Research On Key Technology Of Tomato Picking Manipulator

At present,tomato fruit picking is mainly carried out by artificial methods,but artificial harvesting of tomatoes is laborious,time-consuming and inefficient,which is not feasible for large-scale tomato plantations.Meanwhile,with the increase of labor costs in China,this problem becomes more prominent.In this context,automatic harvesting technology has become a more efficient method to replace manual harvesting.In the face of the actual situation,this paper developed a kind of flexible picking end-effector for the soft and easily damaged characteristics of ripe tomato fruit.In view of the complex picking environment in the visual system,tomato fruit feature extraction was studied.Moreover,mathematical modeling and trajectory planning of the 4-DOF manipulator were studied based on MATLAB,which provided data support for the actual picking operation of the manipulator.The main research contents of this paper are as follows:(1)In view of the soft and vulnerable characteristics of ripe tomato fruit,this paper developed a flexible end effector for picking tomato fruit.The end effector is equipped with a thin film pressure sensor below each finger.When the finger clamping pressure reaches the threshold set in this paper,the pressure sensor can control the corresponding steering gear to stop rotating.In order to determine the value range of pressure sensor pressure threshold,this paper conducted 100 groups of compression failure experiments on "Millennium cherry tomatoes" based on TA.XTC+TA.Touch texture instrument,and finally determined that the pressure threshold of the sensor should be lower than 11.94 N,and calculated that the pressure threshold of the sensor should be greater than 0.46 N by measuring the quality of each group of tomatoes.Based on the results of physical experiments,the hardware and software of the end-effector are designed in this paper.(2)In this paper,tomato fruit feature extraction in the visual system was studied in view of the complex picking environment of the picking manipulator.Firstly,the color model x R-y G was customized according to the growth characteristics of tomato plants and the color differences between mature fruits and leaves,and the model n R-G was obtained by normalization processing.The threshold segmentation experiment was carried out on the tomato fruit images collected in the laboratory to determine that "n" was 0.9(under 8Lux)and 1.1(under 129Lux).Finally,through the comparison experiments of threshold segmentation,edge extraction,ellipse fitting and noise reduction with YUV color model,the characteristics of the suitable environment for each color model are finally determined.(3)Based on MATLAB,the mathematical modeling and trajectory planning of the4-DOF manipulator were studied,which provided data support for the actual picking operation of the manipulator.This article first uses the D-H method to mathematically model the manipulator,and then uses the "standard" second conversion matrix to analyze the forward kinematics of the manipulator model.Based on the results of the positive kinematics analysis,this article uses the algebraic method to perform inverse kinematics on the manipulator.Analyze and verify the correctness of the analysis results of forward and inverse kinematics in this paper through MATLAB kinematics simulation.Finally,trajectory planning simulation of the manipulator was carried out based on the established mathematical model,which laid a solid foundation for the research and development of the 4-DOF manipulator in the later stage,and provided theoretical support for the tomato fruit picking experiment.(4)In order to analyze the actual picking performance of the end-effector developed in this paper,a simulated picking experiment was carried out on the endeffector.The results of the simulated picking experiment showed that the success rate of the end-effector picking experiment could reach 84%,and the average picking time was 6.18 s,respectively.In addition,in each picking experiment,the clamping actuator can control the steering actuator to stop rotating when the sensor pressure reaches the threshold set in this paper.In 42 successful picking experiments,there was no fruit damage phenomenon,indicating that the end-effector designed in this paper has good coordination performance and meets the actual picking requirements.