监控视频运动物体元数据并行提取系统分析-analysis of parallel extraction system for metadata of surveillance video moving objects.docxVIP

AbstractWith the promotion of projects like Peace City, Safe Campus, etc., the number of surveillance videos increase dramatically. Using human resource to handle the video can not reach the requirement. In order to better analyzing the surveillance videos by using computer, metadata of the moving objects, including size, location, color and some other information, should be extracted and stored. Therefore, the extraction of metadata for moving objects has great importance. Currently, metadata extraction of moving objects in surveillance videos relies on the traditional algorithm o f moving object tracking, and the research mainly focus on the issue of how to improve the precision of metadata extraction. However, as surveillance video is characterized with long surveillance time and large data size, it will result in low efficiency o f metadata extraction for insisting on the traditional serial iterative algorithm, which will influence the subsequent processing.A parallel metadata extraction system for moving objects is proposed to improve the speed of metadata extraction. Firstly, split the video to make every video segment require the same time to extract metadata. However, there exists the problem of latter metadata extraction depends on previous metadata. To solve this problem, metadata is integrated on basis of the moving objects appearing mostly in two segments. Different object keywords are set based on the relationship among video segments. Then extract feature to establish the distance matrix. At the same time, calculate overlap area to solve the occlusion problem specific to surveillance video. Using the above method, parallel metadata extraction system can be realized to accelerate the metadata extraction.Based on Linux and Hadoop, parallel metadata extraction system for moving objects in surveillance videos was completed through C++ and openCV library. The experimental result shows that the precision of parallel metadata extraction system is higher than