基于PCA和BM3D的噪声估计方法及其在中子图像去噪中的应用  

Noise Level Estimation Method Based on PCA and BM3D for Neutron Image Denoising

作　　者：乔双[1] 吴晓阳 赵辰一 王泽琦 QIAO Shuang;WU Xiaoyang;ZHAO Chenyi;WANG Zeqi(School of Physics,Northeast Normal University,Changchun 130024,China)

机构地区：[1]东北师范大学物理学院,吉林长春130024

出　　处：《原子能科学技术》2018年第4期729-736,共8页Atomic Energy Science and Technology

基　　金：国家重大科学仪器设备开发专项资金资助项目(2013YQ040861);国家自然科学基金资助项目(11275046;11405027;11201061)

摘　　要：针对中子图像的去噪问题,本文提出了一种弱化残差图像特征值的噪声估计方法。该方法将残差图像作为噪声图像,利用主成分分析(PCA)分析了残差图像特征值,估计了噪声强度。根据噪声强度,应用三维块匹配方法(BM3D)去除了噪声,逐步弱化残差图像特征值,得到了最终噪声估计值。实验结果表明,该方法能估计模拟噪声图像和实际含噪中子图像的高斯噪声,计算效率高,可实现较好的去噪复原效果。To solve the problem of neutron image denoising,a noise level estimation method of weakening residual image eigenvalue was proposed.In this method,the principal component analysis(PCA)was utilized to analyze residual image eigenvalue and the noise level was estimated.According to the noise level,the image was denoised,and the residual image eigenvalue was weakened to acquire the final estimated noise level.The experimental results show that the proposed method can estimate the Gaussian noise level of the simulated image and the real neutron image,and achieve preferable denoising effect with high computational efficiency.

关 键 词：残差图像 噪声估计 高斯噪声 中子图像 图像去噪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]