一种联合图像分割与背景建模的运动目标检测算法

介绍了在混合高斯模型的基础上,采用每一个像素点及其邻域组成的集合作为特征矢量来描述图像,对YUV格式的彩色图像的不同颜色分量分别建立混合高斯模型,从而确定是否有变化发生.为充分利用空间信息,提出将彩色图像分割与背景建模结合起来,得到具有精确边缘的运动目标.实验结果表明,即使在前景纹理、颜色比较一致且与背景对比不是很明显的情况下,本方法也能完整地检测出运动前景.

A Moving Object Detection Algorithm in Video Surveillance

Based on mixture Gaussian model,a pixel and its neighbors are used as an image vector to represent that pixel,and the different chrominance component pixels are used as a mixture Gaussian model,then to detect the moving objects.In order to make full use of the spatial information,color segmentation and background model are combined to extract moving objects with pixel-wise precision.Simulation results show that the algorithm can detect intact moving objects even when the foregrounds' texture and color are homogeneous and have low contrast with background.