Year: 2026
Type: article
Abstract: This study aimed to assess the applicability of plant height estimation in castor bean using remotely piloted aircraft (RPA) imagery as a high-throughput phenotyping approach. RGB images (20 MP; DJI Mavic 3 Enterprise Multispectral; GSD = 1.84 cm pixel–1; 40 m flight altitude) were acquired at four phenological stages for four genotypes. Digital elevation models and orthomosaics were generated in Agisoft Metashape (Educational version), and individual plant height was extracted using FieldImageR in QGIS. Estimated heights showed strong agreement with direct field measurements, particularly at 118 days after sowing (DAS), where the smallest bias (–0.07 m), lowest RMSE (0.11 m), and high correlation (r = 0.94; p < 0.001) were observed. Limits of agreement derived from a mixed-effects framework indicated reduced prediction error during this stage. Estimation accuracy improved when plants exceeded 1.2 m in height and exhibited a denser canopy structure, corresponding to the production stage. Overall, RGB imagery tended to underestimate plant height. These findings highlight the importance of aligning flight parameters, especially GSD and altitude, with plant phenology, as well as considering environmental conditions that may influence image reconstruction accuracy.
Institution: Universidade Federal do Recôncavo da Bahia
Language: English
Subfield: Ecology
Field: Environmental Science
Domain: Physical Sciences
SDG: Zero hunger
Open Access status: hybrid
Estimated APC paid: $4,500
Award: G361406575
OpenAlex ID: W7235526079
