Research On Key Technologies Of Intelligent Cucumber Picking Robot

At present,the picking of agricultural fruits and vegetables in China is still mainly manual,with low efficiency,generally poor picking environment,high work intensity,high physical requirements for picking workers,and poor continuous working ability.The research on intelligent picking robot is still in the stage of testing machine.Based on the existing picking robot at home and abroad,this paper studies the key technologies of cucumber intelligent picking robot through the idea of modularization and structure,in order to meet the requirements of high efficiency and intelligence of the current agricultural picking device,and provide some reference for intelligent picking of fruits and vegetables.The main research work of this paper is as follows:(1)In order to solve the problems of low picking efficiency and slow cutting speed of cucumber vine,a pneumatic cutting device of cucumber vine was designed,and the dynamic simulation analysis of the device was carried out.The data of cucumber with different maturity were measured and analyzed,and the shearing test was carried out.The influence of maturity,blade angle,cutting angle,oblique cutting angle and other parameters on the shearing effect of cucumber vine was researched.According to the above research,it provided the basis for reducing the picking time of cucumber and determining the parameters of the shearing device.(2)In order to solve the problem that the holding device would cause different degrees of damage to cucumber during picking,a pneumatic flexible holding device with low holding damage was designed,and its flexible finger structure was analyzed.Through finite element simulation,the stress of cucumber was compared between the flexible holding device and the traditional hard material holding device under the same load,that this design has less damage to cucumber during holding.Cucumber samples were extracted by clamping test with flexible finger samples with different hardness.The potential damage of the specimen was observed from macroscopic and microscopic point of view,and the optimal parameters of the flexible finger were obtained.(3)In order to solve the problems of the similar color between cucumber and background and the difficult to determine the coordinate position of the clamping point,the recognition and location of cucumber picking robot is studied.According to the characteristics of different neural network algorithms,a multi-path convolution neural network algorithm is proposed.The image is preprocessed by binarization,feature extraction and other methods,and the cucumber background is filtered.The three components of color with high weight are input into the deep learning module to complete the data training and testing.Combined with feature fusion and support vector machine classification,the recognition of cucumber was completed.By analyzing the calibration principle of binocular camera,the parameters of binocular camera were calibrated,and the center of mass of cucumber was determined by ROI region centroid calculation method.(4)In order to simplify the control mode of cucumber picking robot,a gantry motion control system is designed.By analyzing the motion characteristics of the device,the technical parameters of the hardware were determined and the hardware assembly is completed.The connection circuit between each hardware module is designed,and the communication protocol between each module was determined,and the program design of the vision module,the end effector control module and the motion control module were carried out,and the coordinated control among the modules is completed.(5)The cucumber picking experiment was carried out in the daytime using the designed cucumber picking robot,and the results showed that the success rate of this design could reach more than 80%in the daytime.The cucumber picking experiment was carried out at night with the designed cucumber picking robot,and two kinds of test methods were adopted:different light intensity and different light angle.The results showed that:1)At night,the picking success rate was the highest when the light intensity was 250 lux,which could reach 85%;2)At night,the picking success rate was the highest when the angle of light supplement is 0-30°,which can reach 80%.