采用零树结构分类小波系数的红外图像降噪

Infrared Image Denoising Based on Classified Wavelet Coefficients Using Zerotree Structure

红外图像易受噪声污染,为了改善红外图像的质量,提出了一种基于零树结构分类小波系数的红外图像降噪算法。该算法利用小波零树结构表达尺度间的相关性,通过空间自适应阈值将小波系数进行分类,并根据不同类系数的统计特性采用不同的先验分布模型,在贝叶斯框架下实现降噪。实验结果表明,本文算法在峰值信噪比(PSNR)指标上优于传统算法;从视觉效果来看,该算法在有效去除图像噪声的同时能较好地保持空间细节,可以满足当前红外图像降噪的需求。

Infrared image is vulnerable to noise pollution. In order to improve the quality of the infrared image, a denoising algorithm based on classified wavelet coefficients using zerotree structure was proposed. First, the wavelet coefficients were classified via adaptive threshold by expressing the inter-scale dependencies using zerotree structure. Then, various prior distribution models were adopted to represent various statistic characteristics of different class＇s coefficients. Finally, infrared image denoising was implemented by Bayes estimation. Experimental results show that the performance of the proposed algorithm is superior to the traditional algorithms in terms of the Peak Signal to Noise Ratio （PSNR）. As for visual quality, the proposed algorithm could reduce the noise effectively and retain more details simultaneously. Therefore, it can meet the general demand of denoising for infrared image.