基于平滑滤波处理的图像插值通用方法

A General Method of Image Interpolation Based on Smooth Filtering Processing

针对传统图像内插值无法克服图像中噪声的问题,论文利用平滑滤波处理和金字塔分解相结合的形式,提出了一种通用方法来设计具有任何图像平滑滤波处理的图像插值方法,介绍了均值滤波、中值滤波或概率滤波这三种滤波法生成插值的原理,不仅适用于内插值的下采样图像,而且还保留了所形成滤波的特性。通过对灰度图像进行插值仿真分析,验证了所提方法与滤波处理相结合的有效性。

Aiming at the problem that traditional image interpolation can't overcome the noise in image, this paper presents a general method to design image interpolation method with any image smoothing filter processing by combining smoothing filter processing with pyramid decomposition, and introduces the mechanism of generating interpolation by means of mean filter, median filter or probability filter, which is not only applicable to image interpolation. The interpolated down-sampled image can retain the filtering characteristics. The validity of the proposed method is verified by the interpolation simulation of gray image.