Research On Target Matching And Positioning Technology Of Industrial Robot Based On Geometric Primitives

With the increasing demand for automation in modern manufacturing industry,machine vision is widely used in various fields.Many experts and scholars proposed a variety of target matching and positioning methods,but there are still problems of poor robustness,low matching rate and imprecise location in the complex industrial production applications.In this paper,the target matching and positioning method of industrial robot is taken as the research direction.the basic methods of image preprocessing are analyzed and discussed,and the problem of target matching and positioning methods based on feature points and geometric primitives is studied deeply.The main research content includes:Firstly,the basic principles of image smoothing,threshold segmentation,edge detection and contour tracking are summaried.The common threshold segmentation algorithms and edge detection algorithms are compared by simulation.Then the contour points are obtained by tracking the target contour with 8 neighborhoods.Secondly,the extraction process of feature point of SIFT algorithm is improved.The edge of the tool image is detected by the improved Canny operator,and morphological operation is used to remove the edge fracture.An improved SIFT algorithm is proposed.Simulation results show that the improved algorithm outperforms the traditional SIFT algorithm in terms of matching rate and speed.Finally,a method of workpiece matching and positioning that based on the principle of line spread function,and high-accurate geometric primitives extraction is proposed.The parameters of the initial geometric primitives are obtained by the least square method,the line diffusion function is introduced to accurately locate the contour points,and the precise geometric primitive parameters are extracted.Then the angle and distance between the primitives are calculated to complete the matching and positioning of the workpiece.Simulation results show that the proposed method can achieve accurate matching and precise positioning of the workpiece,and ensure real-time performance.