Effective 3D Data Extraction for Multiple Structured Lights Projection Method

Abstract: Computer vision systems have been exploited to detect the three dimensional shape data of objects. Slit-laser scanning or pattern projection on the recording process is executed on these system. Generally, since the recording process takes a time, the temporally stop of physical movement is required for both of the measurement target and the measurement device during the recording process. In order to cope with this problem, the measurement system that projected the multiple structured light is introduced. Multiple structured light projection enables the quick measurement and three-dimensional data on these structured lights can be obtained by a few images. However, in the image processing, each structured light should be separated and identified properly for estimating the three-dimensional position by the triangulation. In this paper, separation method of each structured light from the single recorded image is introduced, and this method improves the robustness of pattern recognition.