Research On Flower Thinning Device Based On Vision Technology

In the orchard management work,flower thinning is one of the main labor-intensive and time-consuming links in orchards,and at present,it is mainly based on traditional manual flower thinning operations,and there are problems such as high cost,low efficiency and high labor intensity.Therefore,the development of intelligent flower thinning device to replace artificial flower thinning,to improve the efficiency of modern orchard flower thinning operations.Starting from the planting characteristics of orchards in China,this study adopts mechanical theoretical analysis and design,hydraulic design and selection,flower collection system construction and related program development to achieve the construction of intelligent flower thinning device based on visual technology,and optimizes the key parameters of flower thinning device by using the test bench.The main contents of this article are:(1)Through the investigation of the modern orchard model,the flower thinning plan was determined according to the orchard thinning requirements.In order to identify flowers on fruit branches,a flower information collection system was developed and built.The flower characteristics of apple trees and peach trees were analyzed,and the mechanical properties of apple trees and peach branches were studied.And design and selection of its variable closed piston pump,motor,hydraulic motor and auxiliary components.(2)In order to obtain the proportion of the petal area of the collected information,two effective flower recognition algorithms are developed: first,the image is preprocessed by the Opencv flower recognition algorithm,and the petal area is obtained by accumulating the contours of each petal,and then calculating The proportion of petal area;second,the flower recognition algorithm using nonlinear regression based on the pixel value of the primitive,that is,the flower recognition algorithm with the pixel value of the primitive as the independent variable,the algorithm using the RGB value of the primitive as the variable,construct a nonlinear primitive classification function,and then convert it into an unconstrained optimization problem through the least square method,and obtain the optimal solution through optimization calculation,which is the coefficient value in the classifier,and finally realize the optimization of each basic primitive numerical representation of.In order to realize the automation of flower thinning,a GUI display page was designed using Python software,and its components and working principles were introduced;related programs were developed to control the camera status,organize and transmit petal information to the lower computer.(3)In view of the complicated walking operation of the flower thinning device,the wireless remote control is used to transmit the work instructions to the single-chip microcomputer,and the four advanced timers TIM1 in the single-chip microcomputer output PWM waves to control the work of the two driving motors respectively,which simplifies the moving operation process of the device and improves the work efficiency.efficiency.In order to realize the adjustment of the motor speed of the thinning shaft,the encoder is used to output the speed of the shaft as pulses,and the relative pulse data of AB is collected and sent to CAN communication.The shaft speed detection is completed and displayed on the GUI page;the host computer communicates through CAN.Send the command to the single-chip speed control module and control the output.(4)In order to determine the quantitative parameters of the flower thinning device and facilitate the study of flower thinning experiments,repeated tests were carried out on the rotation speed of the flower thinning shaft under two rotations of the motor.The influence of light source intensity on the recognition accuracy of image acquisition and video acquisition shows that the recognition effect is the best when the strip light source is 6 and the ring light source is 8.The impact force of three kinds of adhesive strip materials was measured by Aipu digital dynamometer to determine the thinning adhesive strip material.In order to determine the best parameters for the thinning effect,the effects of different strip intervals(4 cm and 5 cm)and the speed of the thinning shaft(1-10 r/s)on the thinning rate of simulated trees were explored in manual mode.The comparative analysis of the test results shows that the optimal parameter combination is that the strip interval is 4 cm and the transmission speed is 6-8 r/s.In view of the fact that the speed of the flower thinning axis cannot be accurately controlled in the automatic flower thinning mode,the compensation coefficient K is introduced to control the speed of the flower thinning axis,and through multiple tests and verifications,the optimal K value of the flower thinning operation in different petal area ratio ranges is determined.scope.