Lightweight SAR Image Ship Target Detection Algorithm Based on Improved YOLOv8

To address the issue of high model complexity in ship target detection algorithms for Synthetic Aperture Radar (SAR) images, which makes deployment on embedded devices difficult, a lightweight ship target detection algorithm based on YOLOv8 is proposed. First, a feature aggregation module is proposed to reconstruct the neck network. Second, deformable convolution and spatial channel reconstruction modules are introduced into the C2f module of the backbone. Finally, a lightweight shared convolution detection head is designed. Experimental results show that the improved algorithm reduces the number of parameters by 23.5% and the computational load by 23% compared to the original YOLOv8 algorithm, while achieving increases in mAP50 of 1.4% and 0.6% on the HRSID and SSDD datasets, respectively, which achieves the balance between algorithmic complexity and detection accuracy.