图像椒盐噪声滤波的自适应开关插值算法

Adaptive Switching Interpolation Algorithm for Filtering Salt and Pepper Noises in Image

针对图像中椒盐噪声的特点，利用牛顿插值，提出了一种白适应开关插值滤波方法．结合极大极小值椒盐噪声检测算法和自适应窗口大小方法标识可能的噪声点，然后用迭代图像插值估计方法，利用与当前噪声点近邻的非噪声点或已修复的噪声点消除当前噪声点．对噪声密度为10％-90％的测试图像的仿真结果表明，该算法能在有效抑制噪声的同时较好地保持图像的细节信息．

An adaptive switching interpolation filter was proposed using Newton interpolation method to remove salt and pepper noises in an image. Maximum-minimum noise detector and adaptive window size method were combined to identify pixels which are likely to be corrupted by noises. Then the image is restored by iterative interpolation to replace each noised pixel with a non-noised or repaired pixel in its vicinity. Simulation results on images with noise densities from 10% to 90% show that the proposed filter performs well in reducing salt and pepper noise in images with good detail-preservation.