Generating to Three Dimensional Models from Taken Photos in Vertical Position with Unmanned Aerial Vehicles: Aksaray University Campus Mosque

Abstract

Unmanned aerial vehicles (UAVs) applications included archaeology, construction engineering, agricultural catchment and forestry. Unmanned aerial vehicle UAVs (Rotary or fixed wing), capable of performing the photogrammetric data acquisition with amateur or digital cameras, can fly in manual, semi-automated and autonomous techniques. Following a typical photogrammetric workflow, three dimensional results like Digital Surface Models or Digital Terrain Models (DSM/DTM), contours, textured 3D models, vector information, etc. can be produced, even on large areas. In this study, a three dimensional model of the campus mosque of Aksaray University was formed with the vertical photos of UAV. Two different resolution cameras at 50 and 100 m altitude were obtained with 80% and 60% overlap rates for both altitudes. Using the PhotoScan software, a four-faced three-dimensional model of the mosque was created. 3D model generated for both data in lowest and medium accuracy for two different areas, and then total errors for all assessments were compared to know which altitude is the best to generating 3D model, problems in the model were discussed and the edges and elevations measured by the geodesic model were compared with those measured by the obtained model. The difference was found to vary between 2-50 cm after the comparison. The average error of the flight was found to be 4.85 cm for 50 m flight height and 5.16 cm for 100 m flight height. Because the Unmanned Aerial Vehicle used cannot take an oblique picture, the Mosque could not be fully modelled. The three-dimensional modelling gaps have been tried supplied with pictures taken from the ground. In the case of unmanned aerial vehicles, it is seen that three dimensional models can be made with appropriate accuracy in case that such structures are suitable picture taking.