中文摘要：

视频结构化组织是建立视频检索和浏览系统的基础,而镜头边界检测是视频结构化的第一步．在本文中,我们提出了基于基色调的镜头边界检测方法．该方法首先利用I帧的比特数信息,减少参与检测的Ⅰ帧：然后在考虑基色调的基础上得到局部的自适应阈值：最后用双阈值方法检测镜头边界．实验证明,该方法可以对长视频序列进行较好的镜头突变和渐变检测,并且减少了计算量,同时能够排除大物体运动和摇镜头对镜头检测的影响．

英文摘要：

Shot boundary detection is one of the critical techniques in video browsing and indexing systems. In this paper, we propose a main color based shot boundary detection algorithm on DC images. We use the bit-rate information to decrease the number of I frames that take part in the shot change detection. Then we obtain local adaptive thresholds based on the main color of a segment of the video sequence. Finally, we detect shot change based on twin-comparison method. Experimental results show that our algorithm can detect shot boundary from a long video sequence with reduced computation, and at the same time solve the problems that are caused by object moving and camera panning in shot change detection.