检索结果-内蒙古大学图书馆

SARATR-X: Toward Building a Foundation Model for SAR Target Recognition

IEEE TRANSACTIONS ON image processing 2025年 34卷 869-884页

作者： Li, Weijie Yang, Wei Hou, Yuenan Liu, Li Liu, Yongxiang Li, Xiang Natl Univ Def Technol Coll Elect Sci & Technol Changsha 410073 Peoples R China Shanghai AI Lab Shanghai 200000 Peoples R China

Despite the remarkable progress in synthetic aperture radar automatic target recognition (SAR ATR), recent efforts have concentrated on detecting and classifying a specific category, e.g., vehicles, ships, airplanes, or buildings. One of the fundamental limitations of the top-performing SAR ATR methods is that the learning paradigm is supervised, task-specific, limited-category, closed-world learning, which depends on massive amounts of accurately annotated samples that are expensively labeled by expert SAR analysts and have limited generalization capability and scalability. In this work, we make the first attempt towards building a foundation model for SAR ATR, termed SARATR-X. SARATR-X learns generalizable representations via self-supervised learning (SSL) and provides a cornerstone for label-efficient model adaptation to generic SAR target detection and classification tasks. Specifically, SARATR-X is trained on 0.18 M unlabelled SAR target samples, which are curated by combining contemporary benchmarks and constitute the largest publicly available dataset till now. Considering the characteristics of SAR images, a backbone tailored for SAR ATR is carefully designed, and a two-step SSL method endowed with multi-scale gradient features was applied to ensure the feature diversity and model scalability of SARATR-X. The capabilities of SARATR-X are evaluated on classification under few-shot and robustness settings and detection across various categories and scenes, and impressive performance is achieved, often competitive with or even superior to prior fully supervised, semi-supervised, or self-supervised algorithms. Our SARATR-X and the curated dataset are released at https://***/waterdisappear/SARATR-X to foster research into foundation models for SAR image interpretation.

关键词： radar polarimetry Foundation models remote sensing Imaging Frequency modulation Buildings Object detection Noise Marine vehicles Atmospheric modeling synthetic aperture radar target recognition object detection foundation model self-supervised learning deep learning masked image modeling

来源：评论

学校读者我要写书评

暂无评论

Spatial-Frequency Information Interaction Diffusion for SAR Colorization

Spatial-Frequency Information Interaction Diffusion for SAR ...

引用

2025 IEEE International conference on Acoustics, Speech, and signal processing, ICASSP 2025

作者： Zhang, Xupei Qin, Hanlin Li, Jingjing Geng, Jinni Gao, Zihan Yu, Yue School of Optoelectronic Engineering Xidian University Xi'an710071 China Xi'an Key Laboratory of Infrared Technology and System Xi'an710071 China

ISBN: (纸本)9798350368741

The inherent speckle noise and grayscale characteristics of synthetic aperture radar (SAR) images pose challenges to information perception and interpretation. To address this issue, we propose a novel conditional diffusion model with spatial-frequency information interaction for SAR colorization, namely SFI4SC. Specifically, we introduce a learnable wavelet transform module to obtain the global frequency information of the input SAR image and then take it as a guiding condition to assist the diffusion model for SAR colorization. Additionally, we improved the network structure for the diffusion model by designing and introducing a spatial and frequency information interaction module to achieve multidimensional information interaction for the network inputs, further enhancing perceptual details and geometric structures of the SAR colorization results. Experimental results in the real-world remote sensing dataset show that our approach can successfully enhance visual clarity and informational content while preserving the unique characteristics of the SAR images. The code will be made publicly available. © 2025 IEEE.

关键词： Diffusion Model image colorization Spatial-Frequency Interaction synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

MDDNet: Multilevel Difference-Enhanced Denoise Network for Unsupervised Change Detection in SAR images

MDDNet: Multilevel Difference-Enhanced Denoise Network for U...

引用

2025 IEEE International conference on Acoustics, Speech, and signal processing, ICASSP 2025

作者： Zong, He Zhang, Erlei Li, Xinyu Zhang, Hongming Ren, Jinchang College of Information Engineering Northwest A&F University Yangling China School of Computing and Engineering Robert Gordon University Aberdeen United Kingdom

ISBN: (纸本)9798350368741

Change detection in synthetic aperture radar (SAR) images is a hot yet highly challenging task in remote sensing. Existing unsupervised SAR change detection methods often struggle with inherent speckle noise and insufficiently utilize pseudo-labels, particularly neglecting uncertain areas. In this paper, we propose a multilevel difference-enhanced denoise dual-branch network (MDDNet), comprising representation learning and change detection branches. First, fuzzy c-means clustering is employed to generate pseudo-labels, categorizing the image areas as changed, nochanged, and uncertain. Second, we design a denoise representation loss function in the representation learning branch to maximize the use of pseudo-labels, while mitigating speckle noise. Furthermore, a multilevel difference computation module is proposed to focus on changes in ground objects and capture more comprehensive change information. Experimental results on three public SAR datasets show that the proposed method outperforms six state-of-the-art methods, achieving the best performance with an average overall accuracy of 98.86% and an average Kappa coefficient of 89.36%. © 2025 IEEE.

关键词： Change detection denoise representation multilevel difference computation synthetic aperture radar (SAR) images

来源：评论

学校读者我要写书评

暂无评论

A Sidelobe-Aware Semi-Deformable Convolutional Ship Detection Network for synthetic aperture radar imagery 7th

A Sidelobe-Aware Semi-Deformable Convolutional Ship Detectio...

引用

7th Chinese conference on Pattern Recognition and Computer Vision

作者： Luo, Hao Lin, Xianming Xiamen Univ Minist Educ Key Lab Multimedia Trusted Percept & Efficient Co Xiamen 361005 Peoples R China Xiamen Univ Inst Artificial Intelligence Xiamen 361005 Peoples R China Nanjing Marine Radar Inst Nanjing 211153 Peoples R China

ISBN: (纸本)9789819784929;9789819784936

This paper focus on SAR ship detection, which has been widely used in tasks such as marine traffic, fisheries management, battlefield posture assessment and military target reconnaissance. One popular solution is to utilise deep learning algorithms in conjunction with SAR image object detection. However, due to the unique imaging characteristics of SAR, existing solutions usually do not distinguish well between ships and interference targets in complex backgrounds. In this paper, we address this problem by proposing a sidelobe-aware semi-deformable convolution that takes full advantage of the combination of both standard and deformable convolution to learn high-quality ship features without significantly increasing the computational complexity. Specifically, it makes the feature extraction location more closely fit to the ship shape, strengthens the extraction capability of the region of the ship target itself and the sidelobe features, while reducing the extraction of background information. The channel attention mechanism is then proposed to enhance the ship local detail information extraction. Experiments on two widely used datasets show that the proposed method outperforms the state-of-the-art methods, which is effective and efficient to improve SAR ship detection.

关键词： Deformable Convolution synthetic aperture radar (SAR) imagery Ship Detection remote sensing image processing

来源：评论

学校读者我要写书评

暂无评论

A Multiscale Convolution SAR image Target Recognition Method Based on Complex-Valued Neural Networks

引用

IEEE Journal of Selected Topics in Applied Earth Observations and remote sensing 2025年 18卷 10657-10673页

作者： Hou, Guangyu Xin, Zhihui Liao, Guisheng Huang, Penghui Huang, Yuhao Zou, Rui Yunnan Normal University Yunnan Key Laboratory of Opto-Electronic Information Technology Kunming650500 China Xidian University National Laboratory of Radar Signal Processing Xi'an710071 China Shanghai Jiao Tong University School of Electronic Information and Electrical Engineering Shanghai200240 China

Recent advances in deep learning have driven significant success in synthetic aperture radar (SAR) automatic target recognition, particularly through convolutional neural network (CNN) based classification algorithms. However, SAR images possess distinctive physical scattering properties, owing to their unique imaging mechanism. Many deep learning algorithms rely solely on amplitude information, ignoring phase information, which may result in the loss of information in the original complex-valued SAR image and suboptimal performance. To tackle these problems, this article introduces a SAR target recognition approach based on complex-valued operations, designated as complex-valued residual mish activation and convolution block attention module (CBAM) net (CRMC-Net). The CRMC-Net effectively utilizes the amplitude and phase information in complex-valued SAR data. Specifically, first, the elements of CNN, including the input and output layers, the convolution layers, the activation functions, and the pooling layers, are extended to the complex-valued domain. Second, in order to further enhance the representation ability of the model, multiscale information of the target is extracted through different convolution kernel sizes. Finally, the network constructs many complex-valued operation blocks to enhance the robustness of the designed network, including the complex-valued residual block, complex-valued Mish activation function, and complex-valued CBAM. The experimental results obtained from the moving and stationary target capture and recognition dataset and OpenSARShip2.0 dataset demonstrate that the proposed network model outperforms the traditional real-valued models. It can further reduce the classification error and enhance performance. © 2008-2012 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Advancing SAR image Robustness: Integrating Diffusion Models for Adversarial Purification and Speckle Noise Suppression

Advancing SAR Image Robustness: Integrating Diffusion Models...

引用

2025 IEEE International conference on Acoustics, Speech, and signal processing, ICASSP 2025

作者： Xie, Ganglin Lu, Haobo Zhang, Xinze He, Kun School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China

ISBN: (纸本)9798350368741

synthetic aperture radar (SAR) is a powerful tool for ground target detection, but SAR images often suffer from coherent speckle noise, which complicates automatic target recognition (ATR). Recent advances in deep learning have shown promise in SAR-ATR, yet there is a notable gap in addressing adversarial attacks and despeckling challenges. This study presents SAR Adversarial Diffusion Purification (SAR-ADP), a novel model designed to enhance SAR image analysis by tackling both adversarial attacks and coherent speckle noise. Our main contributions include pioneering methods for SAR image adversarial defense, simultaneous adversarial noise removal and despeckling, guidance mechanisms within the diffusion model, and a streamlined architecture for SAR image purification. SAR-ADP integrates adversarial defense strategies, diffusion modeling, and despeckling techniques, resulting in significant improvements in image recognition accuracy and post-processing. Empirical results validate its effectiveness, achieving a robustness accuracy of approximately 97% and a standard accuracy of around 99%. © 2025 IEEE.

关键词： adversarial defense despeckling Diffusion model synthetic aperture radar image target recognition

来源：评论

学校读者我要写书评

暂无评论

PolSAR image classification using complex-valued multiscale attention vision transformer (CV-MsAtViT)

引用

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 2025年 137卷

作者： Alkhatib, Mohammed Q. Univ Dubai Coll Engn & IT Dubai 14143 U Arab Emirates

This paper Introduces a novel method for Polarimetric synthetic aperture radar (PolSAR) image classification using a Complex-Valued Multiscale Attention Vision Transformer (CV-MsAtViT). The model incorporates a complex-valued multiscale feature fusion mechanism, a complex-valued attention block, and a Complex-Valued Vision Transformer (CV-ViT) to effectively capture spatial and polarimetric features from PolSAR data. The multiscale fusion block enhances feature extraction, while the attention mechanism prioritizes critical features, and the CV-ViT processes data in the complex domain, preserving both amplitude and phase information. Experimental results on benchmark PolSAR datasets, including Flevoland, San Francisco, and Oberpfaffenhofen, show that CV-MsAtViT achieves superior classification accuracy, with an overall accuracy (OA) of 98.35% on the Flevoland dataset, outperforming state-of-the-art models like PolSARFormer. The model also demonstrates efficient computational performance, minimizing the number of parameters while preserving high accuracy. These results confirm that CV-MsAtViT effectively enhances the classification of PolSAR images by leveraging complex-valued data processing, offering a promising direction for future advancements in remote sensing and complex-valued deep learning. The codes associated with this paper are publicly available at https://***/mqalkhatib/CV-MsAtViT.

关键词： PolSAR image classification Complex-valued (CV) networks Feature fusion CV coordinate attention CV vision transformers (CV-viT)

来源：评论

学校读者我要写书评

暂无评论

Design and in-orbit application of Ku-band high resolution intelligent SAR satellite system

Design and in-orbit application of Ku-band high resolution i...

引用

2024 International conference on remote sensing and Digital Earth, RSDE 2024

作者： Chang, Wujun Jiang, Xiupeng Gao, Enyu Yan, Wei Song, Jiexuan Guo, Shuai Wu, Jiang Li, Yanming Jia, Zhihui Su, Fan Beijing Minospace Technology Co. Ltd. Beijing100094 China Xi’an Institute of Space Radio Technology Xi’an710100 China The Institute of Spacecraft Application System Engineering CAST Beijing100094 China

ISBN: (数字)9781510688223

ISBN: (纸本)9781510688216

TaiJing-4(03) satellite is the world's first high resolution flat-panel Ku-band synthetic aperture radar (SAR) commercial satellite, characterized by the high resolution, light weight design, low cost, multiple imaging modes and long-term operation. A series of key technologies and innovations are adopted, such as lightweight & high resolution Ku-band SAR payload, on-board intelligent imaging processing, non-uniform pulse sparse reconstruction of high resolution and wide width SAR imaging technology, target detection and recognition capability improvement, high precision and high stability attitude measurement and control, 2D guidance composite control, and multi-target imaging efficiency improvement. It achieves high resolution, high image quality and high image acquisition efficiency. The Ku-band SAR satellite has stronger perception of target details, and has multiple workmodes such as spotlight, sliding spotlight, strip, scanning, wide-width scanning, and multi-target imaging, which significantly improves satellite observation performance, coverage capability and application efficiency. It has important applications in high-resolution observation, emergency disaster relief, marine environment monitoring, resource investigation, etc. © 2025 SPIE.

关键词： synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

Characterisation of Resident Space Objects and Synchronisation Error Compensation in Multistatic Interferometric Inverse synthetic aperture radar Imaging

引用

IET radar SONAR AND NAVIGATION 2025年第1期19卷

作者： Rudrappa, Manjunath Thindlu Kaeske, Martin Albrecht, Marcus Knott, Peter Fraunhofer Inst High Frequency Phys & Radar Tech F Wachtberg Germany

This article presents a study on interferometric inverse synthetic aperture radar (ISAR) for three-dimensional imaging and rotational velocity estimation. The study focuses on synchronisation error compensation in a multistatic setup with three ground-based radars and non-orthogonal baselines. The simulation involves CAD models of non-cooperative targets in Keplerian orbits with different orbital parameters, and radar backscattering along the orbit is simulated based on physical optics approximation. The paper also illustrates the signal processing chain for synchronisation error compensation and the generation of interferometric pointclouds for the resident space object models in orbit. The study includes multilabel classification and performance analysis of synchronisation error compensation at varying SNR using Monte Carlo simulations. The interferometric reconstruction and classification accuracy at low SNR conditions is enhanced using multiple receiver or multi channel fusion and the performance of the fusion algorithm is evaluated at varying noise correlation between the fused channels or receivers.

关键词： image fusion inverse synthetic aperture radar (ISAR) multistatic radar object recognition radar imaging radar signal processing radar target recognition remote sensing by radar

来源：评论

学校读者我要写书评

暂无评论

Evaluating multi-modal image matching algorithms for remote sensing applications 6

Evaluating multi-modal image matching algorithms for remote ...

引用

6th International conference on Geoscience and remote sensing Mapping, GRSM 2024

作者： Liu, Zhanquan Han, Yilong Fu, Mingsheng Song, Baiwan Gong, Danchao Ming, Yanfang Sun, Lin College of Geodesy and Geomatics Shandong University of Science and Technology Qingdao266590 China Shandong Province Research Center of Engineering and Technology for GPS RS and GIS Qingdao266590 China Star-Rising Technologies Co. Ltd. Qingdao266590 China Chinese Academy of Surveying and Mapping Beijing100036 China Shandong GEO-Surveying & Mapping Institute Jinan250000 China State Key Laboratory of Geo-Information Engineering Xi’an710054 China Xi’an Institute of Surveying and Mapping Xi’an710054 China

ISBN: (纸本)9781510688018

image matching is a fundamental task in image processing, particularly in photogrammetry and remote sensing. Conventional image matching methods often struggle with multi-modal images due to significant geometric distortions and nonlinear radiometric effects. To address these challenges, researchers have developed a variety of algorithms specifically tailored for multi-modal data. Despite this, there is a notable lack of comprehensive theoretical summaries of multi-modal image matching. This paper aims to bridge this gap by discussing the complexities of matching multi-modal images and providing a succinct overview of existing methods. We evaluate three Conventional algorithms—OS-SIFT, RIFT, and POS-GIFT—and two deep learning methods—SemLA and SuperPoint+SuperGlue. These methods are assessed using six types of multi-modal remote sensing datasets: Optical-Optical, Optical-Infrared, Optical-Depth, Optical-Map, Optical-synthetic aperture radar (SAR), and Day-Night. The results indicate that the POS-GIFT algorithm outperforms others, achieving an average of 1217 correct matches with a success rate of 96.70%. This paper concludes by discussing the challenges of multi-modal image matching and suggesting future research directions, serving as a valuable resource for researchers and practitioners in remote sensing. © 2025 SPIE.

关键词： image matching

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：