一种改进的自适应背景建模视频分割方法

An Improved Video Segmentation Method Base on Adaptive Background Modeling

针对传统的自适应背景建模难以解决背景复杂以及外界光照变化等因素影响视频分割质量的问题,提出了一种改进的自适应背景建模视频分割方法。该方法首先对彩色图像建立高斯背景模型,然后对背景模型进行更新,最后通过对高斯分布准则进行改进使前景目标分割性能得到提高。仿真实验表明,该方法能够准确建立背景模型,准确分割前景目标,较传统算法具有更强的鲁棒性。

An improved video segmentation method on adaptive background modeling is proposed for solving the difficult problems of traditional Gaussian background modeling, and the problems contain complex background, outside illumination changes etc. First, the Gaussian background model is constructed. Then, the background model is updated. At last, the moving target detection can be effectively improved with the update of new distribution generation rule. Experimental results show that the proposed method could construct background model accurately, segment foreground accurately and have strong Robustness compared with traditional method.