基于灰度和直方图的阈值自适应镜头边界检测  

Gray Value and Histogram Based Shot Boundary Detection with Adaptive Threshold

作　　者：黄茜[1] 张海泉[1] 杨文亮[1] 吴元 

机构地区：[1]华南理工大学电子与电信学院,广州510640 [2]School of Informatics,University of Bradford

出　　处：《科学技术与工程》2008年第14期3787-3792,共6页Science Technology and Engineering

摘　　要：镜头是对视频内容进行分析和检索的基础,研究镜头边界检测具有重要的现实意义。在对视频以及视频镜头边界检测的现有方法分析后,提出了一种改进的镜头边界检测算法——基于灰度和直方图的阈值自适应镜头边界检测算法。该算法是在应用自适应阈值的基础上,结合灰度图和直方图的特征总结出来的。实验结果表明这种改进的算法具有很好的突变镜头边界检测效果,同时该算法还对不同性质的视频有很好的适用性。Video shot is the basic of video analysis and retrieval,so it＇s meaningful to research the methods for shot boundary detection, have some general descriptions of the video and current video shot boundary detection algorithms. Then an improved algorithm is presented, which is based on the gray value and histogram. The improved algorithm uses adaptive threshold and summarize the feature of gray values and histogram. The experimental result shows that the improved algorithm can detect abrupt changed shot boundaries well and also has good applicability of many kinds of videos.

关 键 词：镜头边界 突变 特征提取 自适应阈值 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]