Feature based extraction method for airplane door detection and localization to control the aerobridge

Abstract

In this thesis, a features detection system of aircraft door in the automated passengers boarding bridge, also called as aerobridge docking system, is proposed. The target for the automated passenger boarding bridge system is to dock the aerobridge to the airplane after the aircraft landing. The door features detecting is serve as the eye for the aerobridge docking system, which is able to locate the position of the aircraft door, and then to guide the aerobridge docking to the door automatically. This thesis focus on the first level for airplane door detection, it introduces 5 common features in the airplane door and discusses the method to find these features.

At first, some critical technics to pre-process the images, in order to simplify the information, has been introduced, such as smoothing technics and information extracting technics. After that, arrow, door handle, footplate, doorframe, and window have been set as the common features for aircraft door, and then, the different ways to locate these features, based on the physical characteristics of them, have been proposed one by one, and the result of detecting has been presented. Meanwhile, some special situations for features detecting are also discussed in this thesis, such as poor light condition and heavy rain.