视频图像中人体运动目标检测算法研究  

Research on algorithm for human motion target detection in video images

作　　者：李伟[1] 熊建设[1] 

机构地区：[1]中国海洋大学信息科学与工程学院,山东青岛266100

出　　处：《现代电子技术》2013年第3期74-77,共4页Modern Electronics Technique

摘　　要：视频图像中人体运动目标的检测是人体运动视觉分析的主要研究方向之一,在许多方面有着广泛的应用。主要针对摄像头固定即静态背景下的情况,提出一种基于高斯建模的背景减除法来实现运动目标的检测,并对运动目标实现人体分类判别,不仅可以较为精确地从背景中提取出前景区域,获得人体目标信息,又对环境有一定的自适应性,算法具有比较良好的鲁棒性。Moving human body detection in video images is one of the most important research directions in the visual analysis of human motion and has been broadly applied in many areas. For monitoring camera fixed in the environment （i.e. the static background）, a background subtraction method based on Gauss model is given to detect the moving objects and recognize the moving human body classification. The proposed method extracts foreground areas from the background accurately and obtains the human body information, which has the self-adaptability to the environment. The algorithm has Good robustness

关 键 词：视频图像 运动目标检测 人体目标分类识别 鲁棒性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]