基于ViBe算法的改进背景减去法

摘要

ViBe背景减去算法基于RGB色彩空间对像素进行处理,在光照突然改变的情况下,会造成大面积的背景误判为运动前景；同时会将场景中的运动背景大量的误检为前景。针对上述问题,本文提出一种结合(r, g, I)标准色彩空间的改进算法。实验结果表明,改进算法在光照突然变化时对前景的提取具有更好的鲁棒性,同时对于场景中运动的背景像素点,取得了更好的检测效果。%ViBe background subtracted algorithm deals with pixels based on RGB color space. So when the light suddenly changed, a large of background area will be misclassified as moving foreground. In addition, ViBe algorithm considered a lot of the moving background pixels to be foreground in the scene. To solve these problem, this paper proposed algorithm which combined with (r,g,I) normalized color space. The experimental results show that the improved algorithm can reaches the more robust extraction for foreground at the abrupt changes of light. At the same time, the improved algorithm has a better effect on detecting the moving background pixels in the scene.