Detecting and recognizing runway holding position
signs and hold-short lines are of significant safety interest for
runway incursion avoidance. It is an important supplement to
pilots’ education and airport ground radar to enhance safety,
particularly at complex major airports. In this project, we leverage the standards of airport signs and markings for the detection
and recognition purposes. For sign detection and recognition, an
HSV color space based algorithm is developed to detect runway
holding positions signs and Optical Character Recognition (OCR)
is used to recognized the signs. For line detection, Canny Edge
Detection and Hough Transform are used to recognize major
line features in the images, and a heuristic algorithm leveraging
the layout characteristics of the images is used to determine the
presence of hold-short lines among those lines. An accuracy of
95.1% is achieved for runway holding position signs detection and
recognition while an accuracy of 89.2% is achieved for runway
hold-short line detection.