Wavelet image de-noising method based on noise standard deviation estimation
Zhen-Bing Zhao, Jinsha Yuan, Qiang Gao, Yinghui Kong · 2007
Wavelet threshold de-noising method is an important approach for image de-noising. In the paper, the standard deviation estimation of image noise was introduced, and its correction was given. Based on its correction, a wavelet shrinkage threshold method of image de-noising was proposed. The principle of wavelet image de-noising was discussed. The standard deviation of image noise was estimated by the difference between two Laplace templates. According to the different behaviors of the useful signal and the noise in wavelet domain, image de-noising method was designed. It can be seen from the de-noising results of simulation images and temperature field images that the proposed method can estimate the noise standard deviation well, improve the peak signal-to-noise ratio and the visual quality, and remove the noise from the image effectively. And it shows the method is better than other traditional ones, too.