中文摘要：

针对单纯基于块或像素对图像进行背景建模在目标检测中存在高误检率，提出了一种新的分层背景建模和目标检测方法。首先利用基于局部二值模式的纹理信息进行第一层分块背景建模，然后缩小建模粒度，在第一层上选取代表点进行第二层码本背景建模；目标检测时，不同粒度从上到下与所得背景模型分层匹配。对比实验表明了分层背景建模算法的实用性和鲁棒性，不但有效避免了像素背景建模中目标颜色和背景相似导致的误检，也弥补了分块背景建模在边缘处背景信息过多的问题。

英文摘要：

Background modeling methods only based on image blocks or pixels suffer from unacceptable false negative detecting rate. A novel layered background model and object detecting approach is presented. Firstly, every block on the first layer is modeled via textures based on local binary pattern operators. Then, the modeling granularity is deflated and some representative pixels on the second layer are chosen to model background with the codebook. In object detecting, different granularities are matched with the background model from top to down. Experimental results prove that the approach is effective and robust. It efficiently avoids the false negative detection rate in the pixel-based background modeling when the object color is similar to that of the background, and remedies the false positives occurring on the contour areas of the moving objects due to the area model.