Research On Key Techniques Of Strawberry Picking And Optimal Design Of Manipulator

In recent years,strawberries have become an important economic crop and are grown all over the world.my country has become the largest producer and exporter of strawberries,ranking first in the world.Strawberry is rich in nutrients and rich in many vitamins.It has been called the "Queen of Fruits" and is deeply loved by consumers.However,strawberry is a laborintensive product.The planting and picking of strawberry are all done by manpower.The economic cost of fruit growers is high,which affects the income of fruit growers.At present,with the development of agricultural planting and picking automation,the realization of automated strawberry picking is becoming more and more important.In the past three years,my country has the largest number of patents for strawberry harvesting robots.The increase in the number of patents for strawberry harvesting robots reflects the growing interest in the use of robots in agriculture.The degree of agricultural automation is high and the trend of automation is more durable.This paper designs a strawberry picking manipulator according to the growth status of elevated strawberries,and conducts simulation research on strawberry identification,positioning and picking process,as follows:(1)A new type of strawberry picking manipulator is designed,and a new end effector is designed using Solid Works.The manipulator is a rectangular coordinate manipulator that is convenient for numerical control.Then select the material and drive of the entire manipulator according to the physical characteristics of the strawberry,and finally use ANSYS Workbench finite element analysis to verify whether the material of the end effector meets the requirements of picking.(2)The multi-state image of strawberry is collected by the camera and stored in the mobile workstation.The RGB image is converted to the YCbCr color space by MATLAB programming.After the grayscale comparison,the Cr channel is selected to segment the target strawberry;then the FCM segmentation method is selected for segmentation In the homework,the target strawberry image is separated clearly and non-destructively,and the Canny algorithm is used to check the edge area of the target strawberry that has been segmented;after the detection is completed,the edge area in the strawberry image is extracted by the minimum rectangle method to make a recognition sample library.(3)Build a convolutional neural network model based on GooLeNet Inception v2 for strawberry recognition,use sample sets and test sets for recognition training,and then use the validation set to verify the accuracy of the model;after identifying the target strawberry,perform binocular vision positioning using Zhang’s The calibration method calibrates the camera and obtains the internal and external parameter matrix,then corrects the image,then obtains the coordinates of the strawberry centroid point through MATLAB and calculates the strawberry picking point.(4)According to the picking requirements,ADAMS software is used to carry out threedimensional solid modeling,realize the kinematics,dynamics and statics analysis of the picking manipulator,and perform virtual picking of the strawberry picking process to ensure that the strawberry picking manipulator meets the picking requirements.