Aerial mapping and digitization of railway’s objects with geodetic accuracy based on UAV imagery

Abstract

<p>People are interested in the usability of drone recordings in many areas of research around the world. The use of this method for economic, financial, industrial and research purposes is being addressed in more and more areas around the world, and more and more results support the effectiveness of this method. In Hungary, the use of drone recordings has grown exponentially in the last 5-10 years, but there are still more sectors with significant research potential. One of the countless possibilities of its use is to carry out surveys in the environment of railway transport. It is still a relatively new research direction.</p><p>Numerous objects in the railway environment pose a challenge to the surveyor specialist. These field objects (e.g. truck, box, switches, lamps, and overhead wire) can negatively affect the measurement. Measuring on railways is dangerous, the large number of objects slows down the measurement, and the iron structure of bridges degrades the accuracy of GPS. Covering can also be an obstacle, e.g. ruined buildings, areas significantly overgrown by vegetation.</p><p>Nowadays, geodetic surveyors are using RTK (Real-Time Kinematic) GPS in the Hungarian State Railways Co. Ltd. These field surveys involve labour-intensive and time-consuming processes and are hazardous to human lives. They most often take place in an area with scheduled daily traffic. Current measurements with a drone are not affected by these disadvantages. The most important criteria are the accuracy of the measurement result and the resolution. We want to examine the accuracy of aerial photography compared with field measurements. Depending on the result, it is worth considering replacing the measurement technology and weighing between the abovementioned aspects.</p><p>We choose the Vác (Hungary) railway station as the examining area, which is quite important due to the high number and the diversity of the objects and the volume of the passenger traffic. The measurement was performed with a DJI Mavic 2 Pro type drone at a flight altitude of 50 m with an average flight speed of 4 m/s.</p><p>The expected results are encouraging in favour of aerial photography. The end product is an orthophoto, which captures all the objects that can be connected to the railway. A large number of objects could be digitized and registered in a database. This database is similar to the RTK in situ survey results, but the negatives mentioned above does not appear during the flight.</p><p>The orthomosaic made by lots of orthophotos (in this case 375 pieces) seems to be accurate enough. It is possible to determine the difference between the ground control points (GCPs) measured locations in the field and their coordinates on the orthophotos. The expected values of the margin errors could be low enough when we make variations of the GCP’s location. Based on our surveys so far, from the use of drone recordings, at least as much accuracy can be expected as the RTK surveying.</p><p> </p><p><em>FV is supported by EFOP-3.6.3-VEKOP-16-2017-00001: Talent Management in Autonomous Vehicle Control Technologies – The Project is financed by the Hungarian Government and co-financed by the European Social Fund.</em></p>