Ship Detection in SAR Imagery Based on Density and Clustering

In this paper, an algorithm for ship detection based on density and clustering is proposed for SAR imagery. The algorithm includes density map generation, local extremums extraction, target quasi-centers classification, suspected area identification, and detection results determination based on the quasi-centers and the suspected area. The proposed algorithm can effectively reduce the impact of sea clutter, outspread shadows and crossed sidelobes in SAR imagery, and provide more accurate detection results. Finally, the effecttiveness of the algorithm is verified by simulation.