概率密度估计和阴影抑制的运动目标检测

Kenel Density Estimation for Motion Detection and Shadow Suppression

采用核密度估计算法得到可靠背景,通过试验准确地分割出前景物体.利用最近的历史帧数据估计当前像素的概率密度,以适应不同的复杂背景场景.结合阴影抑制技术,通过HSV色彩的阴影抑制处理降低目标检测的虚警率,不仅减轻污染前景的程度,还能得到更加合理的背景模型和前景目标,提高运动目标检测的准确性和鲁棒性.

This paper proposed a method of kernel density estimation （KDE） to get reliable background, and extract foreground object accurately by the experiment. In order to improve adaptive ability for different backgrounds, historical pixels are used to estimate the probability density of current pixels. Combined with shadow suppression, Hue-Saturation -Value （HSV） color information is used to detect and suppress moving cast shadows. This not only alleviate the pollution of foreground, but also get more reasonable background model and foreground object, and improve the accuracy and robustness of motion detection.