Maritime Target Detection Method Based on SDP-YOLOv8

Due to the influence of sea clutter, detecting maritime targets, especially small ones, is extremely challenging. With the continuous development of image recognition technology, techniques that convert one-dimensional maritime radar echo data into two-dimensional images and use neural networks for target recognition have been proposed. This paper applies this technology by converting radar echo data containing targets and pure sea clutter into two-dimensional images using the Symmetric Data Path (SDP) method. The YOLOv8 algorithm is then employed for classification to determine the presence of targets. The effectiveness of the proposed method is validated using a dataset shared by the Naval Aviation University.