Rapid 3D surface profile measurement of industrial parts using two-level structured light patterns

This paper aims to present a rapid 3D shape measurement system based on novel monochromatic structured light patterns. The system consists of projectors shooting the fringe patterns onto the inspected parts and cameras recording the corresponding distorted images. Using the two-level fringe patterns, the correspondence between the projector image and camera image can be established with sub-pixel accuracy. The two-level pattern is based on three spatiotemporal binary stripes, in which the value of the stripe boundary (first-level coding) is determined by the two adjacent stripes patterns over time and the codeword of the strip boundary (second-level coding) depends on its values and neighbor boundary values in space. The proposed pattern is robust to ambience light variation and part texture. Moreover, the occlusion can be overcome and high density measurement can be achieved. Experiments with different 3D parts are conducted to evaluate the robustness and accuracy of the inspection system using the two-level patterns. The results show that the system has desired properties of high accuracy, high density, rapid acquisition, and robustness, which are essential for industrial application.