of the oil content distribution of Camellia oleifera seed and the selection of high-quality its varieties.

Abstract:A heat pump low-temperature circulating grain dryer control system was designed for the low-temperature circulating grain dryer using air source heat pump as the heat source. A multi-mode soft logic system was employed to control the low-temperature circulating grain dryer and heat pump system. In the hierarchical system, the PLC was used as the control core and touch screen as the man-machine interface, programming through the sequential function chart (SFC) and ladder diagram, which guaranteed the multi-mode operation of heat pump low-temperature circulation grain dryer and supported manual adjustment and setting of drying parameters, realtime monitoring of grain moisture content and drying temperature and other functions. The experiments validated the operation of the dryer in manual mode, automatic mode and timing mode. The designed logic control system had advantages of high degree of automation, good stability and high safety and so on. Additionally, experiment results on the control of the drying temperature showed that the mean error of the heat pump outlet temperature was 0.95℃ and the standard deviation was 0.93℃ at different ambient temperatures, which meant the control system had high accuracy and stability. The grain temperature was stable at about 33℃, which satisfied the operating requirements of the low-temperature circulating grain dryer.

Abstract:Due to the irregular shape of shrimp, the automatic orientation is difficult. In view of the problems that the orientation link of shrimp pretreatment still relies on manual operation and lacks of orientation equipment, taking headless Penaeus vannamei as the research object, through the analysis of the shape and structure characteristics of the shrimp body, a head and tail orientation method was proposed, and the head and tail orientation mechanism of shrimp was analyzed. The theoretical study showed that it was feasible to orient shrimp’s head and tail by means of opposite rotation of parallel rollers. According to the shape parameters of shrimp, a directional conveying device for shrimp was designed. Medium-sized and large-scale shrimps were selected as the experimental objects. The single factor and orthogonal experiments were carried out with the roller gap, the conveying speed of pushing plate and the rotation rate of orientation roller as the experimental factors, and the orientation success rate as the index. Through the experiment, the influence law of each experimental factor on the orientation effect was explored, and the main parameters of the test device were optimized. The experimental results showed that the optimal parameters combination of the experimental device were as follows: the conveying speed of pushing plate was 40mm/s, the rotation rate of orientation roller was 90r/min, the corresponding roller gap of medium-sized shrimp was 10mm, and that of large-sized shrimp was 13mm. Under the optimal combination of parameters, the success rates of medium-sized and large-scale shrimp were 97.3% and 94.7%, respectively. The research can provide theoretical and technical support for the development of shrimp orientation equipment.

Abstract:In order to improve the way of nondestructive detection of kiwifruit and provide a new method to investigate whether the kiwifruits is infected with Penicillium expansum(P. expansum), the quality, physiological indicators and the relevant electrical parameters within the frequency range of 100Hz~3.98MHz of Hongyang kiwifruit which were infected with 15μL 1×106CFU/mL P. expansum spore suspension in the storage period were measured, while the control group was infected with sterile water. The characteristic frequency and sensitive electrical parameters were screened out through principal component analysis and correlation analysis, and the mathematical models were further established. The results showed that the characteristic frequency of control group treated with sterile water and the inoculated group treated with P. expansum was 3956.5kHz and 251Hz, respectively, the sensitive electrical parameters were equivalent series resistance Rs and equivalent series capacitance Cs, respectively. As for the control group, β-1, 3-glucanase (GLU) activity was highly correlated with Rs, and as for inoculated group, the titratable acid TA content was significantly correlated with Cs (P<0.01). The mathematical regression equations were also established, which indicated that the electrical parameters can better reflect the quality of kiwifruit. As long as the measurement of electrical parameters, the value of GLU activity or TA content can be calculated to determine whether the kiwifruit was infected with P. expansum. The study provided a theoretical basis for the nondestructive detection of kiwifruit penicilliosis.

Abstract:The frozen Longissimus dorsi (LD) of yak meat were taken as the experimental objects to investigate the effects of different thawing methods on the protein oxidation, functional properties and freshness of yak meat. The results showed that the myofibrillar protein carbonyl groups were the highest of 9.80nmol/mg, surface hydrophobicity was the highest of 48.53μg and the total sulfhydryl groups were the lowest of 41.73nmol/mg and Ca2+-ATPase activity was the lowest of 0.245U/mg protein in the air thawing yak meat, which was significantly different from other thawing methods (p＜0.05), indicating that the meat protein oxidation was the heaviest after air thawing. And the other thawing methods resulting in protein oxidation in descending order were refrigerated thawing, hydrostatic thawing, microwave thawing, and room temperature thawing. The protein solubility was 109.28mg/g, MP solubility was 69.16mg/g and EAI was 31.51m2/g in the air thawing yak meat, all of which were significantly lower than those of other thawing methods (p<0.05), indicating that air thawing was most unfavorable to meat proteolysis and emulsification. Meanwhile, the lipid peroxidation of refrigerated thawed yak meat was the lowest as well as the TVB-N of microwave thawed yak meat was the lowest. There was a significant correlation between protein oxidation and lipid oxidation in thawed yak meat, and also between protein solubility and emulsifying ability (p<0.01). These above results indicated that the air thawing was most detrimental to the quality of thawed yak meat, followed by thawing at room temperature. However, refrigerated thawing, microwave thawing and hydrostatic thawing respectively can delay the onset of meat quality decline process effectively in different levels, and can maintain the defrosted meat quality to some extent. The findings can provide data support and theory reference for illustrating the mechanism of different thawing methods on defrosted meat quality in the future.

Abstract:Water-holding capacity (WHC) refers to the ability of meat to retain moisture when force is applied. As an important processing quality, WHC of beef affects the quality of the final products. The quality of different parts of beef have significant differences. Different WHC not only has an important impact on the flavor, tenderness and juiciness of different parts of beef, but also has an important economic significance. It is helpful to take corresponding measures to control the loss of water by clarifying the WHC of different parts of chilled beef during aging time. Thus, in order to investigate the changes of water distribution during postmortem ageing, the chilled M. longissimus lumborum (LL), M. semimembranosus (SM) and M. psoas major (PM) beef aged for 1d, 2d, 3d, 5d and 7d were investigated. From the perspectives of purge loss, low field nuclear magnetic resonance (LF-NMR) characteristics and muscle structure, the changes of WHC, water distribution and microstructure of the three parts of chilled beef were described. The results indicated that the purge loss of three parts of chilled beef was increased significantly during postmortem aging time. And the WHC of PM were significantly higher than that of LL and SM (P<0.05). The area of relaxation peak (A2) in three parts of beef was decreased significantly (P<0.05) and the differences among P2b and P21 were not significant (P>0.05). The loose arrangement of myofiber and the short length of sarcomere resulted in the poor WHC of SM. PM had longer sarcomere and smaller myofiber diameter. The difference of myofiber structure in three parts of beef resulted in different WHC.

Abstract:Aiming at the problems of difficult leveling, poor stability, complex operation and poor safety of traditional tractor in slope operation, a kind of omnidirectional leveling hillside crawler tractor (abbreviated as hillside tractor) was developed. Firstly, based on the analysis of the leveling principle of hillside crawler tractor under condition of slope, the lateral leveling scheme based on parallel four-bar mechanism and the longitudinal leveling scheme based on double frame mechanism were proposed. Secondly, the key components of the hillside tractor (omnidirectional leveling device, walking system, stepless speed regulation drive system based on hydrostatic drive (HST), multi-functional hydraulic system, slope adaptive hydraulic suspension device) were selected, designed and created. Finally, the performance test of hillside tractor was carried out. The test results showed that the vehicle body leveling could be realized on the 0°~15°lateral slope and 0°~10°longitudinal slope, which could effectively improve the stability and safety of the slope driving and operation;the stepless speed regulation of 0~8km/h could be realized, which could meet the speed requirements of flat driving, climbing, contour line operation and other working conditions;the driving, braking, steering, and omnidirectional (lateral and longitudinal) leveling of hillside tractor and posture adjustment of farm tools could be realized by remote control, which greatly improved the convenience of operation;the ground pressure of hillside tractor was 0.025MPa, which indicated that it had good trafficability on soft road and swamp;the steering performance of hillside tractor was good, and the minimum turning radius was 1728mm, which could be used in relatively narrow slope working environment such as hilly areas;the flat deviation rate of hillside tractor was 55%, and the deviation rate after leveling on 15°slope was 5.75%, which was less than 8.62% without tractor leveling, and all met the requirements of corresponding national standard (≤6%); the maximum lifting force of the hydraulic suspension device was 8.2kN, which can meet the operation requirements;the stability of rotary tillage depth on slope land met the national standard (≥85%).The research result had a good reference for the research and development of agricultural machinery in hilly areas.

Abstract:According to the topology design theory and method of parallel mechanism (PM) based on position and orientation characteristic (POC) equations, a PM which can realize one-translation and two-rotation (1T2R) was designed. The main topological characteristics of the PM, such as POC, degree of freedom (DOF) and coupling degree (κ) were analyzed. It was shown that the PM contained only one sub-kinematic chain (SKC) and its coupling degree was one, but its symbolic direct position solutions could not be obtained. Only the numerical methods could be used to solve its numerical direct position solutions. Therefore, the topology coupling-reducing design was carried out to obtain the 1T2R PM with zero coupling degree, symbolic direct position solutions but constant POC and DOF, because zero coupling degree would benefit the subsequent error analysis, trajectory planning, motion control and dynamic analysis of the PM. Both the symbolic direct position solutions and inverse solutions were derived, and the correctness of them were proved by example. The singularities of the PM were analyzed based on the Jacobian matrix, and some cases of the singular positions of the PM were given by the graphs. Based on symbolic direct position solutions, the workspace of the PM was calculated and analyzed, which was consistent with that obtained by the inverse position solutions. At the same time, it showed that the workspace of the PM was large and the shape was regular. It was found that the workspace based on symbolic direct position solutions had the advantages of less computation and accurate boundaries.

Abstract:The full-actuated parallel manipulator can realize the actuation redundancy of manipulator by increasing one or more constraint branch chain or joint. In order to analyze the optimization of the constrained branch chain and joint to the dynamics performance of the redundantly actuated parallel manipulator, a deep analysis and comparison of dynamics characteristics between redundantly actuated parallel manipulator 6P-US+UPU and non-redundantly actuated parallel manipulator 6P-US were conducted, and the optimization mechanism of constrained branch chain on the dynamics performance of redundantly actuated parallel manipulator was discussed. Based on the principle of virtual work, the dynamics models of non-redundant manipulator 6P-US and redundant manipulator 6P-US+UPU were established respectively. Then according to different influence factors on the dynamics, dynamics performance indexes of redundantly actuated parallel manipulator with clear physical meanings were proposed. The influences of constrain branch chain with and without mass on the different dynamics performance indexes of the manipulator were analyzed subsequently. Prototypes of the two manipulators with specified inertial and geometric parameters were given and numerical comparisons of dynamics performance was carried out finally. The results showed that the branch constraint improved the dynamic performance of the mechanism, while the quality of the branch would reduce the performance.

Abstract:In order to improve the machining accuracy of the cam profile surface and reduce the influence of tool errors on the normal error of the cam profile during the machining process, by analyzing the structural characteristics of the cam profile indexing cam and the existing problems in the machining method, and reducing the extreme difference between the actual profile and the theoretical profile obtained by processing, the principle of multiple envelopes for one-side processing was proposed, The effectiveness of the method was demonstrated and a simulation calculation was carried out. The theoretical research was based on the cam profile equation of the arc surface indexing of the set coordinate system, using spatial meshing principle and rotation transformation matrix, according to the principle of one-side machining and multiple envelopes, the actual contour equation after the machining of the globular indexing cam was deduced and calculated, the tool position compensation and tool control method of one-side machining multiple envelope principle was researched. The problem of tool position control based on the principle of multiple envelopes for one-side machining of a globular indexing cam was solved. Matlab simulation results showed that the method can significantly reduce the normal error of the cam profile surface and improve the machining accuracy of the cam, and it had important practical application value.

Abstract:Aiming at the problem of machining inefficiency caused by narrow path interval in the current five-axis CNC machining process of triangular mesh surface, a tool path generation algorithm for variable path interval in machining was proposed. Firstly, with the data reconstruction of triangular mesh, an intersection-algorithm based on KdTree mesh partition was proposed, which achieved fast grid data point acquisition. Then, with analyzing the geometric characteristics of the toroidal cutter and the triangular mesh surface, a method was proposed to maximize the effective cutting radius by changing the tilt angle and optimizing the inclination angle under the condition of no curvature interference. Finally, a discrete model of toroidal cutter was established based on the characteristics of triangular mesh surface, and the corresponding interference detection and correction method was also put forward. The simulation results showed that the optimization method of path intervals can significantly improve the machining path intervals compared with the existing methods under the same machining conditions and the presented method of checking and correcting the tool-interference can effectively avoid the local interference of cutter bottom and the global interference of cutter arbor. It can be seen that the methods proposed had upper theoretical significance and practical application value in improving the efficiency of five-axis CNC machining for triangular mesh surface.

Abstract:In order to improve the accuracy and generalization of the thermal error model of CNC machine tools, a thermal error model of and long short term memory convolutional neural network based on attention mechanism (AM-CNN-LSTM) was proposed. A thermal error model with two branches was established by using the ability of convolutional neural networks to extract the space state features of high-dimensional data and the ability of long short term memory networks to extract long-term sequence state features, and the extracted features were input into the attention mechanism to reconstruct according to the importance, and then a feature map of original data and thermal error value was established. Finally, the thermal error prediction value was performed through the full connect layer. The G460L CNC lathe was used to collect experimental data, the temperature and thermal error collected in different seasons were used as the model input, and the model was trained using the cyclic learning rate and regularization optimization method. Compared with the thermal error model of LSTM, ConvLSTM and CNN-LSTM, the results showed that AM-CNN-LSTM model had the strongest ability to restore features and the smallest residual error range. It was decreased by 62.09%, and the prediction accuracy of the model was within 2.4μm.

Abstract:In order to improve the stability and measurement accuracy of the magnetostrictive displacement sensor, and reduce the influence of the reflected wave signal on the effective signal of the output voltage, the generation and influence factors of the reflected wave voltage were studied. Based on the analysis of the influence of elastic pressure between damping and waveguide wire and the friction generated on the stress wave propagation, a voltage model of the reflected wave of the magnetostrictive displacement sensor under the action of damping was constructed, and the voltage amplitude of the reflected wave of the magnetostrictive displacement sensor was calculated. The results showed that the change of the damping parameter had a great influence on the amplitude of the reflected wave voltage, and the calculation result was basically consistent with the change trend of the experimental result. Based on the reflected wave voltage model, the optimal values of the damping length and Shore hardness were determined to be 5~10mm and 50~75, respectively, and the damping with a diameter of 10mm, a length of 10mm and a Shore hardness of 50 experimental verification were verified by experiments. Under the damping effect, the amplitude of the reflected wave voltage was reduced from 75mV to 4mV. At this time, the amplitude of the reflected wave voltage was much smaller than that of the effective signal, which greatly reduced the impact on the effective signal of the output voltage. The research result can provide a theoretical basis for the selection of damping parameters of the magnetostrictive displacement sensor.

Abstract:The non-ideal pipe flow would induce unsteady fluctuation under the condition of turbulence, in which large scale fluctuation would influence the measurement accuracy of ultrasonic flowmeter. A pack-of turbulent dissipation grilles was designed paralleling to the acoustical path of the ultrasonic flowmeter to limit the scale and intensity of turbulent fluctuation. The turbulence reduction grilles can optimize the flow field from two aspects: reducing the vortex intensity and controlling the turbulence pulsation scale. And the diameter of the grid bar was close to 1mm, which had little effect on the average flow velocity. A numerical simulation based on detached eddy simulation (DES) was used to study the unsteady internal flow and the simulation was verified by laser Doppler velocimetry (LDV) experiment. The design was compared with the original setup, and the design did not disturb averaged flow field due to the small diameter of the turbulent dissipation grilles. For vorticity magnitude, the turbulent dissipation grilles could smooth the flow and restrict the vorticity magnitude around the acoustical path. In the range of L＝10~50mm, the vorticity magnitude of the design was decreased significantly. For fluctuating scale, the turbulent dissipation grilles broke the mechanism which turbulence propagated toward larger scale. When Re was greater than 10000, the turbulence reduction optimization had a significant effect on the reduction of turbulence error. When Re was greater than 40000, the turbulence error was about half of that before optimization. The introduction of turbulent dissipation grilles had a promising prospect for the improvement of the time domain resolution of ultrasonic measurement and the design of low power consumption.