Estimation of rice leaf angle distribution from geometric features in millimetric-resolution UAV RGB imagery acquired at 45° viewing angle
Abstract
Leaf angle distribution (LAD) is a fundamental determinant of canopy light interception and photosynthetic efficiency, yet its robust quantification remains a major bottleneck. Traditional methods are constrained by either labor-intensive manual measurements or the ill-posed nature of indirect gap probability inversion. Recent advances in deep learning have demonstrated great potential for directly extracting canopy traits from high-resolution imagery. However, whether 2D RGB imagery contains sufficient information to retrieve LAD, a three-dimensional canopy structural trait, is still questionable. We investigated this question by developing a deep learning framework that establishes the relationship between image-derived features and LAD from millimetric-resolution RGB images taken from a UAV. We focused on rice crops representing one of the main cultivated species. A total of 332 rice plants were digitized to construct realistic 3D canopies; we generated 21,600 paired images and GT LAD labels via physically based rendering. In silico experiments demonstrated that deep learning models enable efficient LAD estimation, particularly when taking images inclined at 45 degrees perpendicularly to the rows. Further, our model achieved substantially higher accuracy for average inclination angle (AIA, R-2 = 0.608, RMSE = 1.590 degrees) compared to gap probability-based inversion (R-2 = -0.788, RMSE = 3.397 degrees). Finally, our model showed superior performance from the transplanting to booting stage (R-avg(2) > 0.94 for LAD) when applied to actual observations. Application to over 1000 cultivars across two years confirmed high repeatability, establishing a scalable solution for large-scale monitoring. Interpretability analyses, supported by feature visualization and ablation studies, reveal that the model explicitly extracts geometric features specific to leaf inclination, while disentangling them from confounding canopy variables in millimetric-resolution RGB imagery. This feature-level discrimination underpins the model's superior accuracy and ensures robust generalization in diverse field scenarios.
Related
- Fusion of Multi-Spectral Features from UAVs to Improve the Accuracy of Total Soil Nitrogen Estimation During the Corn Growing Season
- TAPAS: Learned integration of AlphaFold3 confidence and geometric features for TCR-pMHC binding prediction
- Full-length 16S rRNA gene sequencing resolution reveals variable compositional features of medfly gut microbiomes
- A high-throughput UAV millimetric imaging pipeline for plant density phenotyping reveals high heritability in wheat-pea sole crops and intercrops.
- Environmental features or interspecific relations: What determines distribution of Syrian and Great Spotted woodpeckers hybrids?
BibTeX
@article{gao2026estimation,
title = {{Estimation of rice leaf angle distribution from geometric features in millimetric-resolution UAV RGB imagery acquired at 45° viewing angle}},
author = {Yangmingrui Gao and Linyuan Li and Ming Shi and Wenjie Wang and Peiqi Yang and Feng Xiao and Hao Lu and Frédéric Baret and Yanfeng Ding and Shouyang Liu},
journal = {Remote Sensing of Environment},
year = {2026},
month = oct,
doi = {10.1016/j.rse.2026.115697},
url = {https://hal.inrae.fr/hal-05806261v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Estimation of rice leaf angle distribution from geometric features in millimetric-resolution UAV RGB imagery acquired at 45° viewing angle
AU - Yangmingrui Gao
AU - Linyuan Li
AU - Ming Shi
AU - Wenjie Wang
AU - Peiqi Yang
AU - Feng Xiao
AU - Hao Lu
AU - Frédéric Baret
AU - Yanfeng Ding
AU - Shouyang Liu
PY - 2026
DA - 2026/10/19
JO - Remote Sensing of Environment
DO - 10.1016/j.rse.2026.115697
UR - https://hal.inrae.fr/hal-05806261v1
AB - Leaf angle distribution (LAD) is a fundamental determinant of canopy light interception and photosynthetic efficiency, yet its robust quantification remains a major bottleneck. Traditional methods are constrained by either labor-intensive manual measurements or the ill-posed nature of indirect gap probability inversion. Recent advances in deep learning have demonstrated great potential for directly extracting canopy traits from high-resolution imagery. However, whether 2D RGB imagery contains sufficient information to retrieve LAD, a three-dimensional canopy structural trait, is still questionable. We investigated this question by developing a deep learning framework that establishes the relationship between image-derived features and LAD from millimetric-resolution RGB images taken from a UAV. We focused on rice crops representing one of the main cultivated species. A total of 332 rice plants were digitized to construct realistic 3D canopies; we generated 21,600 paired images and GT LAD labels via physically based rendering. In silico experiments demonstrated that deep learning models enable efficient LAD estimation, particularly when taking images inclined at 45 degrees perpendicularly to the rows. Further, our model achieved substantially higher accuracy for average inclination angle (AIA, R-2 = 0.608, RMSE = 1.590 degrees) compared to gap probability-based inversion (R-2 = -0.788, RMSE = 3.397 degrees). Finally, our model showed superior performance from the transplanting to booting stage (R-avg(2) > 0.94 for LAD) when applied to actual observations. Application to over 1000 cultivars across two years confirmed high repeatability, establishing a scalable solution for large-scale monitoring. Interpretability analyses, supported by feature visualization and ablation studies, reveal that the model explicitly extracts geometric features specific to leaf inclination, while disentangling them from confounding canopy variables in millimetric-resolution RGB imagery. This feature-level discrimination underpins the model's superior accuracy and ensures robust generalization in diverse field scenarios.
SN - 0034-4257
SN - 1879-0704
ER -