Automated Exposures Selection for High Dynamic Range Structured-Light 3-D Scanning

Abstract

Structured light (SL) scanning has gained widespread applications in industrial metrology; however, measuring objects that have large surface-reflectivity variations is challenging. For this purpose, the high dynamic range (HDR) technique has become a standard method to fuse features under multiple exposures. Presently, in HDR measurement, manual or cumbersome exposure selection strategies are used. This paper reports a new exposure selection approach. An image quality metric is designed to evaluate captured images and suppress regional overexposure, based on which a multiple-exposures selection strategy is developed. Unlike existing methods for single-optimal exposure selection, in our approach, pixel-filtering is introduced to locate eligible pixels and candidate pixels in next exposures. Experimental results with industrial parts having large surface-reflectivity variations demonstrated that compared to HDR with 15 exposures (equal time intervals), our proposed approach achieved a similar surface coverage rates (97.4% vs. 98.1%) and measurement accuracy (0.043 mm vs. 0.041 mm), but only took 3-4 exposures averagely.