Nonlinear Diffusion and Structure Tensor Based Segmentation of Valid Measurement Region from Interference Fringe Patterns on Gear Systems

Xian Wang,Suping Fang,A. Kubo,Pengcheng Yang,Xindong Zhu,M. Komori,Jing Ji
2017-12-25
Abstract:The extraction of the valid measurement region from the interference fringe pattern is a significant step when measuring gear tooth flank form deviation with grazing incidence interferometry, which will affect the measurement accuracy. In order to overcome the drawback of the conventionally used method in which the object image pattern must be captured, an improved segmentation approach is proposed in this paper. The interference fringe patterns feature, which is smoothed by the nonlinear diffusion, would be extracted by the structure tensor first. And then they are incorporated into the vector-valued Chan-Vese model to extract the valid measurement region. This method is verified in a variety of interference fringe patterns, and the segmentation results show its feasibility and accuracy.
Engineering,Materials Science,Computer Science
What problem does this paper attempt to address?