基于模糊推理的方向中值椒盐滤波算法

Fuzzy reasoning-based direction median filter for removing salt and pepper noise

针对灰度图像中的椒盐噪声,提出了一种基于模糊逻辑推理的方向中值滤波算法。该算法先利用椒盐噪声的特点,将图像像素点分为信号像素点和噪声像素点,然后利用模糊推理在4个方向上推理出最接近理想值的非噪声点代替当前噪声点。同时算法中采用一种简便的方法检测出噪声点所处的滑动窗口中的边和线。仿真结果表明,该算法能在有效抑制噪声的同时较好地保存图像的细节信息。

A fuzzy reasoning-based direction median image. Firstly, each pixel is classified to be signal pixel filter is proposed to remove the salt and pepper noise from gray-scale and noise according to the characteristic of salt and paper noise. Then choose the ideal noise-free pixel to replace the current pixel accordin~ to fuzzy reasoning in four direction, This paper also propos- es a simple approach to determine the edge and line direction in the current sliding window. The results of experiments show that the new algorithm has a good noise removing and detail of the image preserving capability.