High precision calibration strategy and algorithm on binocular visual collaborative tracking

摘要

The binocular visual collaborative video tracking method can simultaneously obtain Far-Range image in global field of view and Close-Range image in following field of view with which can track the target, but the error of PTZ(Pan Tilt and Zoom) motion parameters of the following camera is relatively large and the calibration algorithm is relatively complex. Thus, a new strategy for PTZ calibration is presented, with which the PTZ motion parameters of the following camera are calibrated according to different positions of the global field of view, and the mapping relationship between the target position and the PTZ motion parameters is calculated so that the target locates the central position and displays in zoom-in state in monitoring field. Field experiment results show that the motion parameter error is obviously reduced while the precision of calibration is greatly improved.