轮廓波消噪中消噪效果与噪声强度的关系

Relation between de-noising effect and noise density of contourlet-based de-noising

首先简单回顾了轮廓波变换的基本原理和实现方法;然后阐述了基于轮廓波变换的图像消噪实现方法,针对包括小波消噪、第一版轮廓波变换消噪和第二版轮廓波消噪变换消噪在内的三种变换域消噪方法在不同噪声强度下的消噪效果进行了实验和分析。实验表明,第一版轮廓波变换在噪声强度较大时,比小波有较高的信噪比;在噪声强度较小时,比小波有更低的信噪比;第二版轮廓波变换则总是在信噪比上优于小波。

The basic principle and implementation approach of contourlet transform were first briefly reviewed, after that the contourlet transform based de-noising method was described. Under the conditions of different noise density, three kinds of denoising effect including wavelet, first and second edition contourlet transform, were provided and analyzed. Experimental results show that under higher noise density first edition contourlet has better signal to noise ratio than wavelet and under lower noise level the result is on the contrary, the 2nd edition of contourlet transform always has better performance in de-nolslng applications than wavelet.