Airport detection based on Line Segment Detector

In this paper we propose a new approach for detecting airport on remote sensing images based on Line Segment Detector (LSD). First, LSD is applied on the image to extract straight line segments, and then fragmented line segments generated by LSD are joined to reduce false alarm rate (FAR). As we know that airport regions are more likely to have parallel lines, we accumulate the total length of parallel lines that are close enough, and select the region which has the longest total length of parallel lines as an airport region. This algorithm has been tested on remote sensing images and compared with other methods, the result shows that the approach presented here generates a considerably high detection rate and low false alarm rate in a reasonable time.