一种新的基于统计的背景减除方法

提出了一种有效的彩色视频背景减除的新方法。首先对彩色视频采样得到图像序列，统计序列中各像素的RGB值并归类，用出现概率最高类的RGB均值来构建背景模型;然后根据颜色差异和亮度范围综合条件，结合形态学处理进行背景减除。实验结果表明，此法可以很好地克服灰度视频背景减除中很难识别与背景灰度相近目标的缺陷，同时比传统的彩色视频背景建模快速，且样本中允许运动目标存在。对背景减除的准确性和实时性有一定程度地改进。该文还针对光照和背景变化提出了一些有效的背景更新策略。

New background subtraction algorithm based on statistic

A new effective background subtraction algorithm for color video is proposed.First gets an image sample sequence from a color video.Then analyzes and classifies the RGB value per pixel of the sequence and uses the average RGB value of the class which has the highest frequency in the sequence to construct the background model.At last subtracts background according to both color distortion and brightness range.At the same time，does some morphological work to improve the result.Experiment results indicate that this algorithm can solve the problem that gray video background subtraction algorithms can’t detect the targets which have the similar gray intensity compared to background.Further more，it can construct background model faster than other color video background subtraction algorithms and allows presence of moving objects in the sample images.It improves the accuracy and real-time performance of background subtraction to a certain extent.Finally proposes some effective methods to update background model so as to adapt the illumination and background changes.