The Study Pivotal Technologies Of Automatic Flat Groove Cutting Robot's Vision System

The national number of steel used is increasing year by year,along with the acceleration of the national industrialization,among which flat steel is particularly remarkable.The higher the cutting quality and efficiency of flat steel,the higher the requirements.Hand-held and semi-automatic cutting methods are gradually being phased out.With the development of science and technology,the promotion of new technologies for industrial production is very rapid.In particular,it is applied to practical engineering that the industrial cutting robot with visual functions,with the development and maturity of machine vision technology.The automatic flat groove cutting robot automatically recognizes the position of the workpiece in space and guides the robotic arm to cut workpieces with high precision.In this paper,we conduct research about the target contour extraction,visual calibration,identification and matching with the standard parts' templates on this robot.The main research contents of the targeted research are as follows:(1)Extraction of Workpiece Contour:Based on the depth map and gray image of the workpiece,the target of the workpiece is extracted by extracting the binary contour map of the depth map and multiplying with the gray image.The image of workpieces is preprocessed to reduce noise and enhance image edge information.Find the appropriate threshold on the histogram of the target area,and extract the edge of the workpiece object with Canny edge detection operator.It is useful to solve the optimal contour of the workpiece with combining the contour edges of the binary target area of depth image.(2)Visual calibration and stitchingAfter the principle of camera imaging is analyzed,the camera calibration and the translation parameters of left and right cameras are calibrated.The calibration of the hand-eye relationship of the robot is performed,based on the calibration of the robotic camera.After the cutting robot is visually calibrated,we can establish the position information conversion relationship between the image coordinate system and the world space coordinate system.So,the robot can perform spatial operations according to the image information.The simple stitching method of the left and right cameras is analyzed by the principle of imaging coincidence area,and the contour of workpieces' image is stitched.(3)The contour of workpiece recognition and matchingThe edge contour of workpieces can be obtained by the method simple stitching,and the CAD's Hu moments in the standard library can be calculated and compare with the workpiece'Hu moments,so the similarity between them can be obtained.Based on this,the corresponding CAD template of the matching standard component is correctly identified.Then,we can obtain the process parameters of CAD template,thereby the cutting path of the cutting robot can be corrected and the precise cutting can be realized.