文章基本信息

标题：Stain Detection in Video with Background Restructured
本地全文：下载
作者：Wang, Jia ; Li, Xin ; Liu, Shuai 等
期刊名称：Journal of Multimedia
印刷版ISSN：1796-2048
出版年度：2014
卷号：9
期号：7
页码：963-969
DOI：10.4304/jmm.9.7.963-969
语种：English
出版社：Academy Publisher
摘要：Nowadays, multimedia, especially video, is used widely in many domains. But camera lens are easily stained because they are always outside, such as in traffic and monitoring. Admittedly, the video from these cameras is out of action because it cannot provide clear frames. In this paper, we provide a detecting method from video with background restructure. This method is presented to detect stains in video. Firstly, the first video frame is used as a pseudo-background because original cameras are clear with no stains. Then, background is constructed by the pseudo-background and the continuous frames which are from the video. Moreover, with training in the frames, we restructure background by dropped moving objects in it. Finally, after background restructured, we detect stains in camera lenses by the restructured background, and provide positions of stained area. Experimental results show its robustness and practicability
关键词：Background Restructured;Background Robustness;Stain Detection;Video Stain;Practicability