Abstract: To eliminate the influence of visual difference on color image filtering, a method to compute the visual sensitivity value of three primary colors was given. The filter algorithm detected pixels by rough set first, then used visual sensitivity to correct the test results, finally filtered the noise pixels by the improved Directional Distance Filter (DDF). The experimental results show that the new filter, compared with several other representative vector filters, provides better performance on color keeping, edge detail preservation, and noise filtering rate. In addition, the normalized mean square error of our method is the smallest and the Peak Signal-to-Noise Ratio (PSNR) is the largest under different noise density.

Key words: visual sensitivity, RGB space, adaptive, directional distance, vector filtering, iteration

摘要： 为消除人眼视觉差别对彩色图像滤波的影响，提出了RGB非均匀色彩空间中三基色(红、绿、蓝)的视觉灵敏度的计算方法。滤波算法先用粗糙集理论对噪声像素做初步判断，再结合视觉灵敏度修正判断结果，最后用改进的方向距离滤波器(DDF)滤除噪声像素。仿真实验表明，算法在颜色保持、边缘细节保护、噪声滤除率方面均优于现有滤波算法；且在各种密度噪声情况下，其归一化均方差最小，峰值信噪比最大。

关键词: 视觉灵敏度, RGB空间, 自适应, 方向距离, 矢量滤波, 迭代