Unmanned Systems Technology 022 | XOcean XO-450 l Radar systems l Space vehicles insight l Small Robot l BMPower FCPS l Prismatic HALE UAV l InterDrone 2018 show report l UpVision l Navigation systems

86 the changeable weather. In general, UpVision’s flights were plotted to align with the direction of the wind, but that proved challenging to predict accurately. The mapping was routed in rows, the flights obviously being faster with the wind behind the UAV or slower when flying into the wind. The team generally found that the second morning flight each day consumed far more battery power as the morning turned into afternoon and the winds became stronger. As a result, flight speeds varied between 30 and 70 kph, but thanks to the size of the overlays, and the camera exposure being set according to GPS coordinates, there were no spaces in the mapping. “In Mongolia, the best weather conditions for aerial mapping occur between May and July because of the low rainfall, but owing to the geography and altitude, changes in weather were still considerable,” Karas notes. “On the first day of mapping, the temperature was more than 20 C; on the second day it was snowing,” he says. “The snow melted quickly, and for the third and last day of flying, we were back to 20 C. It took us two days to adapt to the local conditions and identify the best ones and parts of the days for the mapping.” Seven of the nine flights ended by landing on the more eastern of the landing sites. That made for complicated recovery owing to the rough terrain. However, a cache of spare parts for the Sirius was kept in the car used to drive between the various key points of the mapping mission. This included hull sections, motors and propellers. That would prove important after one landing resulted in a broken propeller. The prop was removed and replaced within five minutes after recovery, and the remaining flights and landings proceeded without repairs being needed. After each flight concluded with Fechtner completing the landing of the Sirius, the UAV would be recovered and Karas would check the integrity of the captured data output using a laptop powered by the car battery. By the end of the survey, 10 sq km had been mapped, with over 50 Gbytes of data in the form of roughly 7000 photographs captured by UpVision’s Sirius. Post-mission processing Downloads of the UAV’s data encompassed aerial images and flight logs. Each photo was oriented and aligned by MAVinci Desktop according to its centre coordinates, and previewed to check that the required overlays were met and that the mission area was covered to specification after each flight. A photogrammetric calculation of the entire surface of the mine was then performed on the UpVision computer server using AgiSoft Photoscan Pro. “The main output of this was the digital surface model [DSM] of the whole mine area, which formed a textured point cloud with a density of 400 million points,” Karas explains. Geologists primarily use the DSM model in specialist geology software applications, where they can analyse the slopes of the vertical walls, their stability and where potential problems in the mine might occur in the future. “These outputs enable the generation of a current 3D model of the mine, and for various analyses,” Karas says. “These can involve creating cross-sectional and longitudinal profiles in the mine, and could be used as a basis for planning future mining operations. “The current orthophoto of the whole area was also created as a background and raster layer for digitally drawing the geological and geodetic works.” Based on the results of this first Sirius flight over the mine, future measurements are to be conducted to help create a digital archive of the DSMs and data series to record and predict changes in the mine over time. October/November 2018 | Unmanned Systems Technology The best weather conditions for aerial mapping in Mongolia occur between May and July because of the low rainfall In operation | UpVision