The effect of ground control point network architecture and distribution on the accuracy of UAV image maps.

Understanding the impact of unmanned aerial vehicle (UAV) photogrammetric network design parameters on 3D reconstruction quality is crucial for achieving optimal spatial information production from UAV images in a UAV-based project, considering existing conditions and limitations. The purpose of this research is to assess the impact of ground control point network architecture and distribution on the accuracy of spatial information and maps generated from drone images. The design and distribution of the ground control point network have been examined in two modes and three distinct scenarios for this aim. Control points ranging from 4 to 42 points were designed just in the side models in the first case (A) and both around and in the center position of the photogrammetry block in the second case (B). According to the data, mode B had the highest measurement accuracy in both urban and non-urban regions, as well as the best results and the least error in each of those scenarios. The findings showed that the average accuracy loss in UAV-based outputs in the first scenario is at least 12% in the arrangement of A and B control points. Furthermore, the accuracy loss in the second and third scenarios is at least 0.59% and 0.53%, respectively, when compared to the first scenario. In general, it is not suggested to construct and expand control point distances more than 30 times GSD, and if these distances are exceeded, control points must be designed in central block models.