Particle filter based moving object tracking in dynamic scene

A adaptive algorithm of tracking moving object using particle filter in dynamic scene is proposed in this paper, which based on the work of reference [17] on object detection. At first, we apply six-parameter affine transformation to building motion parameter model and adopt the detection methods for moving object in [17]. Then, Color and edge features are used to describe object model. An adaptive weighted distribution algorithm is used in fusion of object features. According to the weights distribution, we apply particle filter algorithm combined with two object features to tracking the moving object. Experimental results show that the method presented in this paper improves the accuracy of tracking for moving target in dynamic background.