Multiple Luminaire Identification in Airborne Images of Airport’s Approach Lighting Using Mathematical Morphology With Variable Length Structuring Element

摘要

The increasing demand for fast air transportation around the clock has increased the number of night flights in civil aviation over the past few decades. In night aviation, to land an aircraft, a pilot needs to be able to identify an airport. The approach lighting system (ALS) at an airport is used to provide identification and guidance to pilots from a distance. ALS consists of more than $100$ luminaires which are installed in a defined pattern following strict guidelines by the International Civil Aviation Organization (ICAO). ICAO also has strict regulations for maintaining the performance level of the luminaires. However, once installed, to date there is no automated technique by which to monitor the performance of the lighting. We suggest using images of the lighting pattern captured using a camera placed inside an aircraft. Based on the information contained within these images, the performance of the luminaires has to be evaluated which requires identification of over $100$ luminaires within the pattern of ALS image. This research proposes analysis of the pattern using morphology filters which use a variable length structuring element (VLSE). The dimension of the VLSE changes continuously within an image and varies for different images. A novel technique for automatic determination of the VLSE is proposed and it allows successful identification of the luminaires from the image data as verified through the use of simulated and real data.