A Particle Filter Infrared Target Tracking Method Based on Multi-feature Adaptive Fusion

In infrared target tracking it is difficult to predict the existence of background interference as there is a contradiction between the real-time and effectiveness of the algorithms. Hence this paper presents a particle filter tracking algorithm based on adaptive fusion of color features and edge features. Using the natural frame work of a particle filter, with infrared conditions it selects color features and edge features that best represent the target information to construct the multi-feature model of the target. According to the different feature separability of target and background, the weight of each feature component of a multi-feature is adaptively adjusted. This dynamic space model improves the particle filter tracking algorithm to predict the motion state of the particles to overcome the effects of environment mutations on tracking stability. Experimental results show that the proposed algorithm can overcome interference from all kinds of background clutter and noise ensuring tracking robustness and accuracy.