Subaperture Decomposition for Ship Detection and Velocity Estimation from SLC SAR data

A spaceborne synthetic aperture radar (SAR) systematically scans the target scene along its trajectory at various positions and azimuth angles, making it a unique asset for many remote sensing applications. The contrast between ship targets and the surrounding sea can be improved in SAR images by splitting the bandwidth into subapertures (SAs) and then averaging to attain a high azimuth resolution. This letter proposes the use of subaperture decomposition (SD) for detecting moving ships. Furthermore, the ship velocity parameter in the azimuth direction is computed. The proposed SAs dual-pol descriptor (rSA) employed the constant false alarm rate (CFAR) approach followed by a statistical Gamma (Γ) distribution to determine the threshold value for distinguishing ship and sea. Thereafter, the ship's edges are quantified using Sobel and Canny edge detectors. The eigenvalue descriptor (λ) of polarimetric SAR provides the total size of the target; hence, the ship sizes are compared with λ. The velocity of moving ships (VS) is evaluated by utilizing the mean difference of the Doppler centroid (ΔfDC) information between pairs of all SAs and is validated using the SA displacement correlation method. The experimental results demonstrate that the proposed method is feasible for real-time moving ship monitoring with 95% accuracy.