复杂工业现场智能监控系统中的人体检测方法

Human detection technology in intelligent monitoring system for complex industrial spot

针对工业生产线上存在的安全隐患,提出了一种智能监控人体目标的方法。该方法首先通过高斯背景建模,背景差分,形态学处理,快速提取感兴趣区域的前景动态目标,然后通过目标所处区域进行目标尺寸复原,再结合目标形态比例特征和归一化后目标面积特征进一步识别人体,最后对非前景区域进行背景更行。实验证明,该方法在复杂工业现场能有效准确、快速的监测和识别出区域人体目标,且具有较强的光照变化适应能力。

Aiming at dealing with the hidden unsecure danger in the industrial production line, a method of intelligent monitoring human target was put forward. This method firstly used Gauss background modeling, background difference and morphologie processing to rapidly extract the foreground moving target of interest, and then by the target area the target size was recovered. Afterwards combined with the target proportion of form features and normalized target area feature, human was further recognized. Finally the prospect of regional background was performed. Experiments show that the method can effectively and accurately monitor regional human target and make rapid identification in complex industrial field, and has strong ability to adapt to the change of illumination.