Researches On The Identification Of Tender Leaves And Bionic Plucking Fingers For High-quality Green Tea

The high-quality green tea is very particular about the tender leaves.The existing tea plucking machines are based on the working principle of shearing type to harvest tea without selection,which cannot meet the requirements of plucking tender leaves for making the high-quality green tea.At present,the harvest of tender leaves of the high-quality green tea can only rely on manual plucking.However,labor shortage has become a bottleneck restricting the development of tea industry.Therefore,it is urgent to develop a tea plucking robot that can realize selectively pluck tender leaves of the high-quality green tea.This paper takes the early spring high-quality green tea as the research object.Based on the existing problems in the current researches on the intelligent plucking of tender leaves,this paper researches the two key technologies that the identification of tender leaves and the design of end effecters.The main research contents are as follows:In the aspect of tender leaves recognition,based on the three-channel components' pixel distribution of the background and the tender leaves in RGB space,a tender leaves segmentation algorithm,(R-B)>80&&(G-B)>80,is proposed;The three color components of R,G and B are taken into account based on the iterative method of color factor,and a tender leaves segmentation algorithm,-1.01*G+0.8*B+0.19*R>99,is proposed.Then,the area filtering method based on the connected region marker is used to remove the non-interest region and the small-area unopened tender leaves in the segmentation processed image.,which will effectively preserve the pickable sprouts.Finally,the paper compares and analyzes the tea tree images of different varieties,different plucking seasons and different light intensities.In the design of the end effector,this paper carried out the experimental research on the shape and mechanical properties of the tender leaves,and then based on the experimental research results,Solidworks software was used to model the plucking fingers in three dimensions.Two biomimetic plucking fingers that can be used to simulate the manual plucking action were designed.One is electric drive and the other is pneumatically driven.Dynamic and static analysis are carried out to improve the dexterity and compactness.The geometric parameters of the structure were optimized by genetic algorithm combined with MATLAB software.And the prototype of the bionic plucking fingers is made.Through experimental research,it is found that the two bud segmentation algorithms explored in this paper can effectively segment the buds and backgrounds of different varieties,different regions and different plucking seasons.The algorithm has strong universality and the identification rate of tender leaf is 95%.At the same time,the two algorithms also have better adaptability to changes in light intensity.By optimizing the geometric parameters of the bionic picking fingers,the driving efficiency of the driving force is improved by 26.6%.The preliminary results of the indoor test show that the success rate of harvesting is 90%,which can achieve the expected action,which will be beneficial to the promotion and application of tea intelligent plucking in the future.