基于优化PROSAIL叶倾角分布函数的玉米LAI反演方法  

Leaf Area Index Retrivel for Maize Canopy Using Optimized Leaf Angle Distribution Function of PROSAIL Model

作　　者：苏伟[1] 郭皓[1] 赵冬玲[1] 刘婷[1] 张明政[1,2] 

机构地区：[1]中国农业大学信息与电气工程学院,北京100083 [2]山东农业大学资源与环境学院,泰安271018

出　　处：《农业机械学报》2016年第3期234-241,271,共9页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金项目(41371327);北京高等学校青年英才计划项目(YETP0316)

摘　　要：叶面积指数(LAI)是描述玉米冠层结构的重要参数之一,PROSAIL模型是常用于反演LAI的机理模型,能较为真实地反演植被冠层真实情况,但PROSAIL模型中使用的叶倾角分布函数假定区域内所有作物叶倾角分布是相同的,不能反映玉米植株真实的叶倾角分布情况。本研究以高分一号遥感影像和地基激光雷达点云数据作为数据源,充分利用地基激光雷达(TLS)在获取植被结构参数上的优势,通过体素化的方法对玉米叶片回波点云进行分割,获取每个拟合叶片单元的叶倾角,进而得到玉米植株真实的叶倾角分布,结合椭球分布函数得到玉米精确的叶倾角分布函数,实现对PROSAIL模型中叶倾角分布函数的优化。研究过程中分别基于未改进的PROSAIL模型和经过TLS优化后的PROSAIL模型反演黑龙江825农场主要玉米种植区的LAI。LAI反演结果表明:2种反演方法得到的LAI与实测LAI都具有较好的相关性,决定系数R2分别为0.557 6和0.858 3,模型可信度较高;但基于PROSAIL模型反演所得LAI结果偏低,在利用TLS数据提取叶倾角对模型进行优化后,反演LAI的估算精度由26.53%提高到96.23%。由此可知,通过引入TLS点云数据改进农作物叶倾角分布函数能大幅度提高LAI反演的准确性。Leaf area index（ LAI） is one of the important parameters to describe the corn canopy structure. PROSAIL model is a mechanism model for retriving LAI,which can express canopy situation more truly. But the leaf angle distribution function used in PROSAIL model assumed that the leaf angle distribution is constant during whole crop growth period,and it can＇t reflect the actual leaf angle distribution of the corn plant. This paper studied the extracting method of the maize LAI based on the PROSAIL model,using GF-1 images and terrestrial Li DAR data. In order to get the leaf angle distribution of maize,the point cloud of maize was separated into small leaf units through voxel method,and then the surface was matched according to the point cloud in each voxel. The accurate leaf angle distribution function was acquired from statistics data of each leaf unit angle. Combining with the ellipsoid distribution function,the accurate leaf angle distribution function was got,which is used to optimize the PROSAIL model. In this research,the maize canopy LAI was retrived in Farm No. 852,Heilongjiang Province,through traditional PROSAIL model and optimized PROSAIL model respectively. The main conclusion is as follows： all of the two methods of inversion of LAI have a good correlation with measuredLAI as coefficient is 0. 557 6 and 0. 858 3 respectively,which proved that this model is credibly. But the result of inversed LAI based on unimproved model is low. After optimized model with TLS data,the inversion of LAI estimation accuracy was improved from 26. 53% to 96. 23%. Therefore,it can greatly improve the accuracy of LAI inversion by introducing the TLS point cloud data to improve crops leaf angle distribution function in PROSAIL model.

关 键 词：玉米 PROSAIL模型 叶倾角分布函数 叶面积指数 地基激光雷达 高分一号 

分 类 号：P237[天文地球—摄影测量与遥感]