论文部分内容阅读
以杨树-小麦间作系统为研究对象,采用Field Spec Pro FR2500地物光谱仪采集的小麦叶层光谱180份,选取基于单作麦田冠层氮含量估测模型中精度较高的9个指数用于估测林农复合系统中小麦冠层氮的含量,探讨不同水平枯落物覆盖量、林分密度、施肥量条件下小麦叶层氮含量的光谱特征。随机选取116份样本作为训练集基以9个指数分别建立估测模型,其余48份样本作为预测集对估测模型进行适应性检验。结果表明:FDNDNI、SDr-SDb两种指数的P-R2与C-R2达到了0.839、0.777与0.844、0.758,此系统预测杨麦间作系统中小麦叶片冠层的氮含量的精度较高,其余7个指数预测精度不理想。以9个指数所建立的估测模型的精度均低于其对单作麦田氮含量的估测精度。
Taking poplar-wheat intercropping system as the research object, 180 spectral lines of wheat leaf collected by Field Spec Pro FR2500 land-based spectrometer were used to select 9 indexes with high precision based on single-cropping canopy nitrogen content estimation model Estimating the canopy nitrogen content in the wheat-agroforestry system and discussing the spectral characteristics of wheat leaf nitrogen content under different levels of litter coverage, stand density and fertilizer application. Randomly select 116 samples as the training set base with 9 indices to establish the estimation model respectively, and the remaining 48 samples as the prediction set to fit the estimation model. The results showed that the P-R2 and C-R2 of FDNDNI and SDr-SDb reached 0.839, 0.777 and 0.844, 0.758 respectively. The accuracy of this system in predicting the nitrogen content of wheat leaf canopy in the Yangmai intercropping system was high. The forecast accuracy of the seven indices is not satisfactory. The accuracy of the estimation model established by nine indices is lower than the estimation accuracy of the nitrogen content of monoculture wheat field.