Construction of LMA retrieval model based on optimal spectral index
Leaf mass per area(LMA)is the most fundamental functional trait of leaves,closely related to many physiological responses of plants.The rapid and accurate estimation of driness in plant leaves is of great significance for monitoring crop growth and health.This paper selects 330 sample data from the lopex93 database to construct the following five typical spectral feature indices,namely Difference Index(DI),Ratio Index(RI),Normalized Difference Vegetation Index(NDVI),Modified Normalized Difference Index(mNDI),and Modified Simple Ratio(mSR),extracts highly sensitive band feature combinations under different LMAs by using correlation matrix method,constructs an inversion model by using linear regression,and evaluates the model by using the coefficient of determination(R2)and Root Mean Square Error(RMSE).The results showed that the variable input combination composed of three characteristic spectral indices,namely RI(1 870,2 280),NDVI(1 870,2 274),and mSR(1 874,2 281)was the optimal.The univariate correlation was greater than 0.8,and the model determination co-efficient is 0.784,the smallest root mean square error is up to 0.001 2.It indicates that the model has high accuracy and can provide useful reference for large-scale monitoring of LMA.
hyperspectral remote sensingLMAestimating modelspectral characteristic index
万方数据
维普
