首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于改进的VIBE算法在铁路智能视频监控入侵检测的研究
引用本文:杨海林,蔡启仲,薛圣利,周加全.基于改进的VIBE算法在铁路智能视频监控入侵检测的研究[J].科学技术与工程,2016,16(21).
作者姓名:杨海林  蔡启仲  薛圣利  周加全
作者单位:广西科技大学 电气与信息工程学院;深圳市普禄科视频技术有限责任公司,广西科技大学 电气与信息工程学院,广西科技大学 电气与信息工程学院,广西科技大学 电气与信息工程学院
摘    要:视觉背景提取(visual background extractor,Vi BE)算法应用在铁路沿线复杂背景环境下的入侵检测存在比较明显的问题,即Vi BE算法里的参数固定不能自适应背景的变化;当火车经过时"鬼影"现象长时间存在;检测的运动目标有空洞。提出了一种改进的Vi BE算法可以根据背景变化自适应参数;在火车经过时切换到特殊模式;对检测的运动目标进行形态学处理。试运行结果表明,改进的Vi BE算法能很好地适应于铁路沿线复杂的背景环境,且实时性和鲁棒性较高。

关 键 词:VIBE  智能监控  入侵检测  铁路安全
收稿时间:2016/3/13 0:00:00
修稿时间:2016/4/28 0:00:00

Research on Railway Video Monitoring and Moving Target Detection Based on Improved ViBe Algorithm
YANG Hai-lin,CAI Qi-zhong,XUE Sheng-li and JI Zhong-sheng.Research on Railway Video Monitoring and Moving Target Detection Based on Improved ViBe Algorithm[J].Science Technology and Engineering,2016,16(21).
Authors:YANG Hai-lin  CAI Qi-zhong  XUE Sheng-li and JI Zhong-sheng
Institution:School of Electrical and Information Engineering,Guangxi University of Technology,School of Electrical and Information Engineering,Guangxi University of Technology,School of Electrical and Information Engineering,Guangxi University of Technology
Abstract:VIBE algorithm in intrusion detection under complicated background environment along the railway has obvious problems.In the VIBE algorithm parameter is fixed, cannot the auto-adapted background change;When the train passed the "ghost" phenomenon exists for a long time;Detected targets have the cavity.An improved VIBE algorithm.The algorithm can change the auto-adapted parameter according to the background;Switch to special mode when the train passed;Morphological processing for detection of moving targets. Test results show that,improved VIBE algorithm is well adapted to the complex background environment along the railway, and in real time and high robustness.
Keywords:VIBE  Intelligent Video Monitoring  Moving Target Detection  Railway safet
本文献已被 CNKI 等数据库收录!
点击此处可从《科学技术与工程》浏览原始摘要信息
点击此处可从《科学技术与工程》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号