Integration of autofocus and object tracking in an infrared stereo vision-based video surveillance system with multi-lens module

The main purpose of this paper is to develop an infrared stereo vision-based video surveillance system with multi-lens module. The system consists of two high-resolution color cameras, two infrared cameras, a color zoom camera, and a platform with two rotating axes. Based on a novel stereo vision algorithm, the system not only can improve the accuracy of depth estimation, but also can convert a stereo image pair into a disparity map using the speed-up robust features algorithm, then establish the 3D image reconstruction of the monitoring field. In addition, the stereo vision algorithm and an autofocus search method are combined into a hybrid autofocus approach to estimate zoom in ratio, so the system has the ability to track and provide more clear images of moving objects. Finally, a 3D thermal imaging technique is applied to construct 3D thermal images of the tracked moving object and the monitoring field. Experimental results of this paper have demonstrated the applicability of the proposed system.