摘要
为了快速准确地检测人群活动的变化,给出一种基于RGB动态纹理的人群视频图像序列检测算法。该方法首先提取视频序列图像RGB三个通道的动态纹理特征并级联起来,然后用滑窗法量化视频序列之间的差异,从而构造人群活动变化曲线。实验结果表明,所给方法能够有效地检测出人群活动的变化,并快速浏览人群行为的变化趋势。
A new method is proposed to detect the change of crowd sequence of video images based on RGB dynamic textures.Firstly,dynamic textures are respectively extracted from the RGB three channels of video sequence images and cascade them together.Then,quantify the difference between sequences by sliding window,thus,the change curve of crowd activities can be obtained.Experimental results show that,the change curve can fast browse and effectively detect the changes.
出处
《西安邮电大学学报》
2016年第6期29-34,共6页
Journal of Xi’an University of Posts and Telecommunications
基金
国家自然科学基金资助项目(41504115)
公安部科技强警基础专项资助项目(2015GABJC51
2015GABJC50)
陕西省自然科学基础研究计划资助项目(2015JQ6223)
陕西省教育厅科研基金资助项目(16JK1691)
关键词
RGB
动态纹理
人群活动
变化检测
RGB
dynamic texture
crowd activities
change detection
