A Method Of Partition Path Planning For Laser Cleaning Based On Machine Vision Detection

In the industrial field,in order to prevent the mechanical properties of metal surface from being reduced due to corrosion,surface coating is often used to achieve the purpose of protection.When the paint layer is partially damaged or peeling,it is necessary to periodically remove the original damaged surface paint layer and recoat the paint layer to maintain the mechanical properties.Compared with mechanical,chemical,and ultrasonic paint removal methods,which have many disadvantages of,such as,heavy pollution,loud noise and obvious damage,the technology of laser cleaning has been widely used because of its advantages of no contact,no pollution,low energy consumption,high safety and little damage to health,etc.In order to achieve high quality and efficient laser cleaning,it is necessary to develop a reasonable scanning path planning method.In view of the problem of removing local damaged paint layer on workpiece surface,it is an effective method to improve the cleaning quality and efficiency to implement the laser partition cleaning in high efficiency.Using the method of machine vision detection to capture the surface image of workpiece,in the meantime,quick and accurate identification of local damaged paint layer area contour is the premise of automatic partition cleaning.However,in the process of image acquisition,due to the low-exposure of the camera,the defect of low illumination appears in the captured image,which makes it difficult to identify the contour of the paint layer.Therefore,a method of partition path planning for laser cleaning based on machine vision detection is proposed in this paper;the model of low-light image enhancement method is established to restore the color and brightness information of paint image effectively;the threshold segmentation method of paint image is proposed to realize partition recognition and regularization;the method of partition scanning path planning is established;the surface cleaning quality of different scanning paths is analyzed.The main research contents and conclusions are as follows:(1)To solve the problem of color distortion in low-light image enhancement,a camera response model is adopted;aiming at the problem of low illumination of the enhanced image of camera response model,the normalization method of brightness based on mean grayscale of color three channels is adopted.PSNR、SSIM、MSE、LOE and MG of enhanced image are19.638、0.870、706.809、215.411 and 1.100,respectively,which prove that the image quality has been improved and the problems of color and brightness distortion in low-light image enhancement has been solved(2)Aiming at the problem of complex contour precise identification of paint layer area,a method of contour identification based on chroma coordinate value is proposed;the experimental results show that this method achieves accurate identification of complex contour of paint layer area,and improves the identification accuracy compared with the traditional iterative method.In order to realize laser partition cleaning in high efficiency,a method of complex contour regularization of paint layer area based on minimum envelope rectangle is adopted;in order to improve cleaning quality,a method of generation scanning path in variable distance is used;a method of partition scanning path planning is established(3)The scanning paths of spiral,parallel reciprocating and contour offsetting are used to clean the surface paint layer of the samples,and the surface cleaning quality of spiral scanning path is compared with that of parallel reciprocating and contour offsetting scanning path,the surface microstructure,hardness,roughness and shape variables of different scanning paths are obtained and then analyzed;the experimental results show that the sample shape variable of the spiral scanning path is smaller,and the surface cleaning quality of the spiral scanning path is better.