Collaborative observation algorithm for multiple moving targets using a PTZ camera assisted by a wide-angle camera

Abstract

To deal with the contradictions between camera field of view and resolution in visual observation, a collaborative observation algorithm for multiple moving targets using a Pan-Tilt-Zoom camera assisted by a wide-angle camera is designed. Firstly, the low resolution images are captured by the wide-angle camera, based on that multiple moving targets are detected, tracked and predicted. Then, for the multiple observation requested generated from the multiple targets, a visual tracking algorithm for the PTZ camera is developed. The factors, such as the image resolution, vision coverage range and response to tracking requests, are taken into account in the algorithm, and the desired pose of the PTZ camera is worked out optimally. Finally, the vehicle tracking experiments both in a simulated airport and actual traffic scenes are carried out, and the experimental results demonstrate that the proposed algorithm in this paper is accurate and time-efficient.