为了提高无人机遥感对冬小麦叶面积指数(leaf area index,LAI)反演模型的精度与泛化能力,该研究利用无人机搭载多光谱相机获取不同氮素处理和不同复种方式的冬小麦生长实测数据,结合PROSAIL辐射传输模型生成包含机理信息的模拟数据,基...为了提高无人机遥感对冬小麦叶面积指数(leaf area index,LAI)反演模型的精度与泛化能力,该研究利用无人机搭载多光谱相机获取不同氮素处理和不同复种方式的冬小麦生长实测数据,结合PROSAIL辐射传输模型生成包含机理信息的模拟数据,基于不同组合方式建立了5种LAI反演混合数据集,结合多种机器学习方法,以期构建经验与机理相结合的LAI高精度反演模型。由于LAI反演受近红外波段(near infrared,NIR)反射率影响大,该研究筛选7种与NIR波段相关的植被指数提取冬小麦光谱特征,构建与混合数据集LAI的相关系数矩阵,进一步探究不同光谱特征对冬小麦LAI的影响程度。在此基础上,采用具有代表性和普适性的4种机器学习方法,即贝叶斯岭回归模型、线性回归模型、弹性网络模型和支持向量回归模型,构建不同冬小麦LAI反演模型,用以评估基于半经验半机理数据反演冬小麦LAI的可行性,进一步探索其对不同氮素水平和复种方式的冬小麦长势评估能力。结果表明:1)筛选的与NIR波段相关的植被指数与冬小麦LAI之间存在较强的相关性,其中归一化差异植被指数、增强植被指数、归一化差异红边指数、比值植被指数、红边叶绿素植被指数、土壤调节植被指数与LAI呈正相关,结构不敏感色素植被指数与LAI呈负相关;2)辐射传输模型中体现了冬小麦LAI影响太阳光线传播的机理,结果表明,与实测数据混合建立的模型,具有较强的鲁棒性和泛化能力。相比于其他3种模型,支持向量回归模型在各种数据组合下均取得了较好的LAI预测性能,在C1、C2、C3、C4这4种训练-测试组合的训练集中R^(2)依次为0.86、0.87、0.88、0.91,RMSE依次为0.47、0.45、0.45、0.41;在测试集的R^(2)依次为0.85、0.19、0.89、0.87,RMSE依次为0.45、1.31、0.49、0.50;3)使用支持向量机生成试验区LAI反演图,对4种氮素水平和2种复种方式的冬小麦长势评估,结果表明,适当的施加氮素处理能提高冬小麦LAI值,麦-豆复种方式下的冬小麦LAI值普遍高于麦-玉复种的LAI值。该研究为冬小麦LAI的反演提供了一种有效的方法,并为高效评估冬小麦长势研究提供了参考。展开更多
In this paper, we calculated multi-scale residual geoid anomalies with the method of geoid separation processing, according to EGM2008 ultra-high order gravity field model, remove-restore technique and Stokes integral...In this paper, we calculated multi-scale residual geoid anomalies with the method of geoid separation processing, according to EGM2008 ultra-high order gravity field model, remove-restore technique and Stokes integral. The East Asian continental margin was selected as the study area. The residual geoid anomalies have been calculated by programming. On the basis of residual geoid anomalies at various orders, the interlayer geoid anomalies at different depths were calculated to depict the spatial distribution characteristics of the residual geoid. Finally, we conducted a detailed geophysical interpretation for the study area according to the geoid anomalies in combination with other geophysical datasets. Four conclusions can be outlined as follows: 1) it is impracticable that geoid anomalies are used in the interpretation of the shallow objects due to the influence of the terrain; 2) the anomalies of residual geoid can reflect the intensity of small-scale mantle convection in the asthenosphere; 3) the interlayer geoid anomalies can reflect the magmatic activities associated with the mantle convection and mantle plume in different scales; 4) the study of the geoid may provide an approach for the research of the subduction zone, mantle convection and mantle plume.展开更多
文摘为了提高无人机遥感对冬小麦叶面积指数(leaf area index,LAI)反演模型的精度与泛化能力,该研究利用无人机搭载多光谱相机获取不同氮素处理和不同复种方式的冬小麦生长实测数据,结合PROSAIL辐射传输模型生成包含机理信息的模拟数据,基于不同组合方式建立了5种LAI反演混合数据集,结合多种机器学习方法,以期构建经验与机理相结合的LAI高精度反演模型。由于LAI反演受近红外波段(near infrared,NIR)反射率影响大,该研究筛选7种与NIR波段相关的植被指数提取冬小麦光谱特征,构建与混合数据集LAI的相关系数矩阵,进一步探究不同光谱特征对冬小麦LAI的影响程度。在此基础上,采用具有代表性和普适性的4种机器学习方法,即贝叶斯岭回归模型、线性回归模型、弹性网络模型和支持向量回归模型,构建不同冬小麦LAI反演模型,用以评估基于半经验半机理数据反演冬小麦LAI的可行性,进一步探索其对不同氮素水平和复种方式的冬小麦长势评估能力。结果表明:1)筛选的与NIR波段相关的植被指数与冬小麦LAI之间存在较强的相关性,其中归一化差异植被指数、增强植被指数、归一化差异红边指数、比值植被指数、红边叶绿素植被指数、土壤调节植被指数与LAI呈正相关,结构不敏感色素植被指数与LAI呈负相关;2)辐射传输模型中体现了冬小麦LAI影响太阳光线传播的机理,结果表明,与实测数据混合建立的模型,具有较强的鲁棒性和泛化能力。相比于其他3种模型,支持向量回归模型在各种数据组合下均取得了较好的LAI预测性能,在C1、C2、C3、C4这4种训练-测试组合的训练集中R^(2)依次为0.86、0.87、0.88、0.91,RMSE依次为0.47、0.45、0.45、0.41;在测试集的R^(2)依次为0.85、0.19、0.89、0.87,RMSE依次为0.45、1.31、0.49、0.50;3)使用支持向量机生成试验区LAI反演图,对4种氮素水平和2种复种方式的冬小麦长势评估,结果表明,适当的施加氮素处理能提高冬小麦LAI值,麦-豆复种方式下的冬小麦LAI值普遍高于麦-玉复种的LAI值。该研究为冬小麦LAI的反演提供了一种有效的方法,并为高效评估冬小麦长势研究提供了参考。
基金financially supported by the National Natural Science Foundation of China(No.41606044)the Special Fund for Ocean Scientific Research in the Public Interest(No.201305029-02)
文摘In this paper, we calculated multi-scale residual geoid anomalies with the method of geoid separation processing, according to EGM2008 ultra-high order gravity field model, remove-restore technique and Stokes integral. The East Asian continental margin was selected as the study area. The residual geoid anomalies have been calculated by programming. On the basis of residual geoid anomalies at various orders, the interlayer geoid anomalies at different depths were calculated to depict the spatial distribution characteristics of the residual geoid. Finally, we conducted a detailed geophysical interpretation for the study area according to the geoid anomalies in combination with other geophysical datasets. Four conclusions can be outlined as follows: 1) it is impracticable that geoid anomalies are used in the interpretation of the shallow objects due to the influence of the terrain; 2) the anomalies of residual geoid can reflect the intensity of small-scale mantle convection in the asthenosphere; 3) the interlayer geoid anomalies can reflect the magmatic activities associated with the mantle convection and mantle plume in different scales; 4) the study of the geoid may provide an approach for the research of the subduction zone, mantle convection and mantle plume.
