摘要
根据逗留物体运动的特点,对传统的混合高斯背景图像建模方法进行改进,通过自适应混合高斯背景图像模型将作为前景图像的逗留物体从背景图像中提取出来;再用形态学滤波法消除逗留物体图像因阈值分割产生的意外噪声,然后采用颜色直方图匹配的算法实现对逗留物体的跟踪;当逗留物体在监测区内的累积停留时间超过预设的时间后启动报警。实验结果表明:该方法可以很好地适应监测范围内目标物体缓慢移动或静止的场景,能够快速、准确地对逗留物体进行检测、跟踪和报警。
Traditional MOG background modeling method was improved based on the motion characteristics of loitering （abandoned） objects. By means of adaptive MOG background model, loitering （aban- doned） objects as the foreground images were extracted from background images. Unexpected noises resulted from threshold segmentation of loitering （abandoned） object images were eliminated by morphological filtering method. Then the tracking of loitering （abandoned） objects was realized by the algorithm matching with color histograms. Alarm was started when the accumulative loitering time in surveillance area exceeded the preset time. Experimental results show that the proposed method can well adapt to the speed of targets （slow motion or motionless） in surveillance area, and can realize the detection, tracking and alarm of loitering （abandoned） objects quickly and accurately.
出处
《中国铁道科学》
EI
CAS
CSCD
北大核心
2013年第4期105-109,共5页
China Railway Science
基金
铁道部科技研究开发计划项目(2009X007-A)
关键词
视频监控
高斯混合模型
逗留物体
颜色直方图
Video surveillance Gaussian mixture model Loitering （abandoned） obiect Color histogram
