摘要
FY-3B卫星具有观测频次高、成像范围广等特点,可为玉米叶面积指数(leaf area index,LAI)反演研究提供长时序观测数据。LSTM((long short term memory)算法具有在多时相数据中提取时间特征的能力,能解决光谱数据与LAI之间复杂的非线性问题。本文基于辽宁省锦州市近地实测春玉米LAI和反射率光谱数据,利用光谱响应函数模拟FY-3B多光谱波段,结合与春玉米LAI相关性较高的28种植被指数,应用LSTM算法建立不同隐藏层的预测模型,并与偏最小二乘法(partial least-squares regression,PLSR)建立的模型进行预测精度对比。结果表明:隐藏层的层数对LSTM模型的拟合效果有较大影响,三层LSTM模型将LAI估算精度的决定系数由0.8183(单层LSTM)、0.7800(PLSR)提升至0.8692;对应地,将均方根误差由0.5091、0.4906降低至0.3726,模型精度提升明显。
The FY-3 B satellite has the characteristics of a high frequency of observation and wide imaging range,which can provide long-term observation data for maize leaf area index(LAI)inversion research.Long short-term memory(LSTM)algorithm has the ability to extract temporal features from multi-period data and solve complex nonlinear problems between spectral data and LAI.The study was conducted based on LAI and reflectance spectrum data of spring maize in Jinzhou City,Liaoning Province measured near the ground.To construct the LAI inversion model,the spectral response functions were used to simulate the FY-3 B multi-spectral band data combined with 28 vegetation indices highly correlated with spring maize LAI.The inversion models were conducted using LSTM of different hidden layers,and the accuracies of the LSTM models were compared with the accuracy of the partial least-squares regression(PLSR)model.The results showed that the number of hidden layers greatly influences the fitting ability of the LSTM model.The three-layer LSTM model increased the LAI estimation accuracy R~2 from 0.8183(single-layer LSTM),0.7800(PLSR)to 0.8692;correspondingly reducing the RMSE from 0.5091,0.4906 to 0.3726.In short,the accuracy of the model was significantly improved.
作者
张霞
陶诗语
张茂
Zhang Xia;Tao Shiyu;Zhang Mao(Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing100101,China;University of Chinese Academy of Sciences,Beijing100049,China)
出处
《吉林大学学报(地球科学版)》
CAS
CSCD
北大核心
2022年第6期2071-2080,共10页
Journal of Jilin University:Earth Science Edition
基金
国家重点研发计划项目(2017YFC1502802)
中国科学院战略性先导科技专项(XDA28080502)
风云卫星应用先行计划专项(FY-APP-2021.0302)