A CFAR-Enhanced Ship Detector for SAR Images Based on YOLOv5s

Xue Wen,Shaoming Zhang,Jianmei Wang,Tangjun Yao,Yan Tang

DOI: https://doi.org/10.3390/rs16050733

IF: 5

2024-02-20

Remote Sensing

Abstract:Ship detection and recognition in Synthetic Aperture Radar (SAR) images are crucial for maritime surveillance and traffic management. Limited availability of high-quality datasets hinders in-depth exploration of ship features in complex SAR images. While most existing SAR ship research is primarily based on Convolutional Neural Networks (CNNs), and although deep learning advances SAR image interpretation, it often prioritizes recognition over computational efficiency and underutilizes SAR image prior information. Therefore, this paper proposes YOLOv5s-based ship detection in SAR images. Firstly, for comprehensive detection enhancement, we employ the lightweight YOLOv5s model as the baseline. Secondly, we introduce a sub-net into YOLOv5s, learning traditional features to augment ship feature representation of Constant False Alarm Rate (CFAR). Additionally, we attempt to incorporate frequency-domain information into the channel attention mechanism to further improve detection. Extensive experiments on the Ship Recognition and Detection Dataset (SRSDDv1.0) in complex SAR scenarios confirm our method's 68.04% detection accuracy and 60.25% recall, with a compact 18.51 M model size. Our network surpasses peers in mAP, F1 score, model size, and inference speed, displaying robustness across diverse complex scenes.

environmental sciences,imaging science & photographic technology,remote sensing,geosciences, multidisciplinary

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve

This paper primarily aims to address the issue of ship detection in Synthetic Aperture Radar (SAR) images, specifically:

1. **Ship Detection in Complex Backgrounds**: Due to the complex backgrounds in SAR images, traditional methods face numerous challenges in detecting ships, especially in the presence of significant noise interference. This paper proposes a rotation-invariant ship detection method based on the YOLOv5s framework, aiming to improve detection performance in complex backgrounds.

2. **Combining Handcrafted Features with Deep Learning**: Traditional handcrafted features (such as Constant False Alarm Rate, CFAR) are reliable but perform poorly in complex scenarios; while deep learning-based methods are powerful, they lack interpretability and are sensitive to noise. This paper innovatively combines handcrafted features (by learning CFAR features through a subnet) with deep learning models to enhance the robustness and interpretability of the model.

3. **Incorporating Frequency Domain Information**: Most existing methods only utilize spatial domain information of SAR images, neglecting the importance of frequency domain information. This paper introduces frequency domain information into the channel attention mechanism, further improving detection accuracy and robustness.

Through the aforementioned improvements, the proposed model in this paper achieves significant performance enhancement on public SAR ship detection datasets, particularly excelling in detection accuracy and recall rate in complex scenarios. This indicates that the method has high practical value in real-world applications.