基于非抽取小波变换的遥感图像贝叶斯去噪

图像去噪是遥感图像处理的一个重要方面。文中基于非抽取小波变换，提出了一种贝叶斯图像去噪方法。对小波系数采用广义高斯分布建模，根据贝叶斯估计理论，得到贝叶斯收缩阈值，采用软阈值收缩去噪。实验结果表明：该去噪方法能够有效地抑制正交小波变换产生的人为干扰和伪Gibbs现象，与正交小波变换阈值去噪方法相比具有明显的优越性。

Bayesian Denoising for Remote Sensing Image Based on Undecimated Di screte Wavelet Transform

Image denoising is an important aspect for remote sensing image processing. A new Bayesian denoising algorithm based on undecimated discrete wavelet transform （UDWT） is presented in this paper. The BayesShrink threshold is derived in a Bayesian framework, and the prior model used on the wavelet coefficients is the generalized Gaussian distribution （GGD）.Image denosing is finished by using Donoho＇ s soft threshoiding. Experiment results show that the new algorithm can reduce the artifacts and the pseudo-Gibbs phenomena from the orthogonal wavelet transform, and has obvious superiority as compared with orthogonal wavelet denoising method.