基于Mean Shift算法的运动目标检测

Mean Shift, which is widely used in many fields of image processing during recent years, is an adaptive gradient ascending algorithm used to seek the local maximum. Moving object detection is one of the basic techniques in computer vision. Result of detection has great influence on post processing steps such as target classification, object tracking and behavior understanding. In this paper we use a novel Gaussian mixture model based on Mean Shift to represent the background. And corresponding algorithms of model updating, object detection and result correction are also designed. Experiments show that the proposed method performs well in different scenarios.