水平转动摄像头下的移动物体切割.pdfVIP

Abstract  Abstract With the extensive application of multimedia, more and more things shown by video, the research about video is increasingly becoming a hot topic. Segmentation, as one of the fundamental problems in digital video processing, which provides the basis for other processing and application, is more important. In the past, videos of segmentation almost are captured by a fixed camera, and segmentation technology can be divided into three basic methods, i.e. Background Subtraction, Frame Difference and Optical Flow. Background Subtraction, with the background to segment the foreground, is widely used, and many methods improve the procedure used to update the background model. Frame Difference does not define the background model, but rather uses the neighboring frame to segment, then gets the boundary of the foreground. So the result of Frame Difference needs to fill the black part. Optical Flow assumes that the intensity of moving object remains constant along a motion trajectory. Because of relative motion between the foreground and the background, Optical Flow can segment the foreground and background through calculating motion vector between two successive frames. For the moving camera, these methods can’t segment effectively the foreground from the background. With the movement of the camera, the background also changes, but the difference between the background and the foreground, which is caused by relative motion, does not change. According to the basic segmentation methods above, the proposed algorithm in this paper extracts the motion vectors from the video, and calculates the means and variance of the