Cargando…

Target Recognition in SAR Images by Deep Learning with Training Data Augmentation

Mass production of high-quality synthetic SAR training imagery is essential for boosting the performance of deep-learning (DL)-based SAR automatic target recognition (ATR) algorithms in an open-world environment. To address this problem, we exploit both the widely used Moving and Stationary Target A...

Descripción completa

Detalles Bibliográficos
Autores principales: Geng, Zhe, Xu, Ying, Wang, Bei-Ning, Yu, Xiang, Zhu, Dai-Yin, Zhang, Gong
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9863010/
https://www.ncbi.nlm.nih.gov/pubmed/36679740
http://dx.doi.org/10.3390/s23020941
_version_ 1784875231208275968
author Geng, Zhe
Xu, Ying
Wang, Bei-Ning
Yu, Xiang
Zhu, Dai-Yin
Zhang, Gong
author_facet Geng, Zhe
Xu, Ying
Wang, Bei-Ning
Yu, Xiang
Zhu, Dai-Yin
Zhang, Gong
author_sort Geng, Zhe
collection PubMed
description Mass production of high-quality synthetic SAR training imagery is essential for boosting the performance of deep-learning (DL)-based SAR automatic target recognition (ATR) algorithms in an open-world environment. To address this problem, we exploit both the widely used Moving and Stationary Target Acquisition and Recognition (MSTAR) SAR dataset and the Synthetic and Measured Paired Labeled Experiment (SAMPLE) dataset, which consists of selected samples from the MSTAR dataset and their computer-generated synthetic counterparts. A series of data augmentation experiments are carried out. First, the sparsity of the scattering centers of the targets is exploited for new target pose synthesis. Additionally, training data with various clutter backgrounds are synthesized via clutter transfer, so that the neural networks are better prepared to cope with background changes in the test samples. To effectively augment the synthetic SAR imagery in the SAMPLE dataset, a novel contrast-based data augmentation technique is proposed. To improve the robustness of neural networks against out-of-distribution (OOD) samples, the SAR images of ground military vehicles collected by the self-developed MiniSAR system are used as the training data for the adversarial outlier exposure procedure. Simulation results show that the proposed data augmentation methods are effective in improving both the target classification accuracy and the OOD detection performance. The purpose of this work is to establish the foundation for large-scale, open-field implementation of DL-based SAR-ATR systems, which is not only of great value in the sense of theoretical research, but is also potentially meaningful in the aspect of military application.
format Online
Article
Text
id pubmed-9863010
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-98630102023-01-22 Target Recognition in SAR Images by Deep Learning with Training Data Augmentation Geng, Zhe Xu, Ying Wang, Bei-Ning Yu, Xiang Zhu, Dai-Yin Zhang, Gong Sensors (Basel) Article Mass production of high-quality synthetic SAR training imagery is essential for boosting the performance of deep-learning (DL)-based SAR automatic target recognition (ATR) algorithms in an open-world environment. To address this problem, we exploit both the widely used Moving and Stationary Target Acquisition and Recognition (MSTAR) SAR dataset and the Synthetic and Measured Paired Labeled Experiment (SAMPLE) dataset, which consists of selected samples from the MSTAR dataset and their computer-generated synthetic counterparts. A series of data augmentation experiments are carried out. First, the sparsity of the scattering centers of the targets is exploited for new target pose synthesis. Additionally, training data with various clutter backgrounds are synthesized via clutter transfer, so that the neural networks are better prepared to cope with background changes in the test samples. To effectively augment the synthetic SAR imagery in the SAMPLE dataset, a novel contrast-based data augmentation technique is proposed. To improve the robustness of neural networks against out-of-distribution (OOD) samples, the SAR images of ground military vehicles collected by the self-developed MiniSAR system are used as the training data for the adversarial outlier exposure procedure. Simulation results show that the proposed data augmentation methods are effective in improving both the target classification accuracy and the OOD detection performance. The purpose of this work is to establish the foundation for large-scale, open-field implementation of DL-based SAR-ATR systems, which is not only of great value in the sense of theoretical research, but is also potentially meaningful in the aspect of military application. MDPI 2023-01-13 /pmc/articles/PMC9863010/ /pubmed/36679740 http://dx.doi.org/10.3390/s23020941 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Geng, Zhe
Xu, Ying
Wang, Bei-Ning
Yu, Xiang
Zhu, Dai-Yin
Zhang, Gong
Target Recognition in SAR Images by Deep Learning with Training Data Augmentation
title Target Recognition in SAR Images by Deep Learning with Training Data Augmentation
title_full Target Recognition in SAR Images by Deep Learning with Training Data Augmentation
title_fullStr Target Recognition in SAR Images by Deep Learning with Training Data Augmentation
title_full_unstemmed Target Recognition in SAR Images by Deep Learning with Training Data Augmentation
title_short Target Recognition in SAR Images by Deep Learning with Training Data Augmentation
title_sort target recognition in sar images by deep learning with training data augmentation
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9863010/
https://www.ncbi.nlm.nih.gov/pubmed/36679740
http://dx.doi.org/10.3390/s23020941
work_keys_str_mv AT gengzhe targetrecognitioninsarimagesbydeeplearningwithtrainingdataaugmentation
AT xuying targetrecognitioninsarimagesbydeeplearningwithtrainingdataaugmentation
AT wangbeining targetrecognitioninsarimagesbydeeplearningwithtrainingdataaugmentation
AT yuxiang targetrecognitioninsarimagesbydeeplearningwithtrainingdataaugmentation
AT zhudaiyin targetrecognitioninsarimagesbydeeplearningwithtrainingdataaugmentation
AT zhanggong targetrecognitioninsarimagesbydeeplearningwithtrainingdataaugmentation