介绍了背景的提取与自适应更新的算法,提出了一种基于直方图统计与多帧平均混合的算法。

这种算法克服单纯的多帧图像在车流量多、车速低时出现拖尾和模糊的现象,且相对于混合高斯模

型,其计算复杂度较低。该算法已通过TIDM642 DSP硬件平台实现,实验图像结果表明,该背景提

取算法的速度快,且符合实际场景

-The paper introducesbackground extractionswith self-adaptive update algorithm and puts forward an

algorithm ofbackground extraction based on statistics ofhistogram combiningwithmulti-fr a me average. Itavoids the

image trail-blurphenomena using puremulti-fr a memethod on traffic jam, and has relative low computation complex-

ity comparingwith themixGaussmode.l It can run on the TIDM642 DSP hardware platform.Experimental image

shows that the algorithm gives a fast extraction speed and gets good background image.