结合FT与NSCT的自适应阈值去噪算法

Adaptive Threshold De-noising Algorithm Combined with Fourier Transform and Non-subsampled Contourlet Transform

为更好地对图像进行稀疏表示,以改善去噪效果,提出一种傅里叶变换与非下采样轮廓波变换(NSCT)相结合的自适应阈值去噪算法。在傅里叶域中对含噪图像去噪,在NSCT域中利用分层噪声估计的贝叶斯阈值算法,结合多尺度多方向的能量阈值修正方案自适应地滤除剩余噪声。实验结果表明,该算法的去噪性能较好。

In order to express a sparse image better and achieve better de-noising effect, an adaptive threshold de-noising algorithm combined with Fourier Transform（FT） and Non-subsampled Contourlet Transform（NSCT） is proposed. The noisy image is de-noised in Fourier domain, and the remaining noise is filtered out in NSCT domain. It is based on stratified noise estimation and adaptive Bayes threshold, combined with a flexible multi-scale and multi-directional energy correction threshold scheme, Experimental results show that the proposed algorithm can improve de-noising performance.