Flight Plan Optimization, Scanning and 3D Reconstruction Using an Unmanned Aerial Vehicle

In the domain of aerial photogrammetry, a prominent challenge arises from the absence of standardized procedures in the data integration phase, which constitutes the initial step in the process of three-dimensional (3D) reconstruction. This study scrutinizes the impact of the camera-to-subject distance during recording, varying overlap percentages in photography, and distinctions among three methodologies for programming flight plans. The study made use of a DJI drone Phantom 3 in conjunction with Pix4Dmapper to conduct scanning, 3D modeling and finding optimal flight plan. The primary emphasis was on the historical town of Medvedgrad, situated near Zagreb, Croatia. The goal of this experiment is to bring awareness of how should large 3D objects be scanned and what should the protocol for 3D remodeling be, with the goal of improving workflow efficiency. This involves maintaining a quantitatively assessed quality determined by both the total count of building elements (2D key points for pictures and 3D for point clouds). The study's findings indicate the feasibility of recording data at various distances without any significant problems in the term of 2D elements. The experiment establishes clear connection between image overlap and 3D key points. As a result, the authors recommend adopting a novel flight plan that incorporates optimized parameters derived from this experiment, potentially enhancing the overall efficiency of the aerial photogrammetry process.