基于小波包变换的非局部均值去噪方法

摘要

Based on the non-local means(NL-means) filter algorithm, in order to improve the image quality, a NL-means algo-rithm based on wavelet packet transform is proposed .Firstly, the image is transformed by the wavelet packet , and the wavelet do-main coefficients are applied to estimate the Gaussian noise parameters of the image , then the similarity of the high-frequency sub-band ’ s wavelet coefficients is calculated as the weights to adjust the wavelet coefficients , finally the image is reconstructed by the inverse wavelet packet transform .Experiment results show that this algorithm can preserve the edge detail information effectively , and get a superior denoising performance than the original non-local mean algorithm .%在非局部均值滤波的基础上，为了更有效地去除图像噪声，提出一种基于小波包变换的非局部均值去噪算法。首先对图像进行小波包变换，通过小波域系数估计图像的高斯噪声参数，然后计算经小波包分解后高频子带内小波系数的相似度，并以此作为权值来对小波系数进行调整，最后通过小波包逆变换对图像进行重建。实验结果表明与传统的非局部均值去噪算法相比较，该算法能在去噪的同时有效地保持图像的边缘细节等信息，取得更好的去噪效果。