Object-to-track association in a multisensor fusion system under the TBM framework

Due to the increased interest in multiple-objects tracking, various methods have been recently proposed and applied in different applications such as: pedestrians identification and tracking, road vehicles detection and tracking, airplanes classification and tracking, etc. However, in presence of inter-object occlusion and sensor gaps, most of these methods result in tracking failure due to object-to-track association failure. This paper presents a new algorithm on object-to-track association in multi-sensor fusion systems under the transferable belief model framework. The proposed approach quantifies the belief on associating each detected object to each existing track, and takes into consideration the creation of new tracks by the non-associated objects.