An Improved Mixture Gaussian Models to Detect Moving Object Under Real-Time Complex Background

With video inspecting applied, many research efforts have focused on subtracting the moving object. In this paper, we propose an improved Mixture Gaussian Model to subtract the real-time complex background. The experimental results have testified the proposed method can greatly extract the background model from the actual scene through the real-time monitor, extract the moving vessels automatically, filtrate the video abnormal incidents occurred, and achieve the monitoring the dock at no duty of intelligent monitoring.