快速、大面积获取水稻缺氮状况对实现水稻精准施肥具有重要意义。而现有的研究大都集中于利用无人机遥感对水稻进行氮营养诊断，对水稻缺氮程度的定量描述研究较少，同时很少探究临界氮浓度对水稻光谱变化的影响。因此，本研究基于无人机高光谱遥感获取冠层光谱数据、通过田间采样获取水稻氮素含量，研究水稻临界氮浓度曲线构建方法，在此基础上确定水稻缺氮量；以临界氮浓度状态下光谱为标准光谱，分别对光谱反射率数据进行比值与差值变换，并通过连续投影法SPA（successive projections algorithm）对光谱数据进行特征提取，最后以二者提取的特征波段为输入变量，缺氮量为输出变量，分别构建基于多元线性回归MLR（Multivarate Linear Regression）、长短期记忆网络LSTM（Long Short-Term Memory）、极限学习机ELM（Extreme Learning Machine, ELM）与第三代非支配遗传算法优化极限学习机NSGA-III-ELM（The Non-dominated Sorting Genetic Algorithm III Extreme Learning Machine）三种算法的水稻缺氮量反演模型。结果分析表明：1）缺氮量数据与原始光谱的相关性较差，但差值变化与比值变换均能提升其与缺氮量的相关性；2）基于比值光谱与NSGA-III-ELM算法的反演结果最佳，训练集与验证集的R2分别为0.852、0.810，RMSE为0.291、0.308；从光谱数据层面看，比值光谱的反演精度明显优于原始光谱与差值光谱；从算法层面看，基于LSTM算法的模型反演结果过拟合现象严重，反演效果较差；而基于NSGA-III-ELM算法的反演精度明显优于MLR算法与ELM算法的反演精度。因此，基于比值光谱与NSGA-III-ELM算法的反演模型可以对水稻缺氮量进行有效反演，为基于水稻氮营养状况的精准追肥提供了重要的技术支持。

Rapid and large area acquisition of nitrogen (N) deficiency status is important for achieving the optimal fertilization of rice.  Most existing studies, however, focus on the use of unmanned aerial vehicle (UAV) remote sensing to diagnose N nutrition in rice, while there are fewer studies on the quantitative description of the degree of N deficiency in rice, and the effects of the critical N concentration on the spectral changes in rice have rarely been explored.  Therefore, based on the canopy spectral data obtained by remotely-sensed UAV hyperspectral images, the N content in rice was obtained through field sampling.  The construction method of the rice curve for the northeastern critical N concentration was studied, and on this basis, N deficiency was determined.  Taking the spectrum of the critical N concentration state as the standard spectrum, the spectral reflectivity data were transformed by the ratios and differences, and the feature extraction of the spectral data was carried out by the successive projections algorithm (SPA).  Finally, by taking the characteristic band as the input variable and N deficiency as the output variable, a set of multivariate linear regression (MLR), long short-term memory (LSTM) inversion models based on extreme learning machine (ELM), and the non-dominated sorting genetic algorithm III extreme learning machine (NSGA-III-ELM) were constructed.  The results showed two key aspects of this system: 1) The correlation between the N deficiency data and original spectrum was poor, but the correlation between the N deficiency data and N deficiency could be improved by a difference change and ratio transformation; 2) The inversion results based on the ratio spectrum and NSGA-III-ELM algorithm were the best, as the R² values of the training set and validation set were 0.852 and 0.810, and the root mean square error (RMSE) values were 0.291 and 0.308, respectively.  From the perspective of the spectral data, the inversion accuracy of the ratio spectrum was better than the accuracy of the original spectrum or difference spectrum.  At the algorithm level, the model inversion results based on LSTM algorithms showed a serious overfitting phenomenon and poor inversion effect.  The inversion accuracy based on the NSGA-III-ELM algorithm was better than the accuracy of the MLR algorithm or the ELM algorithm.  Therefore, the inversion model based on the ratio spectrum and NSGA-III-ELM algorithm could effectively invert the N deficiency in rice and provide critical technical support for accurate topdressing based on the N status in the rice.