基于Contourlet变换和形态学的图像去噪方法

Image de-noising approach based on contourlet transform and morphology

提出了一种Contourlet变换与数学形态算子相结合的红外图像去噪方法。充分利用Contourlet变换后系数的分布特性以及尺度内和尺度间的依赖性,结合数学形态算子的特点,利用数学形态算子对变换系数进行处理,使得重要变换系数与非重要变换系数分离,对非重要系数子集进行软阈值处理,然后再将两个子集合起来,进行逆变换重建。实验结果表明,与传统小波相比,该算法具有更好地去噪效果,同时更有效地保留了图像的细节信息。

An image de-noising algorithm based on Contourlet transform and mathematical morphology was proposed. The statistics of the Contourlet coefficients of an image were studied, the coefficients of Contourlet transform of an image were manipulated by morphological operator, and morphological dilation was applied to extract the clustered significant coefficients in each subband,which result in the partition of each subband into significance clusters and insignificance space. Then, the smoothing image of the input image was using gotten the soft-thresholding method in insignificance space. Finally,two subsets were combined and the image was reconstructed. The experimental results demonstrate that compared with the traditional wavelet transform, this algorithm can denoise effectively,and keep the detail information. The method can improve the signal-to-noise ratio.