【摘要】 针对传统小波阈值函数在阈值处不连续性、不渐近性、不光滑性等问题，导致去噪后的图像失真严重。因此，文中以小波变换为基础，提出了一种改进的阈值函数，与硬阈值、软阈值以及已存在阈值函数比较，得出该函数具有优越的数学特性，并用峰值信噪比和均方值误差客观评价法验证了该算法的优越性。通过Matlab仿真实验表明，新改进的小波阈值去噪方法不仅保留了原始图像的细节，而且有效抵制了噪声，最终使图像质量得到明显提升。更多还原

【Abstract】 The traditional wavelet threshold function is non-continuous,non-asymptotic and non-smooth on the points of threshold,which leads to serious image distortion after denoising. An improved threshold function is proposed based on wavelet transform. In comparison with the hard threshold function(HIF),the soft threshold function(STF)and the existing threshold function,it is concluded that this function has superior mathematical characteristics,and the superiority of this method is verified by objective evaluation of peak signal-to-noise ratio and mean square error. The Matlab simulation experiments show that the new proposed wavelet threshold denoising method not only can keep the original image details,but also effectively resist the noising. The final image quality has been significantly improved.更多还原