UAV flight path impacts on three-dimensional photogrammetric product accuracy

摘要

The accuracy of unmanned aerial vehicle (UAV) products is affected by many parameters, such as ground control points (GCPs), flight plans, overlap rates, flight altitude, etc. Flight patterns could be another factor that influences UAV product accuracy, but is not given as much attention in the literature. The impact of seven combinations derived from Single (S), Double (D), and Circular (C) flight patterns on point formation is analysed in this study. It was also evaluated the positional accuracy and number of points in both horizontal and vertical planes within three-dimensional (3D) models. For each model, 15 common comparison surfaces (CSs) (10 vertical and 5 horizontal) were selected, and a point cloud generated by a robotic total station was used as a reference to generate reference planes (RPs), which were utilized to compute surface scoring and root mean square (RMS) error of the models. Ultimately, two separate analyses were conducted. Each model's demonstration level of the chosen horizontal and vertical planes is rated in the first analysis. In the subsequent analysis, the RMS values were determined based on the distance between the model points and the generated RPs. The study's results indicate that the D pattern offers the best option for only one flight. Furthermore, the D+C flight combination yielded the best results among the evaluated models when multiple flights were conducted.