Research On Image Contour Extraction And Matching Of Groove Cutting Robot

With the continuous development of robotics and digital image processing technology,the robot system with vision processing is gradually applied to industrial production.In the field of industrial cutting,the groove cutting of the workpiece occupies a large proportion,traditional method of manual cutting and semi-automatic cutting has been unable to meet the requirements of modern industrial production.Therefore,it has important significance to research the automatic groove cutting robot which can automatically identify the type and precise location of workpiece and cut automatically.The groove cutting robot need to move several times to collect images of workpiece on cutting platform,then stitching these images into a complete image of cutting platform.Then extract the contours of the workpiece to be cut in the image,and compare it to the standard workpiece contour template in the workpiece library,thereby identify the type of workpiece and get its standard size information.Finally the robot use this size information to program the cutting track.This paper studies the method of workpiece contour extraction and matching in image processing of groove cutting robot by practice.The main contents are as follows:(1)Study the basic principle of image mosaic.Introduce the basic process and geometric basis of image mosaic.Analyzing the methods of image registration and image fusion.The image splicing method based on SIFT feature points is studied,and use the commonly method of weighted average to fuse overlapping parts of the image.Then use this method to stitch multiple pieces of workpiece image into a complete image of cutting platform.(2)In terms of contour extraction,the method of image contour extraction is studied,which include image preprocessing,image segmentation,morphological processing,contour extraction and contour tracking.For the characteristics of the images in this paper,the contour extraction method based on edge detection is selected.By comparing the effects of various differential edge detection operators on the edge detection of the workpiece image,this paper choose Canny operator to detects the edges of image.For the problem of singular point in the contour,a contour tracking algorithm is proposed to eliminate singular points and using this method to obtain a closed single pixel wide contour.(3)In terms of contour matching,the main characteristics and defects of the modified histogram-based chain code are analyzed in describing the contour which based on the Freeman chain code.For the shortcomings of the modified histogram-based chain code for the description of any angle rotation of the contour are not invariant,this paper use the method combining the smallest circumscribed rectangle of the contour to solve the above problem.And the method is used to match the workpiece contour and the standard template to verify its feasibility.Finally,the method of comparing the contour area is used to select the mismatched contours with the same shape and different dimensions,and identify the type of workpiece.The method used in this paper realized the image mosaic,contour extraction and contour matching of the workpiece in the groove cutting robot,which has certain feasibility and practical engineering significance.