一种基于Shearlet变换域改进自适应中值滤波算法

New Improved Adaptive Median Filter Algorithm Based Shearlet Transform Domain

为了有效滤除图像中大量存在的脉冲噪声,提出了一种基于Shearlet变换域改进自适应中值滤波方法。首先在对Shearlet变换进行深入分析的基础上,给出了Shearlet分解和重构基本步骤;然后实现对含噪图像进行多尺度Shearlet变换,对获得多个尺度下的分解系数采用从噪声检测、噪声滤波等环节改进的自适应中值滤波算法(IAMF)进行噪声抑制;最后实现滤波后分解系数重构。分别与经典中值滤波(MF)、自适应中值滤波(AMF)以及Shearlet变换域阈值法进行比较,实验结果表明,该滤波算法滤波性能较好。

In order to filter the salt & pepper noise in digital image,a new adaptive median filter algorithm based Shearlet transform domain is put forward. Firstly,the theory on Shearlet transform is in-depth analyzed,and specific steps of Shearlet transform is put forward. Then,the noise image is conducted muti-scale Shearlet transform,the classical adaptive median filter is improved in noise testing strategy and noise filtering so as to deal with the decomposition coefficient. Finally,the the coefficient is reconstructed. The experiment results show that the performance of the algorithm in this paper is better than median filter( MF),adaptive median filter( AMF) and Shearlet transform domain threshold denoising method.