Ordered Boundary Extraction Algorithm of 3D Point Cloud for Guiding Robot Spraying

Abstract

The target digital model is provided for the robot controller through the three-dimensional reconstruction technology when the robot is guided by the visual system for intelligent spraying. In order to acquiring the reconstruction model of the target, the point cloud data obtained by the 3D vision sensor is processed by the 3D reconstruction technology. As an important geometric feature of the reconstruction model, the accurate extraction of the ordered boundary will be the core of the3D reconstruction technology. It is found that the current research on 3D reconstruction technology is generally to construct triangular meshes for discrete point clouds. However, in practice, the point clouds obtained from binocular stereo vision have matrix structure and can be directly converted to triangular meshes. However, the results contain large randomness and error when using grid characteristics to sort complex grid boundaries. To solve the above problems, we propose an integrated boundary extraction algorithm based on binocular stereo vision point cloud. The point cloud with matrix structure collected by binocular stereo camera is converted into point cloud with triangular mesh structure and the boundary is extracted according to the relationship between the edge and the triangle in the triangular mesh. It is stipulated that the searching of boundary must be found by the path containing the triangle in the boundary sorting so that the boundary search has a unique solution and provide a method to obtain an accurate and orderly boundary.