摘要
基于Donoho经典小波阈值去除图像噪声基本思路,分析常用硬阈值法和软阈值法在图像去噪中的缺陷。针对这些缺陷,提出一种改进的阈值去噪法,该方法不仅可克服硬阈值不连续的缺点,还能够有效解决小波分解预估计系数与真实小波系数间存有的恒定误差。通过Matlab仿真实验,使用改进的小波阈值法对图像去噪处理后,除噪效果比较理想,在去噪性能指标上,PSNR(峰值信噪比)和EPI(边缘保护指数)均好于传统阈值方法。
The basic idea of removing image noise by studying the Donoho classic wavelet threshold, in the analysis of the commonly used hard thresholding and soft thresholding method for image noise removal using defects, aiming at these defects, put forward a new denoising method improved the threshold, the method not only overcomes the characteristics of hard threshold is not continuous, but also effectively solve the wavelet decomposition coefficient estimate and real wavelet coefficients between the constant error, through Matlab simulation experiment, the wavelet threshold method using improved denoising image processing, denoising effect is ideal, the denoising performance index on PSNR （peak signal-to-noise ratio） and EPI （edge protection index） are better than the traditional threshold methods.
出处
《软件导刊》
2018年第1期89-91,共3页
Software Guide
