Free-Form Surface Analysis and Linking Strategies for High Registration Accuracy in Quality Assurance Applications

Abstract

Robot-based optical inspection systems have become more affordable and show great potential for measuring geometric features of sheet metal parts in production sites. To deal with the lack of positioning accuracy of current industrial robots for the registration of range images, a stitching algorithm is often applied. This paper presents a concept to improve the fine registration accuracy in order to reduce the effect of error propagation. A geometric analysis of the surface is introduced to identify suitable regions for applying adapted texture projections onto the part. In addition, an approach to develop a corresponding linking strategy is described. The proposed concept promises to measure geometric features on both plane and highly featured surfaces with high accuracy.