Vehicle Tracking Based on Fuzzy C-Means Clustering and Grey Model

To detect the moving vehicles correctly, we proposed a vehicle tracking method based on background model, fuzzy C-Means clustering and Grey model. In this paper, the background image is updated adaptively to accommodate vehicle detection in variable outer environment. The Grey model is employed to predict the center and number of C-Means clustering, solving the dependency of fuzzy C-Means upon clustering center of initial frame. The result of clustering provides updated data of Grey model for the position of target in the next frame. The experimental results show that, the method proposed improves both of the speed and accuracy of C-Means clustering, and promote the robustness and real-time of vehicle tracking system.