中文摘要：

依据脉冲噪声特点以及图像的像素关联性，提出了一种改进的脉冲噪声检测与处理算法。该算法首先依据脉冲噪声与其邻域多数像素在强度上具有明显差异的特点检测出疑似脉冲噪声点，然后再利用4个方向模板进一步区分疑似脉冲噪声点，最终建立脉冲噪声标记矩阵。之后，对检测到的脉冲噪声，提出了一种改进的加权中值滤波算法，该算法仅利用滤波窗口内的有效信号对窗口中心像素进行加权中值滤波，其中，加权系数的确定不仅依赖于像素间的灰度关联性，而且还依赖于像素间的位置关联性。实验结果表明，本文算法不仅可以准确检测和有效滤除图像中的脉冲噪声，而且还可以较好地保护图像中的细节。

英文摘要：

According to the characteristics of the impulse noise and the relationship between pixels ,an improved algorithm for detecting and removing impulse noise was proposed . The suspected impulse noise is firstly detected in algorithm according to the characteristics that the impulse noise is obviously different from their neighborhood pixels in intensity ,and then four direction templates for further distin-guishing the suspected impulse noise are used ,eventually a mark matrix about impulse noise are built . Later ,an improved weighted median filter algorithm was proposed in order to filter out the detected im-pulse noise ,w hich only uses the effective signal in filtering window to participate in filtering operation . In addition ,the determination of the weighted coefficients both depends on the gray correlation between pixels ,and depends on the position correlation between pixels .The experimental results show that the proposed algorithm not only can accurately detect and effectively filter out the impulse noise in image , but also can protect image detail well .