In this study, we address the challenges associated with Video Synthetic Aperture Radar (Video SAR) shadow tracking, a technique used for continuous monitoring of ground moving targets. Due to challenges such as changes in shadow appearance, low contrast between shadow and background, and scene occlusion in Video SAR, existing methods often encounter extensive matching errors in the data association process, resulting in unsatisfactory tracking performance. To overcome these issues, we propose a novel method, GNN-JFL, which is based on joint motion-appearance feature extraction and graph neural data association. This method uses the detector as a flexible plugin and introduces two key improvements in the tracker section to enhance tracking accuracy. Firstly, we introduce joint feature learning to extract the complementary appearance and motion features from shadow shapes and positions, obtaining more robust feature representations to improve tracking performance under intricate challenges. Secondly, by organically integrating Multi-object Tracking (MOT) problems and Graph Neural Networks (GNN), we propose a novel GNN-based shadow tracking architecture, which utilizes graph relationships to learn the associations between shadows for more accurate tracking predictions. Our method is validated using two measured datasets and demonstrate superior performance in terms of multi-object tracking accuracy (MOTA). It outperforms the suboptimal method by 4.2% and 3.6% in the two datasets, respectively. This research contributes to the advancement of continuous monitoring techniques employing Video SAR shadow tracking.

At present, most of the ship detection methods use horizontal detection bounding box, which results in the interference of dense ships and decreases the detection accuracy. In addition, most of the current ship detection methods remain in single-category detection, and do not achieve multi-class ship detection, which limits the further promotion and application of these detection methods. In this work, we propose a novel network for high-precision detection of rotated multi-class ships by rotated bounding box, called Rotated Multi-Class Detection Network (RMCD-Net). In RMCD-Net, we adapt a rotated anchor-feature alignment module (RAAM) to solve the misalignment problem between rotated anchors and horizontal features. In RMCD-Net, we adapt double detection head mechanism for better regression and classification. Also, we apply focal loss to classification task. Experimental results on the public dataset SRSDD show that mAP of RMCD-Net is 61.62% that is better than the second-best model by 5.39%.

We report a saliency-guided attention-based feature pyramid network (SA-FPN) for ship detection from synthetic aperture radar (SAR) images. The two key contributions are – 1) the saliency-guided technique and 2) the attention-based means. The former offers one unsupervised visual saliency map that can guide FPN to focus more on regions of interest (ROIs). The latter offers one supervised non-local feature self-attention map that can improve FPN’s global representation ability. We offer an effective combination scheme of the two. Experimental results on the open SSDD dataset reveal SA-FPN’s advanced SAR ship detection performance. Furthermore, the ablation studies can confirm the two contributions' effectiveness.

Ship detection from Synthetic Aperture Radar (SAR) images is of great importance. However, the diversity of ship target scales increases the difficulty of detection. To solve this problem, we propose a novel FPN which is enhanced by de-formable convo-lution, called Deform-FPN. Deformable convolution realizes multi-scale adaptive geometric deformation modeling of ships, and can extract multi-scale features of ships with strong ex-pression ability. The multi-level deformable convolution layers enhance the feature extraction and feature fusion capabilities. Specifically, we add deformable convolution to the backbone and lateral connection of Deform-FPN to improve the feature extraction ability. Experimental results on the SAR ship detec-tion dataset (SSDD) reveal the state-of-the-art performance of Deform-FPN, in contrast to other methods based on convolu-tional neural network (CNN). The experimental results show that Deform-FPN offers a 56.5% mAP that is superior to the suboptimal model DCN by 1.5%. In addition, we conducted ablation experiments to verify the effectiveness of the structure of the Deform-FPN we proposed.

Researchers can gain more intuitive information by converting synthetic aperture radar (SAR) images to optical images using generative adversarial networks (GANs). However, their GANs have poor feature extraction ability, which leads to color conversion errors and loss of details. Therefore, to solve this problem, we add an atrous spatial pyramid pooling (ASPP) module to GAN to enhance the feature extraction ability, i.e., ASPP-GAN. ASPP module can extract multi-resolution feature responses, enabling the network to focus on both overall and detailed features for better feature extraction ability. The experimental results on public SEN1-2 datasets show that ASPP-GAN has a significant improvement over the traditional GAN, i.e., Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM) values are improved by about 20%.