Runway detection using unsupervised classification

Recent advances in the field of remote sensing technology have opened new prospects and impelled the way of analyzing images from remote sensing satellites to detect or identify an object, or a place which is selected as area of interest. The detection of airport becomes a motivating topic recently because of its applications and importance in military and civil aviation fields. This paper presents an approach for airport detection using remote sensing images by implementing conventional K-means unsupervised classification and implementing unsupervised classification based on decomposed polarimetric features that includes Entropy (H), Anisotropy (A), and Alpha angle (α). The obtained preliminary results reveal that classification based on decomposed polarimetric features provided better results than the conventional unsupervised classification for the detection of target. In this work, the effectiveness of the algorithms was demonstrated using quadpolarimetric L-band Polarimetric Synthetic Aperture Radar (polSAR) imagery from the NASA Jet Propulsion Laboratory's (JPL's) Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR). The study area is Louis Armstrong New Orleans International Airport, LA, USA.