Moving targets visual tracking in complex scenes based on PCR6 combine rules

The aim of this article is to investigate multi-moving targets visual tracking in complex scenes based on PCR6 (proportional conflict redistribution 6) combine rules, and improve the poor tracking performance in complex scenes. A tracking model of multi-moving targets was established by combining the color, edge and texture features of the targets, and the corresponding tracking algorithm was designed based on the framework of PF (particle filters) and PCR6 rules. The tracking process of moving targets including different scenes of mutual occlusion, proportion or illumination change was analyzed to validate the reliability and stability of the introduced method. The results show that the number of particles is significantly reduced, which helps to decrease the computational complexity and storage cost for tracking multi-targets of complex scenes. Meanwhile, the adaptive ability of fusion high conflict evidences is improved, and the multi-targets tracking performance is greatly elevated based on bad tracking surroundings. The research will further extend the applied scopes of evidence theory for PCR6 combine rules, and will meets the practical demand of multi-targets tracking in complex scenes. Especially, it has very important engineering application value for improving the artificial intelligence algorithm of visual tracking.