一种基于核回归的改进型图像去噪方法

为了消除图像处理中的噪声,同时尽可能地保留图像细节,提出了一种基于核回归的图像去噪算法。该方法的基本思想是在经典方法以像素位置决定权值的基础上,引入像素灰度值。即核函数在计算权值时考虑两个因素:空间距离和灰度距离。通过计算控制核来做到自适应,最后引入一个迭代过程。实验结果表明该算法能够在滤除图像中的高频噪声的同时尽可能保留了图像的细节特征,获得了较为理想的去噪效果。

A Improved Image De-Noising Method Based on Kernel Regression

In order to eliminate noise in image processing while preserving image detail as more as possible,a algorithm for image de-noising based on kernel regression is presented.The basic idea of this method is introduce pixel gray value instead of merely the pixel location decide the right.In the calculation of weight spatial distance and gray-scale range are taken.To make it adaptively by calculating steering kernel,an iterative process is introduced at last.The final study shows that the algorithm is able to filter out the the high-frequency noise of image while retaining the details of the image characteristics as much as possible.