Sparse coding based airport detection from medium resolution Landsat-7 satellite remote sensing images

A simple but effective method for airport detection from medium resolution Landsat-7 satellite remote sensing images based on sparse coding is presented. It consists of three phases: dictionary construction, sparse coding, and airport detection. Firstly, an over-complete dictionary is constructed using a set of airport training samples. Secondly, test images are scanned using multi-scale windows and each scanned window is sparsely coded in terms of atoms of the dictionary. Finally, sparsity concentration index of each scanned window is calculated based on the coding coefficients, which is used to decide the airport detection. Evaluations on publically available satellite images and comparisons with state-of-the-art approaches have demonstrated the superiority of the presented work.