夜间近红外线视频监控图像人体目标检测

Human detection in nighttime near-infrared video

人体目标检测是很多机器视觉应用的难点,如安全视频监控和车辆辅助驾驶等,其检测方法多针对可见光图像,对夜间图像则少有涉及。夜间监控常使用近红外线摄像机,然而其所摄视频图像光线不均、噪声高、对比度差、色彩信息不足,此外夜间监控常伴有感应式照明设施,因此图像人体目标检测难度较大。提出双模式高斯混合模型,即分开建构开、关灯环境的背景高斯混合模式,并由检测感应式照明设施的开、关灯切换高斯模式。试验证明,双模式可较好地对近红外线视频图像中人体目标进行检测,但未对重叠人体目标检测进行分析,这有待于进一步研究。

Human detection in images is a challenging problem in the applications of machine vision,such as intelligent video surveillance and vehicle assistance driving.There are many methods for visible but not for invisible light.And near-infrared viedos are always used in nighttime surveillance system.Due to uneven brightness,high noise,low contrast and less color information,and sensing lighting equipments,it is difficulty to detect human.In this paper,a two-mode GMM is proposed which separately constructs background GMM for different lighting conditions and switches GMM modes by event detection.Experimental results show that by this two-mode GMM event detection is efficient,but more research should be done for overlapped targets.