Airport runway detection based on long linear structure

To meet the needs of precise runway detection in remote sensing images, a method was proposed to extract the runway from airport based on its long linear structure. The potential center line was detected from remote sensing images, since the runway had high contrast and topological straight edge to the background. Using the points on the center line, latent runway-like edges can be found by extending pixels in the direction perpendicular to the line. Removing regions with small area, parallel lines can be identified by use of Hough transformation under certain constraint. As the standard for building airport was concerned, the threshold markings were high gray-level and broad enough, so the latent regions in the direction perpendicular to the parallel lines were projected and the peak to define the length of runway was searched. The results of concrete surface and pitch surface airport runway show that the method can remove complex background effectively and has high precision.