Determining the 3D position and orientation of a robot camera using 2D monocular vision

摘要

An algorithm for determining the three-space position and orientation of a robot camera relative to a rectangular object of known size has been developed. Given the image plane positions of the vertices of the rectangle, each projected diagonal of the rectangle is isolated and processed in turn to find the 2D position of the camera in the plane that contains the diagonal and the line between the centroid of the mark and the focal point of the camera. The two sets of results are then combined using heuristics based on human vision and processed to determine the unique three-space position and orientation of the camera relative to the rectangle. No a priori constraints are placed on the positional or orientational relationship between the camera and the rectangle.