CAJ | 학술논문

当前，智慧城市成为信息时代城市建设的一个基本目标，智能视频安防监控是其中重要一环，希望从视频图像提取出有效的信息，提供有效的治安防控业务信息。由于视频监控系统广泛使用于各行各业，监控视频数据已成为一类典型的大数据，因此，如何对监控视频大数据进行高效的处理成为一个重要挑战。为此，本文在分析视频处理特点的基础上，提出并实现了一种基于Hadoop MapReduce计算框架的分布式离线视频处理方法，该方法根据视频处理的特点进行优化，提升了监控视频大数据的处理效率。
당전，지혜성시성위신식시대성시건설적일개기본목표，지능시빈안방감공시기중중요일배，희망종시빈도상제취출유효적신식，제공유효적치안방공업무신식。유우시빈감공계통엄범사용우각행각업，감공시빈수거이성위일류전형적대수거，인차，여하대감공시빈대수거진행고효적처리성위일개중요도전。위차，본문재분석시빈처리특점적기출상，제출병실현료일충기우Hadoop MapReduce계산광가적분포식리선시빈처리방법，해방법근거시빈처리적특점진행우화，제승료감공시빈대수거적처리효솔。
Nowadays, smart city becomes a basic aim of city construction in the age of information, and the intelligent security monitoring is one of the most important parts. In the typical intelligent video surveillance systems, the useful information is extracted from video images for providing security service effectively. Because the video surveillance systems are widely used in all walks of life, efifcient processing of big surveillance video data has become an important challenge. Therefore,this paper designs and implements a Hadoop-based distributed oflfine video processing method. This method can signiifcantly improve the processing efifciency of big video data through optimizing the MapReduce framework based on the video analysis features.