视频监控中人群流量和密度估计算法分析

Survey of Crowd Flow and Density Estimation in Video Surveillance

对视频监控领域的群体分析技术进行了研究,包括基于像素统计和纹理分析方法的人群密度估计算法和基于分割和统计回归的在景人数统计的算法。这些算法在不同的背景下各具优点,但在实际应用中仍存在许多缺陷,需要进行改良和提升。最后,预测了对群体分析技术的发展趋势。

With the development of economy and the increasing of crowd activities, crowd analysis algorithm emerges prominently due to the broad applications in surveillance. The crowd density estimation algorithm based on pixel statistics and texture analysis is investigated; and the pedestrians counting algorithm based on segmentation and statistics regression. The paper indicates both the strong point and defects of present algorithms. Improvement and promotions of the algorithms are needed in future research work. The developing trend of crowed analysis field is given in the last section.