基于几何形状特征的运动目标判决方法

为了室内安全监控异常报警任务的需要,文章在对运动检测后的二值图像进行形态学后处理,消除细小噪声、平滑物体边界,对处理后仍保留的大面积伪运动目标区域,通过提取和分析面积、纵横比等几何形状特征参数,提出了一种运动目标判决方法.该方法有利于消除虚假报警,对基于安全报警系统的异常检测具有一定的意义.

Approach of Moving Object Decision Based on Geometric Shape Features

In order to meet the requirement of intrusion alarming of indoor security surveillance,morphology method is used to refine the binary images of moving detection,so the problem of small noises and holes in the binary images can be solved.To distinguish the false target regions,an approach of moving object decision is proposed via the extraction and analysis of geometrical shape features parameters such as area,width-height ratio,etc,which reduces false alarm and provides a good foundation for abnormity detection based on security alarm system.