视频监控系统中的图像分割算法综述

Survey of object segmentation algorithms in vision-based monitoring system

视频监控系统在智能安防、人机交互、交通、娱乐、军事等领域有着广泛的应用空间,是近来研究热点之一。目标的分割是视频监控系统中的首要任务,其分割的有效性对于后续的目标识别、跟踪、行为理解等处理至关重要。从时间分割法和空间分割法的分类角度出发,对国内外的视频监控研究中的若干目标分割方法进行了归纳总结。

Vision-based monitoring system （VMS） is currently one of the most active research topics, there are a lot of applications using VMS, such as smart surveillance, perceptual user interface, traffic surveillance, entertainment, military and so on. Object segmentation is the first step of VMS, and it is very important to object recognition, object tracking and behavior analysis. This article devided the segmentation algorithms into two kinds, the time-based segmentation algorithms and spatialbased segmentation algorithms, made a survey of the object segmentation algorithms in VMS.