A new ship detector for ScanSAR imagery

Due to its extensive and acceptable resolution, the use of scanning synthetic aperture radar (ScanSAR) image for ship detection has become a common application in ocean monitoring studies. In this paper, a new ScanSAR ship detector, which is based on the Hessian of Laplacian, is proposed. The radar echoes of ships and the antenna gain (AG) phenomenon of the ocean in ScanSAR image are investigated. To address the AG phenomenon and the ship-background contrast, comparison experiments of 3 measures of the Hessian matrix (the determinant, the eigenvalue and the trace) are conducted, and the determinant of the Hessian matrix in scale-space is selected as the detector. To improve the performance of the detector (i.e., infrequent false alarms and low undetected rate), a new 3D (x axis, y axis and scale axis) detector is developed with a protected distance. The proposed method is validated with experimental images, and several comparisons are made to prove its robustness and efficiency.