标题：Crowd Target Extraction and Density Analysis Based on FTLE and GLCM
作者：Qiao, Wei; Wang, Huiyuan; Wu, Xiaojuan; Liu, Pengwei; Wang, Huiyuan
作者机构：[Qiao, Wei; Wang, Huiyuan; Wu, Xiaojuan; Liu, Pengwei] Shandong Univ, Sch Informat Sci & Engn, Jinan 250100, Peoples R China.; [Wang, Huiyuan] Univ 更多
会议名称：2nd International Congress on Image and Signal Processing
会议日期：OCT 17-19, 2009
来源：PROCEEDINGS OF THE 2009 2ND INTERNATIONAL CONGRESS ON IMAGE AND SIGNAL PROCESSING, VOLS 1-9
关键词：crowd object detection; finite time Lyapunov exponent; gray level; co-occurrence matrix
摘要：Detecting and tracking up to 8-9 targets is already an arduous task for traditional methods. So it is even more challenging to deal with crowd objects. In this paper, segmentation and density analysis for video crowd flow is realized by a chaotic dynamics based approach. Crowd motion system is treated as a chaotic dynamic system, and crowd targets are treated as particles. As a result, Finite Time Lyapunov Exponent (FTLE) field is obtained. After morphological processing, motion region extraction and classification, texture information can be got from Gray Level Co-occurrence Matrix (GLCM). In GLCM analysis, density information can be achieved. Experimental results show the effectiveness of the proposed approach.