Abstract: This paper proposes a new B-spline & median filte based on human vision system.This algorithm fully considers the characteristics of humanity visual characteristic to detect the noise pixel，the filter window size is adaptively adjusted according to the window size，uses the b_spline function and median filter to estimate noise pixels.This paper produces the experiments and the result analysis.This algorithm can effectively remove noises and also retain image details very well.It also provides a new way for image filtering.

Key words: human vision system, 0epper &, Salt noise, B-spline, median filter

摘要： 论文提出了一种新的利用人眼视觉特性的B样条中值滤波，该技术充分考虑了利用人眼视觉特性确定噪声点，根据窗口内噪声点个数自适应选择合适的滤波窗口大小，采用B样条函数和中值滤波方法对噪声点进行逐点滤除。论文最后给出模拟实验和分析，结果表明该技术是有效的，既能有效地去除图像噪声点，又能较好地保持图像细节部分，为去除图像中的噪声提供了一种新的方法。

关键词: 人眼视觉特性, 椒盐噪声, B样条, 中值滤波