基于轮廓波变换的滤波器消噪性能

De-nosing capability of a filter based on contourlet transform

分析了轮廓波变换的基本原理和实现方法,根据自然图像的统计特征,讨论了基于轮廓波变换消噪的基本方法.在分析实际滤波器参数的基础上,结合实际滤波器的特性,通过实验探讨了在2个不同版本的轮廓波变换消噪方法中,滤波器的选择对于消噪效果的影响.结果表明,采用pkva12和9/7小波能充分利用自然图像中的正则性,具有比其他滤波器更好的消噪效果.

The basic principle and implementation of contourlet transform was discussed. The basic method of contourlet-based de-noising was also discussed according to the statistical feature of natural image. On the basis of the parameter and feature of practical filter, the effect of filter selection on contourlet-based de-noising was discussed for two different versions of contourlet transform by experimental means. The experimental results indicat that it could make better use of the integrity of natural image by means of wavelets pkva12 and 9/7,which can have the better de-nosing effect than other filters.