基于缨帽三角-植被等值线分布模式的多类型作物叶面积指数反演  

Inversion of the leaf area index for the multiple types of crops based on the tasselled cap-vegetation isoline patterns

作　　者：陈子涵 鲁蕾[1] 柴东岳 岳玮 崔霞[1] 周孝明 王靖雯 CHEN Zihan;LU Lei;CHAI Dongyue;YUE Wei;CUI Xia;ZHOU Xiaoming;WANG Jingwen(College of Earth and Environment Sciences,Lanzhou University,Lanzhou 730000,China;State Key Laboratory of Infomation Engineering in Surveying,Mapping and Remote Sensing,Wuhan University,Wuhan 430072,China;Gansu Academy of Eco-environmental Sciences,Lanzhou 730000,China;School of Civil Engineering,Lanzhou University of Technology,Lanzhou 730050,China)

机构地区：[1]兰州大学资源环境学院,兰州730000 [2]武汉大学测绘遥感信息工程全国重点实验室,武汉430072 [3]甘肃省生态环境科学设计研究院,兰州730000 [4]兰州理工大学土木工程学院,兰州730050

出　　处：《农业工程学报》2025年第7期182-191,共10页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家自然科学基金项目(42061056);甘肃省科技计划项目(21JR7RA504,22JR5RA443);兰州大学中央高校基本科研业务费专项(lzujbky-2023-35)。

摘　　要：获取作物叶面积指数(leaf area index, LAI)及其动态变化信息对作物长势监测和产量估测等具有重要意义。基于辐射传输模型的物理模型反演是LAI遥感反演最常用的方法,但该方法存在反演值不唯一的问题。此外,现有研究通常只针对单一作物类型,缺乏针对多类型作物的精度较高的LAI反演算法。该研究以玉米和水稻为主要作物的农田为例,基于PROSAIL模型模拟数据集,通过分析不同类型作物的缨帽三角-植被等值线分布模式,将植被覆盖度作为先验知识,构建用于反演多类型作物的LAI反演查找表,将其用于多时相GF-1 WFV(wide-field view)影像,反演获得整个生长季不同生长时期的LAI,并利用地面实测数据进行验证。研究结果显示:将植被覆盖度作为先验知识构建的查找表反演的LAI和实测值相关性较显著(R^(2)=0.60),均方根误差(RMSE)为0.75,反演的整个生长期LAI的变化趋势与实测LAI的变化趋势一致。而由未加入先验知识的查找表反演的LAI值和实测值的R^(2)为0.47,RMSE为0.85。该研究表明,基于缨帽三角-植被等值线分布模式,在构建涉及多类型作物的农田LAI反演的查找表中引入先验知识,能够显著提高LAI反演的精度,有效获得作物的LAI信息。Leaf area index(LAI)is one of the most important indictors for the crop growth and yield estimation.The LAI dynamics can also greatly contribute to the crops monitoring.The inversion of radiative transfer model has been commonly used for the LAI inversion from remote sensing images.However,the inversed value cannot be unique after inversion.Additionally,previous studies of LAI inversion have concentrated mainly on the individual crop species.The LAI values cannot be concurrently inverted from the multiple types of crops.Hence,it is necessary to estimate the LAIs of different crop species in the mixed agricultural landscapes.This study aims to inverse the LAIs of the maize and rice in fields.The PROSAIL(PROSPECT and Scattering by Arbitrary Inclined Leaves)model was utilized to construct the simulated dataset.The spectral signatures and biophysical characteristics of maize and rice were then collected under varying environmental conditions at different growing stages.The tasselled cap-vegetation isoline pattern was obtained for the different type of crops.A look-up table(LUT)was developed to invert the LAI of maize and rice.The vegetation cover fraction(FVC)was taken as the prior knowledge.Another LUT without FVC was also developed to invert the LAI from the same GF-1 WFV images.Then,the LUT was applied into the multiple temporal images of maize and rice in fields.These images were captured by the Wide Field View(WFV)sensor onboard the Gaofen-1(GF-1)satellite.The entire growing season of the maize and rice was selected to provide the continuous monitoring of LAI dynamics.The inversed LAIs were validated using ground-based measurements.The results showed that there was the significant correlation between the inversed LAI by the LUT-FVC and the ground measurements,with a coefficient of determination(R^(2))of 0.60 and a root mean square error(RMSE)of 0.75.The consistent trend was also found with the great variations in the measured LAIs during the whole growing season.By contrast,a R^(2) of 0.47 and an RMSE of 0.85 fo

关 键 词：模型 反演 叶面积指数 PROSAIL 查找表 GF-1 WFV 

分 类 号：S-3[农业科学]