A stitching algorithm for automated surface inspection of rotationally symmetric components

This paper provides a novel approach to stitching surface images of rotationally symmetric parts. It presents process pipeline that uses feature-based create distortion-free and true-to-life image from video file. The developed thus enables, for example, condition monitoring without having view many individual images. For validation purposes, this will be demonstrated in the using concrete example worn ball screw drive spindle. algorithm aims at reproducing functional principle line scan camera system, whereby physical measuring systems are replaced by approach. evaluation algorithms, metrics used, some which have only been work or supplemented test procedures already use. applicability is not limited machine tool spindles. Instead, method allows general inspection various components can therefore used variety industrial applications. Deep-learning-based detection Algorithms easily implemented generate complete failure on