基于小波和脊波的图像联合去噪方法

为了在图像去噪时更好地保持细节特征,提出了联合小波和脊波的阈值去噪方法。在含噪图像小波分解后,对每一尺度下三个高频子带的细节分量进行单层逆变换,得到该尺度下的细节图像。对细节图像进行脊波阈值去噪处理,然后再进行单层小波分解。用所得的高频子带分别代替先前小波分解所得的高频子带。最后对处理后的图像小波系数进行小波逆变换,得到去噪图像。实验表明,在处理具有直线特征的图像时,该方法要优于单纯的小波或脊波阈值方法。

Image denoising method based on combined wavelet and ridgelet

To preserve more detail during image denoising,a combined wavelet and ridgelet approach is presented.Detail image at each scale can be obtained by inverse transformation of wavelet coefficients of a noise image.Each detail image is first denoised with ridgelet threshold method,and then decomposed with wavelet.The obtained wavelet coefficients are used as substitution for ones of the noise image.Finally,the denoised image is obtained by inverse transformation of the substitution wavelet coefficients.Experimental results showed better denoising performance for image with straight-line characters compared with only wavelets or ridgelts.