基于改进脉冲噪声检测的新型滤波算法*

通过分析脉冲噪声图像的数值特征，为了快速和准确地滤除图像脉冲噪声并能很好地保持图像的细节，提出了基于改进脉冲噪声检测的灰度图像和彩色图像非线性自适应滤波算法。该算法首先通过改进的噪声检测方法把图像中的噪声点标志在噪声标志矩阵中，然后采用改进中值滤波方法并有限制地自适应调整滤波窗口对灰度图像中的脉冲噪声给予有效滤除。在此基础之上，分别采用该方法对彩色图像的三个RGB子图像进行单独滤波，然后利用通道融合技术得到最终的彩色滤波图像。经过实验仿真并与国内外相关文献提出的算法相比，本方法不仅思想简单、快速、易于实现

New filtering algorithm based on improved impulse noise detection

By analyzing the numerical characteristics of impulse noise images, in order to filter impulse noise quickly and precisely as well as preserving image details, this paper proposed non-linear adaptive filtering algorithm about gray image and co-lor image based on improved impulse noise detection. Firstly, the algorithm marked noise points in the noise marking matrix by improved noise detection method. Then, filtered the impulse noise in gray image efficiently by using improved median filtering with adaptive adjustment filtering window. On this basis, in each color channel of an image based on the RGB color space, filtered the noise separately, it could get the final color filtered image by the channel fusion technology. The results compared with other published algorithms show that the proposed algorithms are not only easy and quick, but also have better perfor-mances than the other existing approaches of gray and color image filtering.