JDet-based SAR Images Multi-directional Maritime Ship Target Detection

In complex remote sensing image detection, there are still many challenges in maritime ship detection. In this paper, we combine S2anet, Faster R-CNN, RoI Transformer, FCOS and RetinaNet algorithms from the latest JDet remote sensing image detection algorithm library with satellite remote sensing big data, and use the only SAR image dataset with rotated border annotation, SSDD+, to study the target detection of maritime ships. The experimental results show that the various algorithms in the JDet algorithm library are suitable for remote sensing maritime ship target detection, especially for small target detection, among which S2anet has the most superior results with a mAP of 89.8% In the results, this framework can achieve high accuracy in remote sensing image detection. In addition, we verify the effectiveness, stability and In addition, we verify the effectiveness, stability and accuracy of the above algorithms for rotating border target detection in complex remote sensing scenarios.