Abstract: As the ridgelet transform can process line singularities of images more efficiently than the wavelet transform, an image-denoising method combining the wavelet transform with the ridgelet transform is proposed. The proposed method divides a noisy image into homogeneous blocks and non-homogeneous blocks, and the homogeneous and non-homogeneous blocks are processed by the wavelet transform and the ridgelet transform respectively. The processed image is filtered by the Wiener filter. Experimental results show that the proposed method has higher SNRs for the denoised images than both the ridgelet denoising method and the wavelet denoising method. The proposed method is found more efficient in preserving image details.

Key words: image denoising, wavelet transform, ridgelet transform, homogeneous block

摘要： 由于小波变换不能有效地处理图像中的奇异线，而脊波变换能很好地弥补这一不足，提出了一种基于图像分块的小波和脊波联合去噪方法。该方法把噪声图像分成一定尺寸的图像块并选择其中的同质块和非同质块；利用小波去噪方法处理同质块，而非同质块用脊波去噪方法处理得到去噪后的图像；用维纳滤波器进一步处理去噪后的图像。实验表明，该方法与单纯的小波去噪方法和脊波去噪方法相比，信噪比有了较高的改善，能有效地保留图像的边缘细节信息。

关键词: 图像去噪, 小波变换, 脊波变换, 同质块