E0-1 Hyperion高光谱影像的FLAASH大气校正与评价  

Atmospheric correction and evaluation for E0-1 Hyperion images based on FLAASH model

作　　者：龚建周[1,2] 陈健飞[1] 刘彦随[2] 

机构地区：[1]广州大学地理科学学院,广东广州510006 [2]中国科学院地理科学与资源研究所,北京100101

出　　处：《广州大学学报（自然科学版）》2011年第5期69-75,共7页Journal of Guangzhou University:Natural Science Edition

基　　金：国家自然科学基金项目(41171070);中国博士后科学基金重点项目及面上项目(200902132;20080440511);广东省自然科学基金项目(9151051501000030);教育部人文社会科学研究青年基金项目(10YJCZH031);广州市属高校科技计划项目(08C027;08C026)资助

摘　　要：遥感影像的大气校正是遥感数据应用的基础.文章利用ENVI软件的FLAASH大气校正模块,对Hype-rion高光谱遥感影像进行大气校正,选取典型地物对校正效果进行评价.主观定性评价结果显示校正后的影像地物边界更清楚、对比度增强;光谱曲线对比分析显示,校正前后光谱曲线形状完全不同,林地的反射率图像的光谱曲线形状与USGS标准库的曲线类似;典型地物影像质量的指标评价结果表明,校正后图像的平均梯度值基本都大于校正前影像的值;信息熵的值正好相反,大气校正的效果明显,进一步证实FLAASH模块各参数设置的合理性.Owing to system and occasional errors, atmospheric correction is the basis of application of remote sensing images. Based on FLAASH model in ENVI4.5, the paper implemented atmospheric correction to one scene image of Hyperion data and evaluated the quality of corrected image. Firstly, the boarders between both land cover types were made more clearly from each other in processed images than that in raw images. The local contrast was enhanced from raw images to corrected images. Secondly, double spectral curves of images demonstrated the obvious difference, while spectral curve for forest was similar to that from standard spectral library of USGS. Lastly, both indices of Average Gradient and Shannon were calculated for four typical objects, including forest, water, transportation and urban land cover. Average Gradient gave a good evidence for the effect of atmospheric correction by larger values for corrected images. Shannon had adverse values and gave an indirect check. All in all, image quality evaluation showed that the atmospheric correction method was a likely solution to eliminate the influence of atmosphere on remote data, while the model and parameters selected were corret.

关 键 词：Hyperion高光谱影像 大气校正 FLAASH 评价 

分 类 号：K909[历史地理—人文地理学]