【摘要】 在许多视频相关领域需要做这样的工作:将视频中的感兴趣内容进行提取,比如视频编码标准MPEG-4和MPEG-7的前处理,交通视频流的处理和安全监控等。运动对象的视频分割就是为了解决这些需求。介绍了视频分割的应用前景及主要方法。然后详细介绍了固定背景下的视频分割,先将视频中的每帧进行通道变换、计算获得背景,通过变化检测法,提取前景得到初步结果。再经过中值滤波和形态学膨胀的处理,对视频流中分割出的运动对象使用矩形框进行标定。可以看出分割得到了较好的结果。更多还原

【Abstract】 When the new standard of video code-MPEG-4 and MPEG-7,processing of the traffic flow,and safe monitoring,all need the ROI of the video to be abstracted,such as the moving cars in the traffic flow and people in the movies.And video-segmentation is the technology to resolve these problems.In the beginning,introduces the prospect and main methods of the video-segmentation.Then it shows the details of video-segmentation in the fixed background: first,each frame is converted to single channel;second,the background is obtained by the accumulator;and then,foreground is extracted by the method of change detection.In the end,the median filter and dilation are used to improve above results.At the same time,each moving car from the segmentation is surrounded by the rectangle,and the rectangle is auto-adapted to the size of the car.The result shows the good effect of the segmentation.更多还原