高斯核密度估计背景建模及噪声与阴影抑制

Gaussian Kernel Density Estimation-based Background Modeling with Noise and Shadow Suppression

提出了一种多模态非参数背景模型,用于背景减方法检测运动目标。针对户外监控系统存在背景局部运动以及摄像机抖动、活动阴影等问题,利用像素邻域相关性信息进行多模态高斯核密度估计,并采用HMMD色彩值抑制阴影。通过抖动噪声去除以及阴影抑制处理,降低了目标检测的虚警率。实验结果表明该算法在运动目标检测中具有对噪声和阴影的鲁棒性,可用于户外复杂场景监控系统。

A multimodal nonparametric background model is proposed to detect moving objects by background subtraction. In outdoor surveillance systems, the solution to some of the problems such as illumination changes, initialization of model with moving objects, and shadows are provided. The Gaussian kernel density estimation is exploited to estimate the probability density function of background intensity and to initially classify each pixel as belonging to background or candidate foreground. Pixel's neighbor information is considered to remove noise due to camera jitter and small motion in the scene. The Hue-Max-Min-Diff (HMMD) color information is used to detect and suppress moving cast shadows. That decreases the false positive in object detection. Experimental results demonstrate the robustness to noise and shadow and good detection performance, and it can be used in outdoor environment surveillance systems.