Abstract:

In video monitoring system, the population and density of passenger flow identified automatically are the key indicators of passenger security evaluation. This paper focuses on the method of monitoring area measurement to calculate the density. The method realizes the measurement of effective plane ground of monitoring area based on computer vision technology. It uses towdimensional perspective projection model and the template to calculate the homography, especially for the effectiveness analysis of the feature points picking up from the template with higher accuracy and operability. The paper provides a new way to realize the quick area measurement and calculation, especially on complex scenes. At last, method is applied to calculate the density of passenger traffic in Beijing South Railway Station and provides a way for passenger security warning.

Key words: information technology, area measurement, effective area plane, homography, passenger density

摘要：

视频监控系统中，自动识别出的客流人数和密度，是判断客流安全状态的重要指标.本文重点研究了为获取客流密度而进行的监控面积测量方法.该方法基于计算机视觉的平面测量技术，对视频监控区域有效平面部分进行面积测量.通过采用二维成像透视投影模型和模板的精确标定以计算单应矩阵，特别是针对特征点的有效性进行判断分析，准确率较高，可操作性强.此方法为快速实现复杂场景中面积测量和计算提供了一种新途径.最后基于北京南站具体场景，将此方法应用于客流密度计算，为客流安全预警提供方法支撑.

关键词: 信息技术, 面积测量, 有效平面, 单应矩阵, 客流密度