Synthetic Aperture Radar Images Research Articles

A Structurally Flexible Occupancy Network for 3-D Target Reconstruction Using 2-D SAR Images

A Structurally Flexible Occupancy Network for 3-D Target Reconstruction Using 2-D SAR Images

Inexpensive SAR testbed for 3D mmWave imaging applications

AbstractIn this paper, the authors investigate and develop a low‐cost synthetic aperture radar (SAR) testbed and utilise it to capture and reconstruct high‐resolution 3D mmWave SAR images. Despite much attention from both industrial and academic scholars, cost restraints and complexity have hindered the deployment of SAR systems, particularly in the mmWave domain. This paper outlines the major components and systems of the 3D SAR testbed. The authors build a testbed with low‐cost, commercially available hardware components and include open‐source software that is highly reconfigurable and simple to reproduce. The multiple‐input multiple‐output (MIMO) radar sensor uses a frequency‐modulated continuous wave (FMCW) chirp at the V‐band (40–75 GHz) and synthesises a rectilinear two‐dimensional planar aperture. The authors presented several reconstructed 3D images imitating real‐world scenarios using the testbed. A non‐linear sampling technique that has significantly decreased the amount of time required for data acquisition was implemented. In addition, the authors employed multiple image quality metrics to quantify and compare the images produced by the testbed. The testbed can be used to test and validate new techniques and algorithms.

Frequent Glacial Hazard Deformation Detection Based on POT-SBAS InSAR in the Sedongpu Basin in the Himalayan Region

The Sedongpu Basin is characterized by frequent glacial debris movements and glacial hazards. To accurately monitor and research these glacier hazards, Sentinel-1 Synthetic Aperture Radar images observed between 2014 and 2022 were collected to extract surface motion using SBAS-POT technology. The acquired temporal surface deformation and multiple optical remote sensing images were then jointly used to analyze the characteristics of the long-term glacier movement in the Sedongpu Basin. Furthermore, historical meteorological and seismic data were collected to analyze the mechanisms of multiple ice avalanche chain hazards. It was found that abnormal deformation signals of glaciers SDP1 and SDP2 could be linked to the historical ice avalanche disaster that occurred around the Sedongpu Basin. The maximum deformation rate of SDP1 was 74 m/a and the slope cumulative deformation exceeded 500 m during the monitoring period from 2014 to 2022, which is still in active motion at present; for SDP2, a cumulative deformation of more than 300 m was also detected over the monitoring period. Glaciers SDP3, SDP4, and SDP5 have been relatively stable until now; however, ice cracks are well developed in SDP4 and SDP5, and ice avalanche events may occur if these ice cracks continue to expand under extreme natural conditions in the future. Therefore, this paper emphasizes the seriousness of the ice avalanche event in Sedongpu Basin and provides data support for local disaster management and disaster prevention and reduction.

An Imaging Method for Marine Targets in Corner Reflector Jamming Scenario Based on Time–Frequency Analysis and Modified Clean Technique

In the corner reflector jamming scenario, the ship target and the corner reflector array have different degrees of defocusing in the synthetic aperture radar (SAR) image due to their complex motions, which is unfavorable to the subsequent target recognition. In this manuscript, we propose an imaging method for marine targets based on time–frequency analysis with the modified Clean technique. Firstly, the motion models of the ship target and the corner reflector array are established, and the characteristics of their Doppler parameter distribution are analyzed. Then, the Chirp Rate–Quadratic Chirp Rate Distribution (CR-QCRD) algorithm is utilized to estimate the Doppler parameters. To address the challenges posed by the aggregated scattering points of the ship target and the overlapping Doppler histories of the corner reflector array, the Clean technique is modified by short-time Fourier transform (STFT) filtering and amplitude–phase distortion correction using fractional Fourier transform (FrFT) filtering. This modification aims to improve the accuracy and efficiency of extracting scattering point components. Thirdly, in response to the poor universality of the traditional Clean iterative termination condition, the kurtosis of the residual signal spectrum amplitude is adopted as the new iterative termination condition. Compared with the existing imaging methods, the proposed method can adapt to the different Doppler distribution characteristics of the ship target and the corner reflector array, thus realizing better robustness in obtaining a well-focused target image. Finally, simulation experiments verify the effectiveness of the algorithm.

SVDDD: SAR Vehicle Target Detection Dataset Augmentation Based on Diffusion Model

In the field of target detection using synthetic aperture radar (SAR) images, deep learning-based supervised learning methods have demonstrated outstanding performance. However, the effectiveness of deep learning methods is largely influenced by the quantity and diversity of samples in the dataset. Unfortunately, due to various constraints, the availability of labeled image data for training SAR vehicle detection networks is quite limited. This scarcity of data has become one of the main obstacles hindering the further development of SAR vehicle detection. In response to this issue, this paper collects SAR images of the Ka, Ku, and X bands to construct a labeled dataset for training Stable Diffusion and then propose a framework for data augmentation for SAR vehicle detection based on the Diffusion model, which consists of a fine-tuned Stable Diffusion model, a ControlNet, and a series of methods for processing and filtering images based on image clarity, histogram, and an influence function to enhance the diversity of the original dataset, thereby improving the performance of deep learning detection models. In the experiment, the samples we generated and screened achieved an average improvement of 2.32%, with a maximum of 6.6% in mAP75 on five different strong baseline detectors.

A Spatio-Temporal Deep Learning Model for Automatic Arctic Sea Ice Classification with Sentinel-1 SAR Imagery

Arctic sea ice has a significant effect on global climate change, ship navigation, Arctic ecosystems, and human activities. Therefore, it is essential to produce high-resolution sea ice maps that accurately represent the geographical distribution of various sea ice types. Based on deep learning technology, many automatic sea ice classification algorithms have been developed using synthetic aperture radar (SAR) imagery over the last decade. However, sea ice classification faces two vital challenges: (1) it is difficult to distinguish sea ice types with close developmental stages solely from SAR images and (2) an imbalanced sea ice dataset has a significantly negative effect on ice classification model performance. In this article, a spatio-temporal deep learning model—the Dynamic Multi-Layer Perceptron (MLP)—is utilized to classify 10 sea ice types automatically. It consists of a SAR image branch and a spatio-temporal branch, which extracts SAR image features and spatio-temporal distribution characteristics of sea ice, respectively. By projecting similar image features to different positions in the spatio-temporal feature space dynamically, the Dynamic MLP model effectively distinguishes between similar sea ice types. Furthermore, to reduce the impact of data imbalance on model performance, the dynamic curriculum learning (DCL) method is used to train the Dynamic MLP model. Experimental results demonstrate that our proposed method outperforms the long short-term memory (LSTM) network approach in distinguishing between sea ice types with similar developmental stages. Moreover, the DCL training method can also effectively improve model performance in identifying minority ice types.

Semi-supervised learning based on distance measurement and augmentations comparison for SAR image recognition

ABSTRACT In recent years, deep learning has significantly improved the performance of Synthetic Aperture Radar Automatic Target Recognition (SAR-ATR), but the lack of large-scale labelled training data limits the further development of deep neural networks in this field. Traditionally, annotating SAR data requires manual labour by experts with specialized domain knowledge, which is costly and makes it difficult to obtain a large amount of labelled data. Semi-supervised learning (SSL) algorithms can effectively leverage unlabelled data for training, improving model performance. However, improper use of pseudo-labels can lead to error accumulation and result in performance degradation. This paper proposes an SSL algorithm for SAR image recognition that employs the following strategies to mitigate error accumulation and enhance performance: First, by combining pseudo-labelling and consistency-regularization methods, unlabelled data with different augmentations are compared in terms of classification predictions and feature vectors, improving the model representation capacity. Second, a weighted distance metric combining labelled and unlabelled data is designed to estimate model loss, guiding the model towards intra-class aggregation and inter-class dispersion in terms of data distribution. Experiments conducted on the MSTAR and OpenSARShip datasets demonstrate that the algorithm achieves superior performance in limited labelled sample scenarios, with an accuracy of over 99% when only 30 labelled data points (approximately 10%) per class are available, and an accuracy of over 85% with extremely scarce samples (only 5 labelled samples per class) on the MSTAR dataset.

Dual-Modal Approach for Ship Detection: Fusing Synthetic Aperture Radar and Optical Satellite Imagery.

The fusion of synthetic aperture radar (SAR) and optical satellite imagery poses significant challenges for ship detection due to the distinct characteristics and noise profiles of each modality. Optical imagery provides high-resolution information but struggles in adverse weather and low-light conditions, reducing its reliability for maritime applications. In contrast, SAR imagery excels in these scenarios but is prone to noise and clutter, complicating vessel detection. Existing research on SAR and optical image fusion often fails to effectively leverage their complementary strengths, resulting in suboptimal detection outcomes. This research presents a novel fusion framework designed to enhance ship detection by integrating SAR and optical imagery. This framework incorporates a detection system for optical images that utilizes Contrast Limited Adaptive Histogram Equalization (CLAHE) in combination with the YOLOv7 model to improve accuracy and processing speed. For SAR images, a customized Detection Transformer model, SAR-EDT, integrates advanced denoising algorithms and optimized pooling configurations. A fusion module evaluates the overlaps of detected bounding boxes based on intersection over union (IoU) metrics. Fused detections are generated by averaging confidence scores and recalculating bounding box dimensions, followed by robust postprocessing to eliminate duplicates. The proposed framework significantly improves ship detection accuracy across various scenarios.

Tropical Rice Mapping Using Time-Series SAR Images and ESF-Seg Model in Hainan, China, from 2019 to 2023

Tropical Rice Mapping Using Time-Series SAR Images and ESF-Seg Model in Hainan, China, from 2019 to 2023

Ship target detection method based on improved YOLOv8 for SAR images

ABSTRACT Due to the complex background and less effective information in low-resolution and noisy images, the detection of small ships in SAR images suffers from low detection rate and high false alarm rate. In order to solve the above problems, this paper proposes a method to detect the small ship targets in SAR images based on improved YOLOv8. Firstly, the deformable convolutional networks are added to the leading network to improve feature extraction. Then, the efficient multi-scale attention module is fused with the backbone network to enhance the detection effect of small targets. The weighted intersection over union loss function is used to optimize the regression process of the prediction frame and the detection frame to enhance the localization capacity of small targets. Finally, there is an addition of a specialized small target detection layer, and a reconstruction of the feature extraction and fusion network to enhance the detection performance of small ship targets in SAR images. To verify the effectiveness and robustness of the method, we conduct experiments on SSDD and HRSID. The proposed method achieves high detection accuracy while also offering a more compact model size and less computation time compared to other prevalent methods.

Confidence-Aware Ship Classification Using Contour Features in SAR Images

In this paper, a novel set of 13 handcrafted features derived from the contours of ships in synthetic aperture radar (SAR) images is introduced for ship classification. Additionally, the information entropy is presented as a valuable metric for quantifying the confidence (or uncertainty) associated with classification predictions. Two segmentation methods for the contour extraction were investigated: a classical approach using the watershed algorithm and a U-Net architecture. The features were tested using a support vector machine (SVM) on the OpenSARShip and FUSAR-Ship datasets, demonstrating improved results compared to existing handcrafted features in the literature. Alongside the SVM, a random forest (RF) and a Gaussian process classifier (GPC) were used to examine the effect of entropy derivation from different classifiers while assessing feature robustness. The results show that when aggregating predictions of an ensemble, techniques such as entropy-weighted averaging are shown to produce higher accuracies than methods like majority voting. It is also found that the aggregation of individual entropies within an ensemble leads to a normal distribution, effectively minimizing outliers. This characteristic was utilized to model the entropy distributions, from which confidence levels were established based on Gaussian parameters. Predictions were then assigned to one of three confidence levels (high, moderate, or low), with the Gaussian-based approach showing superior correlation with classification accuracy compared to other methods.

An Integrated Segmentation Framework Utilizing Mamba Fusion for Optical and SAR Images

ABSTRACTSemantic segmentation of multimodal remote sensing images is an effective approach to enhancing segmentation accuracy. Optical and synthetic aperture radar (SAR) images capture ground features from distinct perspectives, offering varied information for ground observation. Effectively fusing these two modalities information and performing multimodal segmentation remains a promising yet challenging task due to their complementary nature and significant differences. The existing methods have the problem of ignoring spatial dimension information and being unable to bridge semantic gaps. However, the proposal of Selective Structured State Space Models (Mamba) provides opportunities for multimodal fusion. Therefore, we propose a segmentation framework based on Mamba fusion of optical and SAR images. The framework introduces a novel fusion module inspired by the principle of Mamba. This module selects effective features from different modalities for cross‐fusion within a global sensing field, facilitating the mutual compensation of optical and SAR image features and reducing the semantic gap. The fused features are then accurately segmented using a decoder that incorporates the Atrous Spatial Pyramid Pooling (ASPP) technique. On the WHU‐OPT‐SAR dataset, this method outperforms other state‐of‐the‐art deep learning approaches, achieving an overall accuracy (OA) of 84.13%, compared to 76.29% for the Kappa statistic.

Enhancing Adversarial Transferability With Intermediate Layer Feature Attack on Synthetic Aperture Radar Images

Enhancing Adversarial Transferability With Intermediate Layer Feature Attack on Synthetic Aperture Radar Images

Synthetic Aperture Radar Imaging Using Computer Simulation of Quantum Algorithm and Circuits

Synthetic Aperture Radar Imaging Using Computer Simulation of Quantum Algorithm and Circuits

Lambda-1 Detector: Adaptive Interference Detection in Synthetic Aperture Radar Images

Lambda-1 Detector: Adaptive Interference Detection in Synthetic Aperture Radar Images

Manifold Learning and Deep Generative Networks for Heterogeneous Change Detection From Hyperspectral and Synthetic Aperture Radar Images

Manifold Learning and Deep Generative Networks for Heterogeneous Change Detection From Hyperspectral and Synthetic Aperture Radar Images

A Two-Stage Oil Spill Detection Method Based on an Improved Superpixel Module and DeepLab V3+ Using SAR Images

A Two-Stage Oil Spill Detection Method Based on an Improved Superpixel Module and DeepLab V3+ Using SAR Images

RDB-DINO: An Improved End-to-End Transformer With Refined De-Noising and Boxes for Small-Scale Ship Detection in SAR Images

RDB-DINO: An Improved End-to-End Transformer With Refined De-Noising and Boxes for Small-Scale Ship Detection in SAR Images

Ricci-Flow Method for Ship Detection in SAR Image

Ricci-Flow Method for Ship Detection in SAR Image

Multifeature Alignment and Matching Network for SAR and Optical Image Registration

Multifeature Alignment and Matching Network for SAR and Optical Image Registration