基于人工神经网络方法的冬小麦叶面积指数反演  

Leaf area index retrieval of winter wheat using artificial neural network

作　　者：马茵驰[1,2] 阎广建[1] 丁文[2] 王跃智[2] 

机构地区：[1]北京师范大学遥感科学国家重点实验室,北京100875 [2]北京市农林科学院水产科学研究所,北京100068

出　　处：《农业工程学报》2009年第12期187-192,I0041,共7页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家自然科学基金项目(40871164);国家重点基础研究发展计划(2007CB714402);国家863计划成果"我国典型地物标准波谱数据库(2002AA130010)"

摘　　要：实践中,大尺度上测量叶面积指数(LAI)很难实现,利用遥感技术进行LAI的定量反演成为当前研究的重点。该文应用MODIS地表反射率数据反演冬小麦叶面积指数,假设MODIS像元由作物和土壤混合,建立了SAILH模型与裸土反射率组成的线性光谱混合模型,基于人工神经网络的方法进行LAI反演,获得了北京顺义冬小麦种植区在2001年4月1个时间序列的LAI。研究表明,此方法能够较好的获取大区域尺度上的LAI,对冬小麦长势监测具有重要意义。In practices,measuring leaf area index（LAI）in large area scale is very difficult.Therefore,retrieving LAI quantitatively based on remote sensing technology is concerned by many researchers.We proposed a BP-ANN based method to retrieve winter wheat LAI using surface reflectance data of MODIS.The MODIS pixel is assumed to be composed by crop canopy and bare soil.The SAILH（Light Scattering by Arbitrarily Inclined Leaves including the Hotspot-effect）model was used to simulate the directional reflectance of crop canopy,and the bare soil was assumed to be Lambertian.Series LAI maps of winter wheat in Shunyi District,Beijing were retrieved using this method during April in 2001.The research indicated that this method can be used well to retrieve LAI in large area scale,which is valuable to monitor crop growth.

关 键 词：神经网络 遥感 监测 叶面积指数(LAI) MODIS数据 反演 

分 类 号：S127[农业科学—农业基础科学]