基于自适应双边滤波的太赫兹图像去噪算法  

A Terahertz Image Denoising Algorithm Based on Adaptive Bilateral Filtering

作　　者：郭俊文 宋贵才[1] 李兴广[2] GUO Junwen;SONG Guicai;LI Xingguang(School of Science,Changchun University of Science and Technology,Changchun 130022;School of Electronics and Information Engineering,Changchun University of Science and Technology,Changchun 130022)

机构地区：[1]长春理工大学理学院,长春130022 [2]长春理工大学电子信息工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2018年第3期111-114,119,共5页Journal of Changchun University of Science and Technology(Natural Science Edition)

摘　　要：针对太赫兹反射扫描成像中介质不均匀及激光器抖动造成图像边缘分辨率较差、噪声严重和图像质量差的问题进行研究,提出了自适应双边滤波算法对太赫兹反射扫描图像进行去噪复原的方法。依据图像整体的高斯噪声分布,利用主成分分析(PCA)的方法估计最优图像噪声方差,达到由噪声方差自适应设置双边滤波的最优亮度标准差。实现提高去噪能力的同时更好的保留图像边缘特征,不必像传统双边滤波算法根据经验设置参数。实验结果表明,算法在峰值信噪比方面提高1.5963d B和结构相似度提高0.06,均优于传统双边滤波。Due to the problem of poor image edge resolution,serious noise and low image quality caused by the uneven medium and laser jitter in terahertz reflection scanning imaging,the adaptive bilateral filtering algorithm is proposed to denoise and reconstruct the terahertz reflection image method. According to the Gaussian noise distribution of the whole image,the principal component analysis（PCA） method is used to estimate the optimal image noise variance,and the optimal luminance standard deviation of bilateral filtering is adaptively set by the noise variance. To improve the ability of denoising while preserving the edge features of the image better,it is not necessary to set the parameters according to experience as the traditional bilateral filtering algorithm. The experimental results show that the proposed algorithm is superior to traditional bilateral filtering; the proposed algorithm the peak signal-noise ratio is improved by 1.5963 d B and the structural similarity is improved by 0.06.

关 键 词：太赫兹 反射扫描成像 自适应双边滤波 图像去噪 

分 类 号：TN911.4[电子电信—通信与信息系统]