Integrating UAV LiDAR-Derived Structural and Spectral Data for Improved Aboveground Biomass Estimation in Konjac
Smart Agricultural Technology · 2026-11 · 102751
DOI: 10.1016/j.atech.2026.102751 ↗Kurz gefasst
Fachartikel in der Zeitschrift Smart Agricultural Technology, 2026-11. Verfasst von Yong Chen und 8 weiteren Autor:innen. Die Arbeit nennt 68 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Accurate estimation of aboveground biomass (AGB) in konjac is crucial for crop growth monitoring, field management, and yield prediction. UAV LiDAR enables the acquisition of three-dimensional crop structural information, offering a novel technological pathway for AGB estimation. However, current research in this domain has predominantly centered on LiDAR-derived structural metrics (SM), while the potential contribution of RGB vegetation indices (VIs) derived from point clouds remains largely unexplored. To address this, we integrate SM and VIs from a single UAV LiDAR sensor, and apply two random forest based feature selection methods (VSURF and Boruta) to identify important predictors. Subsequently, the selected metrics are employed to evaluate the performance of four regression algorithms (support vector regression [SVR], random forest regression [RFR], multilayer perceptron [MLP], and a deep neural network [DNN]) in estimating the AGB of individual konjac plants. Our results revealed that plant height derived from UAV LiDAR agreed well with field measurements (Pearson’s r = 0.90, RMSE = 0.03 m) and served as a significant structural metric for predicting AGB. Compared with Boruta, VSURF more effectively reduced feature redundancy while retaining information relevant to AGB prediction. Over two growing seasons, the SVR model integrating SM+ VIs selected by VSURF demonstrated the most stable overall performance (R²=0.69, MAE = 0.78 g·plant⁻¹). Additionally, the DNN model integrating SM+VIs achieved the highest accuracy during the seedling stage (R² = 0.79, MAE = 0.59 g·plant⁻¹). Our findings indicate that combining SM and VIs derived from UAV-based LiDAR sensors enhances the accuracy of AGB estimation for individual konjac plants. This approach offers a new perspective for estimating growth parameters in other crops using a single UAV-based LiDAR sensor.
Zitieren
APA
Chen, Y., Hu, K., He, W., Zhao, Y., Kou, W., Xu, W., Wang, H., Qi, H., & Lu, N. (2026). Integrating UAV LiDAR-Derived Structural and Spectral Data for Improved Aboveground Biomass Estimation in Konjac. Smart Agricultural Technology, 102751. https://doi.org/10.1016/j.atech.2026.102751
Vancouver
Chen Y, Hu K, He W, Zhao Y, Kou W, Xu W, et al. Integrating UAV LiDAR-Derived Structural and Spectral Data for Improved Aboveground Biomass Estimation in Konjac. Smart Agricultural Technology. 2026:102751. doi:10.1016/j.atech.2026.102751
BibTeX
@article{Chen2026,
author = {Chen, Yong and Hu, Kunrong and He, Weichang and Zhao, Ya’nan and Kou, Weili and Xu, Weiheng and Wang, Huan and Qi, Hongjuan and Lu, Ning},
title = {Integrating UAV LiDAR-Derived Structural and Spectral Data for Improved Aboveground Biomass Estimation in Konjac},
journal = {Smart Agricultural Technology},
year = {2026},
pages = {102751},
doi = {10.1016/j.atech.2026.102751}
}
EUVIMED