Object tracking algorithm based on particle filter with color and texture feature
Dong-Sheng Ding, Zengru Jiang, Chengyuan Liu · 2016
In this paper we present an improved particle filter algorithm for object tracking which combines color and texture features of the object. This new algorithm can overcome the influence of the following factors effectively such as nonlinear motion of the target, occlusion and illumination change. We describe the target feature by the improved color histogram and LBP texture histogram, and establish the reference model of the target. A weighted fusion method is used to realize the linear fusion of color and texture features. In order to adapt to the change of target and environment, an adaptive updating strategy of target template is designed, and the number of particles is dynamically adjusted when the target is seriously disturbed. Experimental results show that compared with the basic particle filter algorithm, the new algorithm proposed in this paper has better robustness and can achieve stable tracking in the presence of occlusion and illumination changes.