摘要
为利用视频监控系统识别铁路客运站的客流,根据车站监控环境的多变性特点,提出将光流速度场算法与背景差分算法相结合的自适应背景更新方法。将光流引入背景建模中,并与背景差分结果进行并运算,再通过"死角"灰度优化处理,实现背景的实时更新。以实录的北京南站视频对给出的自适应背景更新方法进行验证,并与均值背景法和高斯背景法的处理结果进行比较,结果表明,自适应背景更新方法较好地解决了背景的提取、实时更新及运动目标阴影扰动等问题,拟合的背景干净、虚影弱,描述的背景符合实际背景场景,用于动态场景的客流识别取得了较好的效果。
According to the variability of the monitoring environment at station, the adaptive background updating method based on optical flow and background difference was proposed to recognize passenger flow using video monitoring system at railway passenger by introducing optical flow in background modeling, station. Real-time background updating was realized background difference in union operation and ＂blind angle＂ in gray processing. Beijing South Railway Station video was adopted to verify the given adaptive background updating method and the obtained results were compared with those by the average back- ground method and Gaussian background method. The results show that the adaptive background updating method has effectively solved such problems as the extraction and real-time updating of background as well as the shadow disturbance of moving object, etc. The fitting background is clean and the virtual shadow is rather weak. The described background conforms to the actual background scene. Better passenger flow recognition results have been achieved in dynamic traffic scenes.
出处
《中国铁道科学》
EI
CAS
CSCD
北大核心
2014年第6期131-137,共7页
China Railway Science
基金
国家"八六三"计划项目(2009AA11Z207)
高等学校博士科研基金(20110009110011)
关键词
背景更新
光流速度场
背景差分
视频监控
客流识别
Background updating
Optical flow
Background difference
Video surveillance
Passengerflow recognition
作者简介
王爱丽（1987-），女，甘肃白银人，博士研究生。
