一种基于图像处理的人群密度估计方法

摘要

Crowd density estimation in public scene surveillance is an important issue of public security. A method of crowd density estimation based on image processing is presented. The number of people and the level of crowd density are estimated qualitatively and quantitatively combining methods of pixel statistics and texture analysis. To resolve the people number estimation of full scene, this paper divides the large scene into some needed small cells based on the proportion to estimate the density. Experimental results show that this method is simple, effective and practical in application. It also could provide powerful help to the public warming system, such as in airport , subway and station.%公共场景监控下的人群密度估计是公共安全管理中的一个重要内容,因此,对基于图像处理的智能化人群密度估计方法进行了研究.结合使用像素统计和纹理分析的方法,从定性和定量两个方面确定了人群人数和人群密度等级.针对大场景的人群密度监控情况,提出了根据实际场景在图像中的比例将大场景分成所需的小区域,然后对每个子区域进行人群密度估计,从而完成全场景的人群密度估计.实验结果表明这种方法简单、有效、实用,便于在实际中的应用,为机场、地铁、车站等公共场所的预警系统提供有力的帮助.