NOAA卫星角度信息在分类中的有效性研究  

Utilizing the Multi-Angle NOAA Satellite Data to Enhance Terra Species Information Identification

作　　者：龙飞[1] 赵英时[1] 

机构地区：[1]中国科学院研究生院,北京100039

出　　处：《中国科学院研究生院学报》2002年第1期69-74,共6页Journal of the Graduate School of the Chinese Academy of Sciences

基　　金：国家 九五"攀登项目(95 -预 -38);国家重点基础研究资助项目(G2 0 0 0 0 7790 0)

摘　　要：利用连续几天的NOAA卫星数据,经过数据预处理、Ambrals、Rahman模型反演、大气校正,将提取出的角度方向信息,以土地利用图为基础,使多角度遥感理论研究与卫星遥感的实际应用有机结合,评价其在地物识别、提取、分类中对精度提高、方法改善的有效性.研究结果表明,加入角度数据后,无论从类间可分性还是从分类精度上,都要比常规分类数据高,其分类精度提高了 2 %～ 7%.This paper utilizes multi-angle NOAA satellite data and combines several main land cover type in agriculture-grazing neighboring zone of neimeng semi-arid region to study the method of obtaining and appraising the multi-angle directional information. After data pretreatment, atmosphere iterative inversion utilizing Ambrals model and Rahman model ,we register the obtained multi-angle directional information with landuse map, normal NOAA 1B data to prove and evaluate its validation in enhancing precision and improving method on land type identifying, distilling, classifying. All these aim to integrate multi-angle remote sensing, model inversion, experimental simulation with practical application of satellite remote sensing. At last, we combine and appraise the classification ability of different multi-angle data respectively, and compare the classification result of different classification mode. The classification result shows that the whole classification precision can improve 2%～7% compared with normal 1B data by adding angle information.

关 键 词：NOAA数据 多角度遥感 模型反演 BRDF 地物分类 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]