Research On The Technology Of Airport Surface Vehicle Terminal Positioning Based On Signboard Recognition

With the increasing amount of airport ground service vehicles, the airport operating environmentbecomes more and more complex. Besides, the running vehicles have become an important researchobject. The key to accomplish airport surface surveillance system is to get accurate position ofrunning vehicles on the airport surface. On this issue, this paper installed a visible light camera on amoving vehicle and combining it with the laser range finder to detect and identify the signboardbeside the taxiway. In addition, this thesis fused the information with the real-time data provided bythe GPS to get more precise vehicles location information than traditional GPS so as to meet theairport positioning requirements.In this thesis, the research content mainly included two parts: signboard recognition, positioninginformation fusion. In the first part, this paper preprocessed and transformed the color modelaccording to the features of taxiway signboard, pre-positioned and accurately positioned the candidateregion, segmented the characters by the method of connected domain, so as to prepare for featureextraction of the next step.Secondly, this paper extracted feature invariant according to the deformation characteristics ofthe image by using the feature extraction of affine invariant moment, and verified the affine invariantand robustness of the extracted characteristics. Experiment showed that the feature extraction methodwas effective.Finally, this paper used support vector machine (SVM) to design classifier, and simulated themoving vehicles running in the airport taxiway to collect some groups of pictures in different naturalconditions as sample library. An off-line identification experiment was done to verify the effectivenessof the algorithm. The result showed that the method can identify the character effectively. It met theaccuracy requirement well.In the second part, this thesis called the geographic information database of the airport to extractthe coordinates information of the taxiway signboard, and got the current location information of thevehicle based on the distance and azimuth angle between present vehicle and the signboard detectedby the laser range finder. Meanwhile, the paper fused the location information with which tested byGPS to get the accurate positioning information. In addition, this paper simulated the moving vehiclesrunning in the airport taxiway to verify the function of the system. The result showed that theprecision tested by the method of information fusion was more accurate than using traditional GPS positioning. The reliability of the algorithm was verified.