小波分析在图像降噪中的应用

Application of wavelet analysis in image denoising

针对图像在采集与传输中受到的噪声污染,为提高图像信噪比,提升图像准确性与实用性,基于小波分析应用在图像降噪领域的原理与优势,在Donoho阈值降噪方法基础上,提出了一种改进的图像降噪方法。应用改进公式,可以根据图像具体情况选择参数,获得更有效的阈值函数。该方法的优势在于计算小波系数方面,尤其是计算大的系数误差比小的系数误差要小,从而提高了降噪水平。通过Matlab仿真和实际图像降噪结果分析,该方法明显优于传统阈值降噪方法,主要体现在阈值选取灵活、边缘信息处理平滑、降噪效果好等方面。

Images are corrupted by the noises during their acquisition or transmission,so denoising is essential in order to imporve the Signal to Noise Ratio and their accuracy and practicality.An improved image denoising method is proposed based on hard threshold and soft threshold method raised by Donoho,which is an application of the principles and advantages of wavelet analysis used in the filed of image denoising.In order to get a more effective threshold function,the parameters of the improved formula can be selected according to images.The advantage of this method is the calculation of wavelet coefficients,especially in the inaccuraciy errors of large coefficients,they arc smaller than small coefficients,so the level of denoising is improved. From matlab simulation re- suits and denoising of actual images,this method is better than the traditional methods mainly in the aspect of flexible selection of threshold,smooth treatment of marginal imformation and good effect of denoising.