Research Article | Open Access

Volume 2025 |Article ID 0276 | https://doi.org/10.34133/plantphenomics.0276

Estimating Leaf Nitrogen Accumulation Considering Vertical Heterogeneity Using Multiangular Unmanned Aerial Vehicle Remote Sensing in Wheat

Yuanyuan Pan,^1,2 Jingyu Li,¹ Jiayi Zhang,¹ Jiaoyang He,¹ Zhihao Zhang,¹ Xia Yao,¹ Tao Cheng,¹ Yan Zhu,¹ Weixing Cao,¹ and Yongchao Tian ¹

¹National Engineering and Technology Center for Information Agriculture, Engineering and ResearchCenter of Smart Agriculture (Ministry of Education), Key Laboratory for Crop System Analysis andDecision Making (Ministry of Agriculture and Rural Affairs), Jiangsu Key Laboratory for InformationAgriculture, Jiangsu Collaborative Innovation Center for Modern Crop Production, Nanjing AgriculturalUniversity, Nanjing 210095, China
²Jiangsu Academy of Agricultural Sciences Wuxi Branch, Wuxi214174, China

Abstract

The accuracy of leaf nitrogen accumulation (LNA) estimation is often compromised by the vertical heterogeneity of crop nitrogen. In this study, an estimation model of LNA considering vertical heterogeneity of wheat was developed based on unmanned aerial vehicle (UAV) multispectral data and near-ground hyperspectral data, both collected at different view zenith angles (e.g., 0°, −30°, and −45°). Winter wheat plants were evenly divided into 3 layers from top to bottom, and LNA was obtained for the upper, middle, and lower leaf layers, as well as for various combinations of these layers (upper and middle, middle and lower, and the entire canopy, referred to as LNACanopy). The linear regression (LR) and random forest regression (RF) models were constructed to estimate the LNA for each individual leaf layer. Subsequently, models for estimating LNACanopy that considered the impact of vertical heterogeneity (namely, LR-LNASum and RF-LNASum) were established based on the relationships between LNACanopy and LNA in different leaf layers. Meanwhile, LNA models that did not consider the effect of vertical heterogeneity (LR-LNAnon and RF-LNAnon) were used for comparative validation. The validation datasets consisted of UAV-simulated data from hyperspectral reflectance and UAV-measured data. Results showed that LNASum models had markedly higher accuracy compared to LNAnon. The optimal scheme for estimating LNACanopy was the combination of the upper, middle, and lower layers based on the normalized difference red edge index. Among these models, RF-LNASum demonstrated higher accuracy than LR-LNASum, with a validation relative root mean square error of 19.3% and 17.8% for the UAV-measured and simulated dataset, respectively.