Improved Algorithm of Adaptive Gaussian Mixture Model for Moving Target Detection

The background modeling algorithm based on Gaussian mixture model (GMM) is a widely used method in moving objects detection with static cameras. Base on the situation that traditional Gaussian mixture model is very sensitive to sudden illumination variation and is slow for convergence speed, this paper proposed a method to detect the illumination variation and update the single learning rate, in order to build the adaptive updating background model. Using the algorithm of color histogram matching, the proposed method can adaptively adjust the learning rate by introducing the illumination variation factor and the counter for model parameters updating. Meanwhile, adaptive adjustment of the number of GMM components reduced the computational cost and improved the real-time performance.Experimental results show that this proposed approach can adapt scene changes efficiently, and has better accuracy and robustness than traditional Gaussian mixture model.