基于PROSAIL模型的青贮玉米叶面积指数反演

Inversion of Leaf Area Index of Silage Corn Based on PROSAIL Model

精准且高效地估算区域内的玉米叶面积指数(LAI),对于田间管理决策、地物产量预测以及实施精准农业具有至关重要的意义。针对多尺度、大范围遥感反演中存在的尺度效应、精度低、普适性差等问题,本文以张掖市民乐县青贮玉米实验田为研究区,选取青贮玉米为研究对象,基于Landsat-8高光谱和Modis多光谱遥感影像,并结合地面实测数据。通过对PROSAIL模型的输入参数进行局部和全局敏感性分析,构建出青贮玉米在多个生育期内的冠层反射率-LAI的查找表和最小寻优代价函数的反演策略,确定研究区域的最佳LAI反演模型,并利用青贮玉米不同生育期内的实测值完成了反演结果的精度验证及线性拟合。结果表明:LAI反演结果总体较好,拟合精度较高,与实测值之间有较强的相关性,拔节期、抽雄期、成熟期最优决定系数R2分别为0.85、0.91、0.90;均方根误差(RMSE)分别为0.35、0.58、0.51。因此,基于多源高光谱遥感数据结合PROSAIL模型的反演策略可为作物参数反演提供新的科学依据和方法。

Accurately and efficiently estimating corn LAI data within a region is of crucial importance for field management decisions,predicting land yield,and implementing precision agriculture.In response to the problems of scale effect,low accuracy,and poor universality in multi-scale and large-scale remote sensing inversion,taking the silage corn experimental field in Minle County,Zhangye City as the research area,silage corn was selected as the research object,based on Landsat-8 hyperspectral and Modis multispectral remote sensing images,combined with ground measured data.Through local and global sensitivity analysis of the input parameters of the PROSAIL model,the lookup table of canopy reflectance-LAI of silage corn in multiple growth periods and the inversion strategy of the minimum optimization cost function were constructed,and the optimal LAI inversion model for the study area was determined.The accuracy verification and linear fitting of the inversion results were completed by using the measured values in different growth periods of silage corn.The results showed that the inversion results of LAI were generally good,with high fitting accuracy and strong correlation with the measured values.The optimal determination coefficients R~2 for the jointing stage,tasseling stage,and maturity stage were 0.85,0.91,and 0.90,respectively.The root mean square error(RMSE)were 0.35,0.58,and 0.51,respectively.Therefore,the inversion strategy based on multi-source hyperspectral remote sensing data combined with the PROSAIL model can provide scientific basis and methods for crop parameter inversion.