A two-point spatial mapping method for hybrid vision systems

摘要

We propose a new spatial mapping method for hybrid vision systems consisting of an omnidirectional camera and an active perspective camera. We calculate the relative position between the optical centers of the two cameras based on pre-acquisition parameters of two sample points from the observed scene, avoiding the difficulty and error in direct measurement, and at the same time using fewer points. We derive the formula of the relative position between the omnidirectional camera's optical center and the active camera's optical center from the distance between two points, and give a method for calculating the corresponding pan and tilt angles of the active camera from the coordinates in the omnidirectional image. Simulation and real experiments show that the two-point spatial mapping method has higher estimation accuracy than the homography calibration method, and at the same time has higher consistency among test points.