基于M带小波变换与模糊聚类的图像去噪算法

M带小波变换是标准二带小波变换的自然推广,能够分析具有相对窄带的高频信号,而且能更好的集中信号能量,因此在信号处理中应用广泛.本文结合模糊聚类算法,提出了一种新的基于M带小波变换的图像去噪算法,利用模糊聚类算法把小波系数划分成两类:包含信号的小波系数与只包含噪声的小波系数,对只包含噪声的小波系数置为零,将包含信号的小波系数进行利用软阈值法进行收缩,最后对处理后的系数进行M带小波逆变换,得到去噪后的图像.对SAR图像的实验结果表明,该算法有效,而且能较好地保留边缘信息.

An Image Denoising Algorithm Based on M-band Wavelet Transform and Fuzzy Clustering

M-band wavelet transform is a direct generalization of standard 2-band wavelet transform and has been widely applied in signal processing due to its capabilities of analyzing high frequency signals with relatively narrow bandwidth and giving better energy compaction. A novel image denoising algorithm based on M-band wavelet transforms and fuzzy cluster algorithm is proposed. Using fuzzy clustering algorithm wavelet coefficients are separated into two classes, one is coefficients only containing noise information and the other is those containing signal information. All those coefficients only containing noise information are set to zero and all those containing signal information are soft-thresholded. Finally inverse M-band wavelet transforms are performed and the denoised image is obtained. The proposed algorithm is examined on the test image and SAR image. The experimental results show the proposed algorithm is effective and can preserve edge information better.