Research on Multi-camera Collaborative Target Tracking Method Based on Deep Learning

• 1、School of Geographic Sciences,East China Normal University,Shanghai 200241,China
• 2、Shanghai University of Sport, School of Economics and Management, Shanghai 200438，China

Abstract： At present, the general public safety monitoring system adopts a scheme of deploying a large number of camera devices in the monitoring area to achieve monitoring of a large area. These cameras are usually discrete, and the captured monitoring information lacks a fast and comprehensive integration mechanism, resulting in the processing of video information based only on images, and ultimately fails to effectively combine the spatial position information of the video surveillance system. In view of the shortcomings of current video surveillance systems, this paper proposes an improved scheme for video surveillance systems in terms of target detection, target location and multi-camera collaborative tracking based on photogrammetry,Research on Multi-camera Collaborative Target Tracking Method Based on Deep Learning GIS and deep learning. In this way, the target detection and tracking in the multi-camera video surveillance scene is achieved. Finally, the effectiveness of the improved multi-camera collaborative target tracking method for real-time tracking of specific targets is verified by experiments.

Keywords： intelligent video surveillance system multi-camera collaboration coordinate mapping target detection target trackingkey