Rough Surface Wear Analysis Using Image Processing Techniques

Due to the great interest of the automotive industry to quickly and systematically evaluate low wear quantities a new approach based on image processing is presented. By comparing the captured rough surface point cloud using optical interferometry before and after wear, the proposed algorithm was able to estimate the amount of volume wear and mass loss. The proposed algorithm has two main steps. In the first step, the 3D point clouds obtained before and after wear are mapped to 2D images, and conventional edge detection algorithms are applied. Keypoints are determined using the SIFT algorithm on both images (before and after the wear), followed by the use of the RANSAC algorithm to create a good 2D registration between both 2D images. In the second step an enhanced ICP algorithm is used to perform a 3D registration. The main advantage of this method is the visualization of the wear geometry provided by the wear topography. The proposed algorithm was tested and the results are presented.