基于脊波与小波变换的联合图像去噪方法

Image denoising based on ridgelet and wavelet transform

目的通过对脊波与小波变换这两种图像处理方法的分析,指出两种方法各自在图像处理的优缺点并提出了基于脊波与小波变换的联合图像去噪算法(RWT)。方法在去噪之前,首先通过脊波变换中较大的脊波系数,探测并保留二维图像中的主要信息特征,然后在图像的剩余信息中,通过小波变换有效地去除包含在其中的零维噪声信息,可以保留图像中的主要信息特征并完成去噪过程。结果理论分析和实验结果都表明,与传统的小波阈值方法或脊波去噪方法相比较有明显的去噪效果。结论该方法不但可以保持图像的边缘和良好的视觉特性,而且去噪后图像的峰值信噪比可再提高将近2 dB,算法处理的时间复杂度为O(Nlog(N))。

Aim Based on the theoretical analysis of wavelet and ridgelet transform,To discuse the advantage and disadvantage in image processing,and propose a new image denoising method based on ridgelet and wavelet transform.Methods Before denoising,the new algorithm first explores and maintains the main information character in two-dimensional image by using ridgelet,and then effectively denoises the residual information by wavelet transform.Results The theoretical analysis and experimental results show that it has good effect compared with the commonly-used wavelet threshold denoising methods or ridgelet denoising.Conclusion The new denoising method has a rather good numerable and vision result.it can keep image′s edges from damaging and increase PSNR about 2dB with algorithm′s time complexity no more than O（Nlog（N））.