Testing measurement strategies using unmanned aerial vehicles to optimize airborne mission parameters

• Variation in UAV flight altitude and mission count did not inherently improve coordinate accuracy. • All photogrammetric products were computed without ground control points using a commercial UAV system. • Planar coordinates retained high accuracy across datasets, with RMSE for X and Y generally between 2 and 5 cm. • Single mission solutions yielded poor vertical accuracy, with RMSE for Z frequently between 10 and 30 cm. • Averaging coordinates from three independent missions reduced Z RMSE to about 5 cm, comparable to planar accuracy. The aim of this study was to identify the optimal UAV-based measurement strategy for improving that involves combining data from flight missions conducted at different altitudes the accuracy of the coordinates of the surveyed objects based on photogrammetric products. The research was conducted using a commercial DJI Matrice 300 RTK drone equipped with a Zenmuse P1 camera. Importantly, the resulting photogrammetric products were generated without the use of external ground control points (GCPs). The results demonstrated that planar coordinates (X and Y) were determined with consistently high accuracy across all analysed datasets, with root mean square error (RMSE) values falling within the range of 2–5 cm. The most favourable outcome was obtained from the mission conducted at an altitude of 120 m—the lowest tested configuration—which yielded RMSE values of 1.88 cm for the X-coordinate and 3.42 cm for the Y-coordinate. In contrast, the accuracy of elevation (Z) measurements was substantially lower. Photogrammetric products derived from single-flight missions exhibited unsatisfactory vertical precision, with RMSE values ranging from approximately 10 to 30 cm. Integrating image data from varying altitudes did not improve height accuracy and, in some cases, increased RMSE to 35.28 cm. A particularly effective approach to enhancing vertical accuracy involved the arithmetic averaging of coordinates derived from three independent UAV missions (DATASET 4). This method resulted in a substantial reduction of the root mean square error (RMSE) for the Z coordinate, reaching a value of 5.15 cm. It is noteworthy that this level of vertical accuracy is comparable to that achieved for planar coordinates (X: 1.81 cm; Y: 4.76 cm), and it signifies a substantial enhancement over the results obtained from individual flights, where the vertical root mean square error (RMSE) reached up to 29.38 cm. A significant finding of this study is that neither the number of UAV missions nor variations in flight altitude directly improved positional accuracy. Conversely, the enhancement in vertical precision can be attributed to the implementation of a suitable data integration strategy, namely the averaging of coordinates across datasets that have been processed independently. This straightforward yet robust method is especially valuable for generating high-resolution elevation data in the absence of external ground control points.