Structured Light Based High Precision 3D Measurement and Workpiece Pose Estimation

This paper proposes a high precision, robust 3D measurement method and a fast, accurate pose estimation method. The RGB-based Gray code Shifting and Positive/Negative coding method (RGB-GSPN) is introduced to improve the structured light measurement system. This coding method is easy to be decoded. It can eliminate the influence of scattered light and reduce the influence of the objects surface color on the measurement precision. Then, we combine 2D and 3D information to perform pose estimation. SIFT features of 2D images are used to roughly match the position. Subsequently, we send this pose to the Iterative Closest Point (ICP) algorithm as its initial matching position and use ICP for accurate estimation. This can avoid getting into a local optimum and speed up the convergence of the algorithm. Experiments show that the system has a high precision and can estimate the pose of the workpiece accurately.