一种新型的小波阈值图像去噪方法

对含噪图像小波阈值去噪方法中经典的阈值函数选取方式以及阈值选取方式的相关性能进行比较分析,在此基础上,构造出一种新的小波阈值去噪函数。新定义的阈值函数通过增加一个参变量来克服硬阈值函数不连续的缺点和软阈值函数高阶导数不连续的缺陷。理论分析和实验结果均表明,构造的小波阈值去噪函数具有更好的适应性和去噪效果。在图像含有强高斯噪声的情况下,新型阈值去噪方法相对软阈值法其峰值信噪比可以提高4dB到6dB。

Wavelet image de-noising algorithm based on a new threshold function

Based on the comparative analysis in the wavelet image de-noising algorithm on threshold function,and analysis in the classic threshold function selection method as well as the performance of the threshold selection,a novel threshold function is put forward.Thus,it constructs a new wavelet threshold function;the function contains a variable,which can overcome both the shortcoming of the discontinuous in hard-threshold function and the discontinuous of higher derivative on soft-threshold function.The theoretical analysis and experimental results indicate that the new threshold function has better adaptability and de-noising effect.Especially in the strong Gaussian noise,the proposed method can relatively improve PSNR 4dB ～6dB than soft threshold method.