摘要
为了克服光线变化和运动对视频镜头检测的影响,本文提出了一种新的视频镜头检测算法.该算法把视频图像从RGB颜色空间映射到HSV颜色空间,再重新量化图像的H与S分量到127个色度区域,最后通过比较帧间图像在色度区域的重叠程度来检测镜头的变化.同时采取有效的虚假镜头分割分析机制以减少图像快速变化的影响.实验结果表明,本算法具有较高的准确率和查全率,并且有较快的处理速度,在视频镜头分割检测领域具有一定的实用价值.
To overcome the interference produced by illumination changing and movement during shot detection, this paper proposed a new method for shot segmentation. It firstly transforms the video pictures from RGB space to HSV space,then requantitates the H and S components of pictures to 127 chroma areas,finally detects the shot＇s change by computing the coincident degree in chroma areas between frames. In addition,it adopts a false-shot-analysis strategy to weaken the interference brought by fast change of pictures. The experiment results indicated that the algorithm had high accuracy and recall ratio,as well as fast processing speed. So it is suitable for shot segmentation.
出处
《厦门大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第5期665-668,共4页
Journal of Xiamen University：Natural Science
基金
国家863计划项目(2006AA01Z129)
国家自然科学基金(60672018)
厦门大学"985"二期信息创新平台项目(0000-X07204)资助
关键词
HSV颜色空间
镜头分割
镜头检测算法
HSV color model
shot segmentation
shot boundary detection algorithm
