Automatic Target Recognition in Synthetic Aperture Radar Imagery a Stateoftheart Review

Special Issue "Target Recognition in Synthetic Aperture Radar Imagery"

Share This Special Issue

Special Issue Editors

Dr. Luca Pallotta
Email Website
Guest Editor

Department of Engineering, University of Roma Tre, via Vito Volterra 62, 00146 Rome, Italy
Interests: statistical signal processing with emphasis on radar/SAR signal processing; radar targets classification; polarimetric radar/SAR

Dr. Christos Ilioudis
E-mail Website
Guest Editor

Electronic & Electrical Technology, Academy of Strathclyde, 204 George St, Glasgow G1 1XW, Scotland, UK
Interests: automatic target recognition; passive/forrad scattering radars; motions modelling and micro-doppler analysis; articulation radar advice operations; MIMO Radar; cognitive radars and AI

Special Issue Information

Dear Colleagues,

In recent years the interest towards the development of algorithms aimed at automatically classifying targets in Synthetic Aperture Radar (SAR) images is growing more and more. Peculiarly, the knowledge of the types of man-made objects (like missile launchers, vehicles, planes) that are positioned in the observed scene could exist a task of paramount importance in the modernistic surveillance systems to understand possible threats in military contexts, simply also to properly manage some activities in a specific surface area in ceremonious environments.

The telescopic of this Special Event is to provide an overview of betoken processing methods for target recognition. Contributions to the body of knowledge in the field could exist from polarimetric synthetic discontinuity radar (SAR), inverse SAR (ISAR) and passive bistatic radar, with applications of involvement in automatic target recognition (ATR) and its lower level tasks (identification, characterization and fingerprinting).

The application of Artificial Intelligence (AI) techniques to ATR are also very welcomed, as it recently proved to represent an interesting and useful alternate processing strategy. The efforts in this field should highlight the capabilities and limitations of AI for effective application to ATR problems.

Dr. Addabbo Pia
Dr. Luca Pallotta
Dr. Christos Ilioudis
Guest Editors

Manuscript Submission Data

Manuscripts should be submitted online at www.mdpi.com past registering and logging in to this website. Once yous are registered, click here to go to the submission class. Manuscripts tin can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as presently as accustomed) and will be listed together on the special upshot website. Research articles, review articles also as short communications are invited. For planned papers, a title and short abstruse (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should non have been published previously, nor exist under consideration for publication elsewhere (except briefing proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Remote Sensing is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page earlier submitting a manuscript. The Commodity Processing Accuse (APC) for publication in this open access journal is 2500 CHF (Swiss Francs). Submitted papers should be well formatted and use good English language. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

• Synthetic Aperture Radar
• Automatic Target Recognition
• Classification
• Features Extraction
• Artificial Intelligence
• Machine Learning
• Neural Networks

Published Papers (five papers)

Research

Open up Access Article

PeaceGAN: A GAN-Based Multi-Task Learning Method for SAR Target Image Generation with a Pose Calculator and an Auxiliary Classifier

Cited by 1 | Viewed by 730

Abstruse

Although generative adversarial networks (GANs) are successfully applied to diverse fields, grooming GANs on synthetic aperture radar (SAR) data is a challenging task due to speckle noise. On the one manus, in a learning perspective of human perception, it is natural to learn [...] Read more than.

Although generative adversarial networks (GANs) are successfully practical to diverse fields, preparation GANs on constructed aperture radar (SAR) data is a challenging job due to speckle noise. On the i hand, in a learning perspective of human perception, it is natural to learn a task by using information from multiple sources. However, in the previous GAN works on SAR image generation, data on target classes has only been used. Due to the backscattering characteristics of SAR signals, the structures of SAR images are strongly dependent on their pose angles. Nevertheless, the pose angle data has not been incorporated into GAN models for SAR images. In this paper, we advise a novel GAN-based multi-chore learning (MTL) method for SAR target image generation, called PeaceGAN, that has two additional structures, a pose estimator and an auxiliary classifier, at the side of its discriminator in society to effectively combine the pose and form information via MTL. Extensive experiments showed that the proposed MTL framework can assist the PeaceGAN's generator effectively acquire the distributions of SAR images so that it tin better generate the SAR target images more than faithfully at intended pose angles for desired target classes in comparison with the recent land-of-the-fine art methods. Full commodity

► ▼ Show Figures

Open Admission Communication

FEF-Net: A Deep Learning Approach to Multiview SAR Epitome Target Recognition

Viewed past 759

Abstract

Synthetic aperture radar (SAR) is an advanced microwave imaging system of great importance. The recognition of real-earth targets from SAR images, i.e., automatic target recognition (ATR), is an bonny but challenging issue. The bulk of existing SAR ATR methods are designed for single-view [...] Read more.

Synthetic aperture radar (SAR) is an advanced microwave imaging arrangement of great importance. The recognition of existent-globe targets from SAR images, i.e., automatic target recognition (ATR), is an attractive merely challenging event. The bulk of existing SAR ATR methods are designed for single-view SAR images. However, multiview SAR images contain more abundant classification data than single-view SAR images, which benefits automatic target classification and recognition. This paper proposes an finish-to-end deep feature extraction and fusion network (FEF-Internet) that tin effectively exploit recognition information from multiview SAR images and can boost the target recognition performance. The proposed FEF-Cyberspace is based on a multiple-input network structure with some distinct and useful learning modules, such as deformable convolution and squeeze-and-excitation (SE). Multiview recognition information tin can exist finer extracted and fused with these modules. Therefore, splendid multiview SAR target recognition performance can be achieved by the proposed FEF-Cyberspace. The superiority of the proposed FEF-Internet was validated based on experiments with the moving and stationary target conquering and recognition (MSTAR) dataset. Total article

► ▼ Show Figures

Open up Access Feature Paper Article

Automatic Target Recognition for Low Resolution Foliage Penetrating SAR Images Using CNNs and GANs

Cited past 1 | Viewed past 1002

Abstruse

In recent years, the technological advances leading to the production of loftier-resolution Synthetic Aperture Radar (SAR) images has enabled more and more effective target recognition capabilities. Still, high spatial resolution is not always achievable, and, for some particular sensing modes, such every bit Foliage [...] Read more.

In contempo years, the technological advances leading to the production of high-resolution Synthetic Discontinuity Radar (SAR) images has enabled more and more constructive target recognition capabilities. Still, high spatial resolution is non always achievable, and, for some particular sensing modes, such as Foliage Penetrating Radars, depression resolution imaging is often the only selection. In this paper, the problem of automatic target recognition in Depression Resolution Foliage Penetrating (FOPEN) SAR is addressed through the apply of Convolutional Neural Networks (CNNs) able to extract both low and high level features of the imaged targets. Additionally, to address the issue of limited dataset size, Generative Adversarial Networks are used to enlarge the grooming fix. Finally, a Receiver Operating Characteristic (ROC)-based post-nomenclature decision arroyo is used to reduce nomenclature errors and measure the capability of the classifier to provide a reliable output. The effectiveness of the proposed framework is demonstrated through the use of real SAR FOPEN data. Total article

► ▼ Show Figures

Open Access Article

Non-Linear Modeling of Detectability of Ship Wake Components in Dependency to Influencing Parameters Using Spaceborne 10-Band SAR

Cited by 5 | Viewed by 1041

Abstruse

The detection of the wakes of moving ships in Constructed Aperture Radar (SAR) imagery requires the presence of wake signatures, which are sufficiently distinctive from the ocean background. Diverse wake components exist, which institute the SAR signatures of ship wakes. For successful wake [...] Read more.

The detection of the wakes of moving ships in Constructed Aperture Radar (SAR) imagery requires the presence of wake signatures, which are sufficiently distinctive from the ocean background. Various wake components exist, which constitute the SAR signatures of ship wakes. For successful wake detection, the contrast between the detectable wake components and the background is crucial. The detectability of those wake components is afflicted by a number of parameters, which represent the image acquisition settings, environmental conditions or ship backdrop including voyage information. In this study the dependency of the detectability of private wake components to these parameters is characterized. For each wake component a detectability model is built, which takes the influence of incidence angle, polarization, wind speed, wind direction, sea state (significant wave height, wavelength, wave direction), vessel's velocity, vessel's course over basis and vessel'due south length into account. The presented detectability models are based on regression or nomenclature using Back up Vector Machines and a dataset of manually labelled TerraSAR‑X wake samples. The considered wake components are: about‑hull turbulences, turbulent wakes, Kelvin wake arms, Kelvin wake'due south transverse waves, Kelvin wake's divergent waves, V‑narrow wakes and send‑generated internal waves. The statements derived about wake component detectability are mainly in skillful agreement with statements from previous research, simply too some new assumptions are provided. The about expressive influencing parameter is the movement velocity of the vessels, as all wake components are more detectable the faster vessels move. Total commodity

► ▼ Testify Figures

Open Access Article

When Deep Learning Meets Multi-Job Learning in SAR ATR: Simultaneous Target Recognition and Partition

Cited past half dozen | Viewed by 832

Abstract

With the contempo advances of deep learning, automated target recognition (ATR) of constructed discontinuity radar (SAR) has accomplished superior functioning. By not existence limited to the target category, the SAR ATR organisation could benefit from the simultaneous extraction of multifarious target attributes. In [...] Read more.

With the contempo advances of deep learning, automated target recognition (ATR) of synthetic aperture radar (SAR) has achieved superior performance. Past non being express to the target category, the SAR ATR system could benefit from the simultaneous extraction of multifarious target attributes. In this paper, we propose a new multi-chore learning approach for SAR ATR, which could obtain the accurate category and precise shape of the targets simultaneously. Past introducing deep learning theory into multi-chore learning, we commencement propose a novel multi-task deep learning framework with two main structures: encoder and decoder. The encoder is constructed to extract sufficient prototype features in different scales for the decoder, while the decoder is a tasks-specific construction which employs these extracted features adaptively and optimally to meet the unlike feature demands of the recognition and segmentation. Therefore, the proposed framework has the ability to attain superior recognition and segmentation performance. Based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset, experimental results show the superiority of the proposed framework in terms of recognition and segmentation. Full article

► ▼ Show Figures

We use cookies on our website to ensure yous become the best experience.
Read more about our cookies hither.

We take just recently launched a new version of our website.
Help usa to further improve by taking role in this brusk five minute survey here. here.

brownbeirme.blogspot.com

Source: https://www.mdpi.com/journal/remotesensing/special_issues/TR_SAR

Share this post