暴击(t) =总和(c (k) ^ 2 k≤t) + 2 *σ^ 2 * t *(α+ log (n / t))

%例1:信号去噪。负载噪声颠簸信号。加载noisbump;x = noisbump;使用sym6在5级对信号%进行小波分解。wname =“sym6”;列弗= 5;[c、l] = wavedec (x,列弗,wname);使用wnoisest估计1级的噪声标准偏离%细节系数。σ= wnoisest (c、l、1); % Use wbmpen for selecting global threshold % for signal de-noising, using the tuning parameter. alpha = 2; thr = wbmpen(c,l,sigma,alpha) thr = 2.7681 % Use wdencmp for de-noising the signal using the above % threshold with soft thresholding and approximation kept. keepapp = 1; xd = wdencmp('gbl',c,l,wname,lev,thr,'s',keepapp); % Plot original and de-noised signals. figure(1) subplot(211), plot(x), title('Original signal') subplot(212), plot(xd), title('De-noised signal')

例2:图像去噪。%加载原始图像。加载noiswom;nbc =大小(图1);使用coif2在level 3对图像进行小波分解。wname =“coif2”;列弗= 3;[c, s] = wavedec2 (X,列弗,wname);估计噪声标准偏差从%细节系数在1级。det1 = detcoef2(“紧凑”,c, s, 1); sigma = median(abs(det1))/0.6745; % Use wbmpen for selecting global threshold % for image de-noising. alpha = 1.2; thr = wbmpen(c,l,sigma,alpha) thr = 36.0621 % Use wdencmp for de-noising the image using the above % thresholds with soft thresholding and approximation kept. keepapp = 1; xd = wdencmp('gbl',c,s,wname,lev,thr,'s',keepapp); % Plot original and de-noised images. figure(2) colormap(pink(nbc)); subplot(221), image(wcodemat(X,nbc)) title('Original image') subplot(222), image(wcodemat(xd,nbc)) title('De-noised image')