An Adaptive Multiexposure Scheme for the Structured Light Profilometry of Highly Reflective Surfaces Using Complementary Binary Gray Code

The structured light profilometry often suffers from local overexposure and underexposure, resulting in measurement error while applied to highly reflective objects. To address this problem, we introduce an adaptive multiple exposure imaging and fusion scheme to generate desired decoding fringes even with the drastically varying reflective surface. Here, a complementary gray code is utilized in our profilometry to generate the projection stripes, and the responses from the binary positive and negative stripes are first exploited to define a unique image formation model for the structured light system. Based on the model, an optimal exposure sequence is calculated for a specific target using an adaptive mechanism, where the probing exposure is iteratively increased according to a prior difference threshold and some stop criterion. The projected stripes on the target surface are then captured at the optimal exposure time and the high dynamic range (HDR) images are generated by a specifically designed image fusion algorithm which is guided by the difference between the positive and negative stripes. Thus, these HDR response images could be decoded as the common gray code patterns to generate the 3-D profile measurement. It is demonstrated in a comparative experiment using real workpieces that the proposed method could outperform the existing HDR methods in terms of accuracy and completeness of obtained profiles.