FLAASH模型输入参数对校正结果的影响  

Effect on Atmospheric Correction by Inputting Parameters of FLAASH Model

作　　者：杨校军[1,2] 陈雨时[1] 张晔[1] 

机构地区：[1]哈尔滨工业大学图像与信息技术研究所,哈尔滨150001 [2]北京遥感信息研究所,北京100192

出　　处：《遥感信息》2008年第6期32-37,101,共7页Remote Sensing Information

基　　金：国家自然科学基金(60472048)

摘　　要：大气订正模型反演精度的高低从一定程度上来讲依赖于模型输入参数的正确性,然而对于大气辐射校正模型来说,其输入参数是成像时刻的大气状况,显而易见这是无法精确获知的。正是这种模型对输入参数的强依赖特性限制了订正模型在工程上的普遍应用。本文利用部分AVIRIS航空遥感图像为例,对FLAASH模型在使用不同气溶胶模式、不同大气模式、不同能见度初始值设定和是否考虑邻近像元影响等输入条件情况下分别进行了大气订正实验,并与参考值进行了对比分析。实验结果表明,由于采用了尽可能利用图像本身信息的订正思路,使得FLAASH模型对部分输入参数的敏感性并不体现出强依赖,且其邻近效应处理环节大大提高了订正效果,适于工程应用。The retrieval precision of atmospheric correction in some degrees depends on the validity of model inputs. But obviously, the input parameters can not be satisfied due to the model input is very complex and indeterminable i. e. the imaging conditions. The dependence of a model on its input limits the usage in common project field. This paper tested the sensibility of ELAASH inputs using AVIRIS images on different aerosol mode, different atmosphere mode, different initial visibility and if the adjacent effects process presented. According to the data analysis, FLAASH showed weak dependence of input parameters because it considered the information of the image itself in a higher priority. In addition, the process of adjacent effect improved the performance remarkably. This model is adequate for engineering.

关 键 词：FLAASH 敏感度 高光谱遥感 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]