Stereo deflectometry based automatic segmentation technique for measuring structured specular surfaces

Abstract

Techniques have been studied to extend the measurement capability of stereo deflectometry from continuous specular surfaces only to complex structured specular surfaces. Segmentation is a key process in these techniques, which separates a measured structured surface into several continues segments with gradient and coordinate data to achieve high form measurement accuracy. However, it is a great challenge to achieve automatic, fast, and accurate segmentation since specular surfaces have no RGB (red, greed, and blue) texture to be used in the segmentation process. In this paper, we present an automatic segmentation technique based on the characteristics of gradient variation and 3D coordinate data of the measured structured specular surfaces. Different segmentation strategies are applied and discussed for the measured discontinuous specular surfaces and continuous non-differentiable specular surfaces. Compared with the existing segmentation methods in terms of stereo deflectometry, the proposed technique shows significantly advantages in automation, speed, and flexibility. Experimental results verified the effectiveness and accuracy of the proposed method in the measurement of structured specular surfaces.