Skip to content
Open access

Multi-scale ship detection with high-resolution SAR images based on improved YOLOv8

Oct 2026 · Frontiers in Marine Science · Vol 13 · 0 citations · 53 references

TL;DR

An improved YOLOv8 detection method for SAR ship detection tasks is proposed, including CA attention module, improved DSConv convolution module, and Shape-IoU loss function, which achieves superior detection performance under conventional IoU thresholds.

Abstract

To improve the detection accuracy of SAR ship targets under complex background clutter, this paper proposes an improved YOLOv8 detection method for SAR ship detection tasks. Three optimization strategies are adopted, including CA attention module, improved DSConv convolution module, and Shape-IoU loss function. A series of ablation experiments are performed to verify the effectiveness of different embedding positions, independent modules, and multi-module collaborative optimization. Quantitative results show that the proposed method achieves significant improvements in precision, recall and mAP@0.5 compared with the original YOLOv8 baseline. The introduced CA module enhances multi-scale feature representations of ship targets while attenuating responses from irrelevant background regions. The DSConv module extracts fine-grained ship features to reduce missed and false detections. The Shape-IoU loss function introduces shape-aware regularization to optimize bounding-box shape matching for ship targets. It improves multiple metrics when used alone. However, when jointly optimized with CA and DSConv, the complete model achieves prominent gains in precision, recall and mAP@0.5 at the cost of degraded mAP@0.5:0.95 due to multi-module coupling effects. Comparative experiments with multiple mainstream lightweight models demonstrate that the proposed method achieves superior detection performance under conventional IoU thresholds. Visualization results on four polarization modes further verify the strong anti-clutter capability and stable feature extraction ability of the improved model. Cross-dataset generalization tests show that the model trained on the SAR-SHIP (2021) dataset maintains stable accuracy gains on HRSID and SSDD datasets, proving excellent cross-domain generalization ability. Nevertheless, the multi-module optimization brings a certain performance trade-off. The proposed method sacrifices partial fine localization accuracy and inference speed to obtain higher conventional detection precision. It is suitable for high-precision SAR ship detection scenarios with moderate real-time requirements.

Read PDF

Similar papers

Open access Sep 2026

LHR-YOLO: A SAR small-target ship detection method based on improved YOLO11

Synthetic Aperture Radar (SAR) ship target detection holds significant application value in maritime traffic monitoring and marine environmental monitoring. However, due to challenges such as small ship targets, complex marine backgrounds, and speckle noise interference, existing methods still suffer from insufficient...

Miao Zhou, De-Zhi Han, Xiang Shen et al. · 0 citations
Conference Oct 2026

CAA-YOLO: SAR image detection method based on improved YOLO11

SAR image ship detection faces issues such as speckle noise, weak small target features, and arbitrary target rotation. This paper optimizes the algorithm based on YOLO11 by integrating three types of modules. The lightweight Context module expands the receptive field, enhances small target features, and suppresses noi...

Fei Lei, Xin Zhou · 0 citations
Oct 2026

Regional-Scale Object Feature Recognition in High-Resolution Remote Sensing Images Based on YOLOv8

Accurate object detection in regional-scale remote sensing imagery is challenged by significant scale variation, dense small targets, and complex background interference. While mainstream detection frameworks have achieved competitive results, their performance often declines when directly applied to high-resolution ae...

Yi-Ran Zhao, Rosniza Aznie Che Rose, Kuok Choy Lam · 0 citations
Conference Oct 2026

Ship object detection in remote sensing images based on improved real-time detection transformer

An improved algorithm based on RT-DETR is proposed, which achieves a 4.7% improvement in mAP50 on the DOTA dataset, with its computational cost reduced to 40.2 GFLOPs, thereby significantly enhancing both detection efficiency and accuracy.

Rui Zhang, Hao Rao · 0 citations
Open access Oct 2026

Mamba-enhanced RT-DETR for ship detection in optical remote sensing images

The proposed Mamba-enhanced Real-Time Detection Transformer (RT-DETR) framework for high-precision ship detection and identification achieves strong robustness and efficiency, showing high practical value for maritime monitoring, port security, and emergency response.

Qi-Chang Zhao, Dan-Dan Zhi, Yi-Quan Wu · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.