Detection Of Weld Reinforcement And Width Based On Structured-lighted-stripe Vision

At present,the weld was observed mainly relying on the naked eye and was grinded by the use of Chiba film or grinding wheel in the industrial situation.This kind of operation mode is not only inefficient,but also its polishing precision and stability is poor.The development of machine vision technology provides a new way to solve the problem of welding seam polishing,especially the application of laser structured light vision in weld seam detection,which can solve the above problems.The most important measure in the welding process is the reinforcement and the width,so this paper studies the key technology of the weld detection.Firstly,this paper introduces the whole system of laser structured light vision and its principle.According to the actual working conditions,the laser light path is designed and the CMOS camera is selected.And taking into account the characteristics of the welding material,the 650 nm wavelength laser generator is selected,and the narrow band filter is used to filter the ambient light interference.Based on the structure of the visual elements and the size of the end of the robot,the fixture of the vision sensor is designed,and the test system of the weld reinforcement and width is set up.In order to establish the relationship between the pixel coordinates and the three-dimensional space,the vision system is calibrated.Firstly,the mathematical model of the camera is established,and calibration method for camera calibration proposed by Zhang is used to calibrate the camera,and the camera intrinsic parameters is got.And the planar checkerboard is used as the calibration reference.The structured light is calibrated,and the equation of the structure light plane is obtained.The hand-in-eye was calibrated using Two-Step method based on the camera model.The RANSAC algorithm is used to process the weld image and extract the weld reinforcement and width information.Firstly,the weld image is preprocessed to get the binary image.The edge image of the laser is obtained by edge detection,and the single pixel laser center line is obtained by the improved averaging method.The mathematical model can be obtained by RANSAC algorithm,and the ROI can be set automatically.The pre-processing and RANSAC algorithm are used again in this ROI.The single pixel laser center line only containing the weld height information can be obtained.In order to realize the robot polishing,the robot's position and orientation is defined.In order to realize the interaction between the visual information and the robot,the software of the weld height detection system is developed based on MFC and OpenCV.The software system adopts the method of modular,and develop the function modules such as image acquisition,image processing,welding data processing,visual system calibration and visualization.In order to verify the feasibility and stability of the system,the test of weld reinforcement and width is carried out.