Evaluation of Morphological Operators in Gaussian Mixture Model for Moving Object Detection

Moving object detection is fundamental technology for intelligent video surveillance. The Gaussian Mixture Model (GMM) is one of the most commonly used algorithms for moving object detection, but it does not cope with complex scenes efficiently. Thus, morphological operators are always in consideration after applying GMM. This study is to evaluate the influence of morphological operators on the moving object detection result based on the GMM. In our experiment, the open operation is used to remove the small noises outside the silhouette, and then the close operation is used to fill the small holes inside the object region. We evaluated the detection results of different complex scenes with respect to precision and speed. The experimental results show that the performance of the Gaussian mixture model algorithm with morphological operators is inconsistent based on different complex scenes, it is more useful in some specific scenes than some others.