Optimizing UAV Flight Altitude and Image Overlap for RGB Image Post-Processing in Cacao Cultivation
Year: 2026
Type: article
Abstract: The use of UAVs in agriculture provides an efficient approach for acquiring high-resolution imagery for crop monitoring and precision agriculture. This study evaluated the effects of flight altitude and image overlap on RGB photogrammetric reconstruction in a cacao plantation using a DJI Air 2S and Pix4Dmapper. A factorial design combined altitudes of 20, 25, and 30 m AGL, forward overlaps of 70%, 80%, and 90%, and side overlaps of 70% and 80%, at 5 m/s. Each configuration was flown four times over the same 1-ha plot, and image sets were processed using Initial Processing Full, Rapid, and Custom workflows. Theoretical GSD was approximately 0.57 cm/pixel at 20 m, 0.72 cm/pixel at 25 m, and 0.86 cm/pixel at 30 m. Factorial models included repetition as a blocking factor and HC3 heteroscedasticity-robust inference. In Full and Custom, altitude, forward overlap, and the Altitude × Forward and Forward × Side interactions were significant; Custom also showed a significant three-way interaction. In Rapid, only altitude and repetition were significant. Configuration T13 (30 m AGL, 90% forward overlap, and 80% side overlap) achieved the highest observed dataset quality (100%) with Full processing. This configuration provided the most favorable observed reconstruction performance while retaining sub-centimeter theoretical GSD. Without ground control points or independent checkpoints, these findings describe relative reconstruction quality rather than absolute geodetic accuracy.
Source: Agrociencia Uruguay
Language: English
Subfield: Geology
Field: Earth and Planetary Sciences
Domain: Physical Sciences
SDG: Zero hunger
Open Access status: diamond
OpenAlex ID: W7236570167
