红外图像WBC变换方向内和尺度间降噪  

Directional and inter-scale infrared image denoising algorithm based on WBC transform

作　　者：郑佳枫 张炜[1] 叶树亮[1] 

机构地区：[1]中国计量学院工业与商贸计量技术研究所,浙江杭州310018

出　　处：《激光与红外》2016年第3期351-356,共6页Laser & Infrared

基　　金：国家重大科学仪器设备开发专项(No.2013YQ470767)资助

摘　　要：针对传统图像降噪算法不能很好处理含噪红外图像的问题,本文分析红外图像在WBC变换后邻域系数相关性将广义CB形态学应用于邻域系数降噪,并结合基于尺度间相关性的自适应阈值降噪,提出一种基于WBC变换邻域系数相关性和尺度间相关性的红外图像降噪算法。仿真结果表明,该方法相比Bayes估计降噪、WBC尺度间硬阈值降噪和WBC尺度间自适应降噪获得了更好的SNR提升,其SSIM值也更接近于1。As traditional denoising algorithm can＇t deal with the infrared image with noise signal well,an infrared image denoising algorithm of neighborhood coefficient correlation and inter-scale dependency based on WBC transform is proposed by analyzing infrared image＇s correlation of the neighborhood coefficient after the WBC transform. Firstly,the generalized CB morphology is applied to neighborhood coefficient denoising. Then,infrared images are denoised combining with the self-adaptive threshold based on inter-scale dependency. The results show the proposed algorithm can get higher SNR compared with Bayes estimation denoising method,WBC inter-scale hard-threshold denoising method and WBC inter-scale adaptive threshold denoising method. Its SSIM is also closer to 1.

关 键 词：红外图像 邻域系数 尺度间相关性 自适应阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]