2012/09/19 by Prabhu Natarajan, Trong Nghia Hoang, Natarajan, Prabhu +6 · 1 citation
Computer Science · #Artificial Intelligence (cs.AI) #Distributed Control Multi-Agent Systems #FOS: Computer and information sciences #Multiagent Systems (cs.MA) #Multimedia (cs.MM) #Robotic Path Planning Algorithms #Robotics (cs.RO) #Video Surveillance and Tracking Methods
paper · pdf · doi:10.48550/arxiv.1209.4275
openalex publication_date 2012/09/19 · openalex created_date 2022/08/30 · openalex updated_date 2026/07/28
A central problem of surveillance is to monitor multiple targets moving in a\nlarge-scale, obstacle-ridden environment with occlusions. This paper presents a\nnovel principled Partially Observable Markov Decision Process-based approach to\ncoordinating and controlling a network of active cameras for tracking and\nobserving multiple mobile targets at high resolution in such surveillance\nenvironments. Our proposed approach is capable of (a) maintaining a belief over\nthe targets' states (i.e., locations, directions, and velocities) to track\nthem, even when they may not be observed directly by the cameras at all times,\n(b) coordinating the cameras' actions to simultaneously improve the belief over\nthe targets' states and maximize the expected number of targets observed with a\nguaranteed resolution, and (c) exploiting the inherent structure of our\nsurveillance problem to improve its scalability (i.e., linear time) in the\nnumber of targets to be observed. Quantitative comparisons with\nstate-of-the-art multi-camera coordination and control techniques show that our\napproach can achieve higher surveillance quality in real time. The practical\nfeasibility of our approach is also demonstrated using real AXIS 214 PTZ\ncameras\n