基于光流法与水平集方法的运动对象分割

Both Horn-Schunck optical flow model and level set method are analyzed in this paper, and Chan-Vese segmentation model is mentioned. This paper presents an approach based on optical flow and level set method for segmenting the motion objects. In order to segment the image plane into a set of objects containing different information on the basis of only two frames from an image sequence or video. This approach combines Horn-Schunck model with Chan-Vese segmentation model, and evolutes level set based on the different optical flow fields of different objects. Then the different objects are segmented by level set. This paper detailed introduces the two phase’s motion segmentation and the multiphase motion segmentation’s energy function of two level sets and its computational method. To this end, the results of the two phase’s motion segmentation and the multiphase motion segmentation of two level sets are given. Experimental results show that this approach can segment the motion objects from the image sequence or video.