Effects of uav imagery overlap on photogrammetric data quality for construction applications

UAV photogrammetry is a remote sensing method that safely collects spatial data, used in construction inspection, progress tracking, and infrastructure evaluation. This technique generates multiple output products, including point clouds and 3D textured mesh models, using structure-from-motion (SfM) algorithms on UAV-captured photos. Despite its widespread application, the required percentage of aerial imagery overlap—crucial for determining UAV flight patterns, post-processing time, and output quality—remains empirically unsupported and thus arbitrarily defined. Different SfM software can yield varying quality outputs from the same data set, emphasizing the need to evaluate these discrepancies, particularly how overlap percentages affect output quality. This study gathered UAV photogrammetric data from a construction site, created three imagery datasets with varying overlaps, and processed them with two SfM tools, Pix4Dmapper and DroneDeploy. Analysis methods included surface and volume density checks, visual inspections, and defect analysis. Findings suggest that while image overlap percentage crucially impacts output quality, significant differences between high and medium overlaps compared to low overlap were not observed. Optimal results were achieved with a manual UAV operation at 15 meters altitude, 55–60% frontal and 45% side overlaps, a ground sampling distance of 0.4–0.7 cm/pixel, and good lighting, without the need for Ground Control Points.