Research On Visual Guidance Technology For Tea Picking Machinery

Our country’s tea production is huge and has a broad market.The current picking work is mainly manual,which is not conducive to industrial upgrading.The existing mechanical tea picking equipment is only suitable for low-quality bulk tea picking,and a selective automatic picking equipment is needed to improve the quality of tea picking and reduce labor costs.To realize the automation and selective picking of tea picking equipment,two major problems need to be solved: one is efficient picking without damaging the tea leaves,and the other is the precise identification and positioning of buds.To this end,this article has carried out research on the structure of the tea picking manipulator and the visual guidance technology of the manipulator.First,starting from the tea growing environment and the actual process of tea picking,a design of a tea picking manipulator is proposed.The manipulator has four degrees of freedom,and the buds can be identified and positioned by installing a binocular camera at its end.The end effector is designed according to the shape and size of the tea,and the tea picking and collection are synchronized to improve the picking efficiency.The feasibility of picking is verified by the physical prototype of the manipulator produced.Then,establish the kinematics model of the tea picking manipulator,derive the forward and inverse kinematics model of the manipulator,and analyze the working space to ensure that the working range of the manipulator can cover the area where the buds are to be picked and reduce the missed rate of buds.In order to improve the motion stability of the manipulator,trajectory planning is carried out in the joint space.Aiming at the problem of picking path planning for the buds to be picked in the picking process,adaptive improvements were made on the basis of the classic ant colony algorithm,and verified by simulation experiments,the improved algorithm is better in terms of the shortest path distance and the number of iterations.After that,research the visual guidance system of the tea picking manipulator and its key technologies.Complete the binocular camera calibration and hand-eye calibration of the manipulator.According to the characteristics of the bud image,the feature extraction is carried out and improved based on the watershed algorithm.The BM3 D denoising algorithm and the gray-scale stretching method are used to process the segmented bud image to segment the more complete bud image.The SURF algorithm is used to extract the feature points,the RANSAC algorithm and the epipolar constraint are used to remove the mismatched points,and the minimum bounding rectangle method and the binocular distance measurement principle are combined to complete the three-dimensional location of the tea picking point.Finally,the motion control experiment,binocular vision system experiment and visual guidance positioning experiment were carried out on the tea picking manipulator.The experimental results show that the positioning error of the manipulator is within ±2mm,and the average positioning time is 1.88 s.The tea picking manipulator has good position control accuracy and can quickly realize autonomous positioning through the visual guidance system.