Computer Vision System for Cabin Door Detection and Location

This paper describes the outcomes of a framework for aircraft door position identification with visual camera using common features that can be found in all commercial airplanes, which guides the aerobridge to automatically and reliably dock (Precision of 0.5cm), with minimal human intervention. The system has to work day and night, rain or shine under all-weather conditions. The proposed solution consists of a suite of relevant extracted and identified features that characterize aircraft door (e.g. door windows, handle, text, footplate, arrow, frame lines). Furthermore, the work focuses on a final door position confirmation accomplished by the logical structure for the decision process for the identification of the aircraft door using the various features. In that way we are able to build confidence level with appropriate structural (logical component connection) organization. Thus, the aircraft door frame can be accurately identified with a very small error (in the range of a couple of cm) and eliminate risk of injuring the personnel or damaging the airplane.