Backprojection Autofocus of Moving Ships for Synthetic Aperture Radar

In this paper, we present a backprojection autofocus method to image moving ships in high resolution very precisely. This technique might make an automatic ship classification based on Synthetic Aperture Radar (SAR) images without using additional information from the Automated Identification System (AIS) possible. We use a Constant False Alarm Rate (CFAR) detector to estimate the coarse position of a ship in a preview SAR image. A second high resolution image of the ship is generated by the proposed backprojection autofocus algorithm. Our autofocus estimates the time depending distance between the sensor and every pixel in the SAR image, which includes the rotation and the translation of the ship. This estimated relative distance is used to compute a high quality autofocused SAR image of the considered moving ship.