基于高光谱影像的玉米LAI反演模型研究  

Study on Inversion Model of Maize LAI Based on Hyperspectral Images

作　　者：翟羽娟[1] 张艳红[2] 刘兆礼[3] 刘宝江[3] 

机构地区：[1]吉林建筑大学城建学院,吉林长春130111 [2]吉林大学地球探测科学与技术学院,吉林长春130026 [3]中国科学院东北地理与农业生态研究所,吉林长春130012

出　　处：《江西农业学报》2015年第10期58-61,共4页Acta Agriculturae Jiangxi

基　　金：国家自然科学基金项目(40801067)资助

摘　　要：以吉林省长春西部地区合心与合隆镇为研究区,利用高光谱Hyperion遥感影像,结合与其同步实测LAI数据,分别建立影像波段与玉米LAI线性及非线性统计回归模型,最终选出以SAVI为自变量的指数函数模型y=1.717e1.064x为最优反演模型,其估测精度高达96.8%。经实验证实,高光谱遥感可以实现大范围、快速、较精确地获取玉米叶面积指数。In this paper, Hexin and Helong towns of western Changchun city in Jilin province were taken as the research re- gions. According to the hyperspectral remote-sensing Hyperion images and synchronously-measured maize LAI （ Leaf Area Index） data, the linear and nonlinear statistical regressive models describing the relationship between maize LAI and image band were built respectively. The exponential function model y= 1.717e^1.064x, which used SAV1 as independent variable, was finally selected as the optimum inversion model, and its estimative accuracy was as high as 96.8%. This experiment confirmed that the hyperspectral re- mote-sensing could quickly and accurately acquire the LAI of large-area maize.

关 键 词：HYPERION影像 玉米叶面积指数 反演建模 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]