视频摘要系统的研究与实现

Research and implementation of video summarization technology

基于安防领域需要快速检索监控视频中嫌疑人信息需求。本文设计了一种视频摘要系统,深入分析了该系统的设计目标和组成模型。在具体实现方面,该系统通过结合目标形状特征和运动特性来对目标进行分类,并依据目标的运动轨迹提取视频关键帧的方法合成视频摘要。系统完成后通过导入监控视频文件生成摘要验证测试,该视频摘要系统能在69.44%压缩比的前提下达到零误报率。实验结果,该视频摘要系统压缩比高,识别误差小,能极大提高检索视频中信息的效率。

Based on the demand of rapid retrieval of the suspect＇s information in surveillance Video. A video abstracting system is designed in this paper. The system design goals and component model are also deeply analyzed. In terms of concrete implementation, the system classify the targets by the shape and movement characteristic, and the video abstract is synthesized through the method of extracting video key frames based on the trajectory of the target. The abstract verification test is generated by importing surveillance video files after the system competed. This video abstracting system can achieve zero false positive rate in the context of a compression ratio of 69.44%. The experiment results show that the video abstracting system has high compression ratio and little identification error, can greatly improve the efficiency of retrieving video information.