摘要
关于图像传输的优化问题,针对传统小波阈值降噪方法,为了克服硬阈值函数不连续,软阈值函数中估计小波系数与分解小波系数之间存在恒定偏差,构造了一个新的非线性阈值函数,通过调整参数可视化地改变阈值函数的形状,将Dono-ho的硬阈值和软阈值作为两种特殊的情况。算法采用对非重要小波系数的处理,不是均设为0,系数可由多项式调节以接近理想小波系数,并进行仿真。仿真结果表明,算法对图像中的加性噪声能很好地去除,并且较好地克服了传统软、硬阈值存在的振荡和边界模糊偏差缺陷,改善了图像质量。
In order to overcome the discontinuity of the hard thresholding function and the constant deviation between the estimated wavelet coefficients and the decomposition wavelet coefficients in the soft thresholding function of the wavelet threshold de-noising method,this paper presents a new non-linear thresholding function,and the shape of the new thresholding function can be changed visually by adjusting parameters.It takes the Donoho hard-soft threshold as two kinds of special situations.The advantage of the algorithm is non-important wavelet coefficient processing.The coefficients are not supposed to be zero and can be adjusted by polynomial to approach its ideal wavelet coefficients.Simulation experiments confirm that the improved method can remove the noise effectively and overcome the shortcomings of hard and soft threshold with oscillation and borderline blurred.Both the PSNR and the visual effects are superior to the traditional method of the soft and hard threshold.
出处
《计算机仿真》
CSCD
北大核心
2011年第2期295-299,共5页
Computer Simulation
关键词
阈值函数
小波阈值降噪
峰值信噪比
均方误差
Thresholding function
Wavelet threshold
De-noising
Power signal to noise ratio（PSNR）
Mean square error（MSE）
