| As an ornamental plant,hedges are widely planted in municipal parks and other places,but the contradiction between the increasing amount of pruning operations and the labor shortage caused by the aging of China’s population has become prominent year by year.Pruning robot technology is one of the important ways to solve the above contradictions at present.The demand for automatic pruning is more urgent,such as the pruning of special-shaped hedges with spherical,cylindrical,oval,etc.However,the existing profiled cutter special-shaped hedge pruning automation device has poor universality for pruning objects.Moreover,the pruning model construction method of pruning robot based on traditional point cloud registration has poor accuracy and low pruning efficiency.In this paper,a method of machine vision point cloud processing based on robotic arm posture was proposed to construct hedge model and pruning tool trajectory planning,and then,a composite robot system based on mobile robot chassis and robotic arm was built to carry out experimental research on special-shaped hedge pruning.(1)In this paper,a construction method for special-shaped hedge point cloud model based on robotic arm posture fusion was proposed.The point cloud data collected by the camera mounted on robotic end-of-arm effector was matched with the current position of the robotic arm,and the point cloud registration algorithm was used to fuse the two to realize the extraction of the original point cloud of the hedge.Compared with the general nearest neighbor point cloud method,the error of the construction method of the hedge point cloud model was reduced by 30.77%,and the speed was increased by 60.17%.On this basis,the above point cloud data with robot arm pose information was fitted by straight-through-statistical filtering and Random Sample Consensus(RANSAC)method to obtain a complete hedge point cloud model,which provided complete hedge data support for subsequent pruning tool path planning.(2)In this paper,a grid spatial clustering method and pruning trajectory planning strategy based on hedge surface pruning weights were proposed.In order to obtain the hedge model meshing strategy,this paper explored the matching relationship between the surface triangle mesh granularity of the hedge model and the pruning cutter size,and when the mesh granularity was 0.07,the number of meshes and the tool size reached the most balanced state.A grid clustering method was proposed,which was to cluster the spatial grid of hedge surface after planarization treatment by pruning weight,and when the weight λ was 40.26,the local grid formed by the clustering covers the pruning grid that can be reached by the most robotic arms.(3)The composite robot system of mobile robot chassis and robotic arm was carried out for experimental research.A composite pruning robot system integrating mobile robot chassis and robotic arm was constructed,and the pruning experiment was carried out on the spherical hedge of Privet with small-leaved privet as the pruning object.In this paper,the effects of initial parameters such as the number of iterations,pheromone update model and total amount of ant colony genetic algorithm on the time cost of pruning tool trajectory were studied.When the number of iterations is50 and the pheromone update model is ant-circumferential and Pruning path with an optimal displacement cost of 202.01 cm obtained at a total pheromone release of 75... |