改进的BiShrink与DTCWT相结合的遥感图像去噪

Remote sensing image denoising based on the combination of the improved BiShrink and DTCWT

基于双树复数小波变换(DTCWT)良好的平移不变性和多方向选择性,以及尺度内DTC-WT系数的领域相关性,提出了一种bivariate shrinkage(BiShrink)的改进算法,对遥感图像进行去噪处理。实验结果表明,经本文算法降噪后,图像的峰值信噪比(PSNR)得到显著提高,较好地保持图像的边缘和细节信息,并抑制混淆现象。

By considering the advantages of the dual tree complex wavelet transfer（DTCWT）in shift invariance and multi-direction selection,as well as the neighboring coefficients of DCTWT,an improved bivariate shrinkage（BiShrink） algorithm is presented to denoise the remote sensing images.Experimental results show that the proposed algorithm gets better peak signal-to-noise ratio（PSNR） than other methods obviously.In terms of visual quality,the proposed algorithm can achieve images with more details and edges,smooth profiles and restricted aliasing.