基于光谱空间信息的高光谱遥感图像混合噪声评估.PDFVIP

第34卷第2期 红 外 与 毫 米 波 学 报 Vol. 34,No.2 2015年4月 J. Infrared Millim. Waves April,2015 文章编号:1001-9014(2015)02-0236-07 DOI:10.11972/j.issn.1001-9014.2015.02.019 基于光谱 空间信息的高光谱遥感图像 混合噪声评估 * 傅摇 鹏,摇 孙权森 ,摇 纪则轩 (南京理工大学 计算机科学与工程学院,江苏 南京摇 210094) 摘要:综合利用高光谱图像的光谱信息和空间信息,提出了一种新的混合噪声评估方法.首先通过滤波算法进行图 像中均匀图像块的自动选取;然后利用多元线性回归模型,将均匀图像块内像素点的信号值和噪声值进行分离,并 实现了图像中加性、乘性噪声的粗评估;最后根据噪声模型构建似然函数,利用最大似然估计法求解噪声模型参 数.通过仿真图像和真实高光谱图像进行实验,验证了该方法的准确性和鲁棒性. 关摇 键摇 词:混合噪声;噪声评估;乘性噪声;加性噪声;高光谱遥感图像;噪声模型 中图分类号:TP751摇 摇 文献标识码:A A spectral鄄spatial information based approach for the mixed noise estimation from hyperspectral remote sensing images * FU Peng,摇 SUN Quan鄄Sen ,摇 JI Ze鄄Xuan (School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing摇 210094,China) Abstract:A novel mixed noise estimation method based on the spectral and spatial information of hyperspectral images wasreported. Firstly,homogeneousimage blocks were automatically detected using data masking. Then signal value and noise value of each pixel in homogeneousblockswere split with a multipleliner regression model. Meanwhile,rough ap鄄 proximationsof SDand SI noise were obtained. Finally,likelihood function was built based on the mixed noise model, where parametersofthenoisemodelwerecalculatedbymaximum鄄likelihoodestimationapproach. Theproposedmethodis demonstratedtobe accurate androbust by experimentswith both synthetic imagesandreal hyperspectral images. Key words:mixed noise, noise estim