It's important for the safety of air traffic that there are no high obstacles in the vicinity of the airport. However, the environment around an airport is constantly changing. How can you accurately, quickly and efficiently map these obstacles? Drones (Unmanned Aerial Vehicles of UAVs) are the answer.
The ICAO (International Civil Aviation Organization) lays down criteria in its guidelines for airports regarding height restrictions around airports. Every 5 years the airport must identify all natural and artificial objects that cut through the so-called obstacle-limiting surface. This obstacle-limiting plane indicates the height at which no structures may protrude above it, in order to keep the airspace free of obstacles.
Commissioned by the National Geographical Institute (NGI) and the Ministry of Defence, BitsOfData has detected obstacles in the vicinity of the Koksijde airbase and compared them with the restrictions imposed by the obstacle-limiting plane.
Traditionally, this obstacle detection is performed by means of terrain measurements or the interpretation of traditional aerial photographs. Both methods are very labour intensive and the measurement and processing takes several months. Data collection by UAVs is considerably cheaper, moreover, data collected by UAVs has a very short processing chain which reduces the time between flight and result delivery to a few weeks. The entire airbase including adjacent sites (in total 7.7 km²) was flown over by a multicopter equipped with a 24 MP camera. There were more than 4000 images taken in total with an image resolution of approximately 3 cm. The processing of the raw data takes several steps: relative orientation of images (based on GPS and point matches), georeferencing: integration of measured control points, creation of dense point cloud based on point-matching, derivation of height model (digital surface model - DSM) based on dense point cloud and creation (true) orthomosaic. The detection of the obstacles is done by comparing the points in the dense 3D point cloud with the height given in the obstacle limiting plane. The selected points are then interpreted visually and divided into classes (building, lighting mast, pylon or vegetation).
The image processing results are: an orthomosaic (resolution 3-5 cm), dense 3D point cloud (approx. 125 points/m²), accurate height model (DSM) and classification of obstacles (points). We can accurately identify the obstacles around the airport. Our results show that the use of drones is no longer limited to very small areas (in the order of a few hectares), but is also a cost-effective solution for larger areas. The image processing is relatively cheap, can be done quickly and has a high accuracy.