Abstract
Most recent works on multi-target tracking with multiple cameras focus oncentralized systems. In contrast, this paper presents a multi-target trackingapproach implemented in a distributed camera network. The advantages ofdistributed systems lie in lighter communication management, greater robustnessto failures and local decision making. On the other hand, data association andinformation fusion are more challenging than in a centralized setup, mostly dueto the lack of global and complete information. The proposed algorithm booststhe benefits of the Distributed-Consensus Kalman Filter with the support of are-identification network and a distributed tracker manager module tofacilitate consistent information. These techniques complement each other andfacilitate the cross-camera data association in a simple and effective manner.We evaluate the whole system with known public data sets under differentconditions demonstrating the advantages of combining all the modules. Inaddition, we compare our algorithm to some existing centralized trackingmethods, outperforming their behavior in terms of accuracy and bandwidth usage.