A high dynamic range 3D shape reconstruction method for accurate exposure control

Abstract In industrial metrology, significant surface reflectance variations pose substantial challenges for three-dimensional (3D) shape reconstruction. While existing high dynamic range (HDR) measurement techniques can address this issue, their practical application is still limited by inefficient exposure optimization and complex fusion processes. In this paper, a novel HDR method is proposed to address these limitations. First, three sub-models are designed to determine the initial exposure range. Based on these sub-models, the system adaptively calculates the required exposure times for 3D reconstruction by analyzing pixel quality. After that, fringe patterns captured at the specified exposure times are fused using an adaptive strategy to generate high-quality HDR fringe patterns. Additionally, we chose the robust complementary Gray code as the coding method for 3D reconstruction. Experimental results demonstrate that our proposed method achieves comparable surface coverage (96.7% vs. 95.2%) and accuracy (0.049 mm vs. 0.055 mm) to the HDR with 10 exposures method, but only requires an average of 4–5 exposures per measurement.