Monitoring the Nitrogen Nutrition Index Using Leaf-Based Hyperspectral Reflectance in Cut Chrysanthemums

Precise nitrogen supply is crucial for ensuring the quality of cut chrysanthemums (Chrysanthemum morifolium Ramat.). The nitrogen nutrition index (NNI) serves as an important indicator for diagnosing crop nitrogen (N) nutrition. Hyperspectral remote sensing (HRS) technology has been widely used in monitoring crop N status due to its rapid, accurate, and non-destructive capabilities. However, its application in estimating the NNI of cut chrysanthemums has received limited attention. Therefore, this study aimed to use HRS to accurately determine the cut chrysanthemum NNI, thereby providing valuable guidance for managing N fertilization. During several key growth stages, a hyperspectral spectroradiometer was used to capture hyperspectral reflectance data (350–2500 nm) from three leaf layers. Subsequently, cut chrysanthemum canopies were sampled for aboveground biomass (AGB) and plant nitrogen concentration (PNC). The collected AGB and PNC data were then utilized to fit the critical N (Nc) dilution curve of cut chrysanthemums using a Bayesian hierarchical model, enabling the calculation of the NNI. Finally, spectral indices and partial least squares regression (PLSR) were used to establish the NNI estimation model for cut chrysanthemums. The results showed that the Nc dilution curve of the cut chrysanthemums was Nc = 5.401 × AGB−0.468. The first leaf layer (L1) proved to be optimal for estimating cut chrysanthemum NNI. Additionally, a newly proposed two-band spectral index, DVI-L1 (R1105, R700), demonstrated moderate predictive capabilities for the NNI of cut chrysanthemums (R2 = 0.5309, RMSE = 0.3210). Compared with the spectral index-based NNI estimation model, PLSR-L1 showed the best performance in estimating the cut chrysanthemum NNI (R2 = 0.8177, RMSE = 0.2000). Our results highlight the rapid NNI prediction potential of HRS and its significance in facilitating precise N management in cut chrysanthemums.