复杂地形下自动提取多暗像元的TM影像大气纠正方法  

The Atmospheric Correction for TM Image under Complicated Geographic Conditions Based on Automatically Extracted Many Dark Objects

作　　者：何颖清[1] 邓孺孺[1] 陈蕾[1,2] 齐志新[3] 秦雁[1] 朱家敏[1] 

机构地区：[1]中山大学地理科学与规划学院,广东广州510275 [2]国家海洋局南海海洋工程勘察与环境研究院,广东广州510300 [3]香港大学城市规划与设计系

出　　处：《遥感技术与应用》2010年第4期532-539,共8页Remote Sensing Technology and Application

基　　金：863计划项目(2006AA06Z416);国家自然科学基金项目(40671144);广东省水利科技重点攻关项目(ysk200602)

摘　　要：复杂地形条件下气溶胶的空间分布变化较大,用单一的气溶胶实测参数或单一的暗像元进行大气纠正都难以获得好的效果,手动选取暗像元还受限于专业人员的经验。通过程序自动提取图像中的浓密植被像元以及位于山区阴影的植被像元作为暗像元,使得暗像元均匀分布于图像的各个区域、各个海拔,更好地模拟复杂地形下的大气状况。根据辐射传输模型,利用迭代、插值等方法获取TM1和TM3波段光学厚度,进而推算成像时刻的气溶胶波长指数及混浊度系数,从而得到各个波段的大气光学厚度以实现大气纠正。Under complicated geographic conditions,the spatial distribution of aerosol is comparatively significiant.It is hard to achieve desirable result only by using measured parameter of Aerosol or Dark Objects to carry out the atmospheric correction.Moreover,manually selecting dark objects requires sufficient expertise.Program is used to automatically find out dense vegetation pixels and vegetation pixels under shadow in mountainous areas as dark objects,which are evenly distributed in different district and evaluation,as many as possible.Next,a good imitation was made for complicated geography of atmospheric condition.Based on the radiative transfer model,some mathematical methods such as iteration and interpolation are used to obtain the optical depth of TM1 and TM3 bands.Then,the aerosol wavelength exponent and turbidity coefficient are calculated.Finally,the atmospheric optical depth of each band is calculated to implement the atmospheric correction.

关 键 词：大气纠正 多暗像元 自动提取 复杂地形 TM影像 

分 类 号：TP752.2[自动化与计算机技术—检测技术与自动化装置]