Article
Engineering, Electrical & Electronic
Paul Connetable, Allan Aasbjerg Nielsen, Knut Conradsen, Ernst Krogager, Henning Skriver
Summary: The increased amount of information from fully polarimetric synthetic aperture radar (SAR) provides additional knowledge about ground scatterers. Utilizing the polarimetric information effectively is crucial for target detection. Different approaches have been proposed to summarize the information into polarimetric features, which have been studied in relation to the physical properties of the scatterers. This article aims to review the polarimetric features used for target detection and combine them optimally for vehicle detection in open fields, based on a large airborne dataset in X-, S-, and L-bands.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Engineering, Electrical & Electronic
Peng Chen, Hui Zhou, Ying Li, Bingxin Liu, Peng Liu
Summary: The paper introduces a new hybrid model to improve the detection accuracy of ship targets in complex marine environments, using a multitask loss function combined with classification, localization, and segmentation based on the improved IoU for boundary-box localization. Experimental results show that the model achieves high detection accuracy in complex scenarios with a significantly lower false-positive rate compared to other models.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Engineering, Electrical & Electronic
Paul Connetable, Knut Conradsen, Allan Aasbjerg Nielsen, Henning Skriver
Summary: This paper investigates the removal of speckle and the characteristics of the local covariance matrix in polarimetric synthetic aperture radar (SAR) images. Test statistics based on block-diagonality and change detection are used to detect man-made structures with high potential. The results show that the method is effective in highlighting buildings and urban areas, as well as in ship detection at longer wavelengths.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Shen Tan, Xin Zhang, Han Wang, Le Yu, Yanlei Du, Junjun Yin, Bingfang Wu
Summary: A novel self-supervised SAR denoising model is proposed, which can be trained without a noise-free image and restore spatial details through a hybrid loss function. Experiments show that the method outperforms traditional approaches in terms of noise reduction and feature preservation.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Junjie Wang, Feng Gao, Junyu Dong, Qian Du, Heng-Chao Li
Summary: This study proposes a dual path denoising network (DPDNet) for SAR image change detection, which utilizes random label propagation and distinctive patch convolution techniques to enhance computational efficiency and accuracy.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Geunhyuk Youk, Munchurl Kim
Summary: This article introduces the research status of deep-learning-based target recognition in synthetic aperture radar (SAR) images. It is expensive to collect large numbers of labeled SAR images for training high-performance classification networks. To solve the problem of insufficient SAR data, electromagnetic computational tools are used to synthesize measured SAR target images from data modeling. However, there is a large domain gap between synthetic and measured SAR images, leading to poor classification performance on measured SAR target images. To bridge this gap, a novel transformer-based synthetic-to-measured SAR target image translation network, SAR-SMT Net, is proposed.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Engineering, Electrical & Electronic
Matthew Dvorsky, Mohammad Tayeb Al Qaseer, Reza Zoughi
Summary: This paper describes a synthetic aperture radar (SAR) 3-D polarimetry method using a radially polarized antenna. By developing a 3-D target scattering model and an image reconstruction algorithm, 3-D polarimetry imaging and orientation estimation of targets can be achieved.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Geochemistry & Geophysics
Yuzhen Niu, Yuezhou Li, Jiangyi Huang, Yuzhong Chen
Summary: In this study, an efficient encoder-decoder network with estimated ship direction was proposed for ship detection in SAR images. The network achieved state-of-the-art performance by extracting multiple-level features and addressing the challenge of overlapped annotations.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Gang Xu, Bangjie Zhang, Hanwen Yu, Jianlai Chen, Mengdao Xing, Wei Hong
Summary: Synthetic aperture radar (SAR) image formation is an ill-posed linear inverse problem, and the resolution is limited by the data bandwidth in traditional imaging techniques. Sparse SAR imaging technology using compressed sensing (CS) has improved the performance by achieving superresolution and feature enhancement. Recently, machine learning (ML), including deep learning (DL), has been further studied for sparse SAR imaging and has shown great potential, but there are still gaps between the two groups of methods and their connections have not been established.
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE
(2022)
Article
Geochemistry & Geophysics
Wei Xu, Weida Xing, Chonghua Fang, Pingping Huang, Weixian Tan
Summary: The article proposes a modified two-step notch filtering approach combined with linear prediction to improve the SAR image quality. This method effectively mitigates narrowband RFI energy while recovering missing range spectral components.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2021)
Article
Geochemistry & Geophysics
Biao Hou, Zitong Wu, Bo Ren, Zhonghua Li, Xianpeng Guo, Shuang Wang, Licheng Jiao
Summary: A semisupervised SAR ship detection network, SCLANet, is proposed in this study, which improves the algorithm performance by utilizing unlabeled data. SCLANet trains the unlabeled data through consistency learning and achieves high accuracy in ship detection.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Luyang Zhang, Haitao Wang, Lingfeng Wang, Chunhong Pan, Chunlei Huo, Qiang Liu, Xinyao Wang
Summary: A new filter convolution design is proposed to suppress coherent speckle noise in SAR images while extracting features, and it demonstrates excellent applicability.
Article
Engineering, Electrical & Electronic
Jason M. Merlo, Jeffrey A. Nanzer
Summary: This article presents a low-frequency SAR radar for automotive use and showcases the opportunity to infer landmarks through measurements of co-polarized and cross-polarized scattering.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2022)
Article
Geochemistry & Geophysics
Yiru Lu, Biao Zhang, William Perrie
Summary: We propose an automatic classification approach for Arctic sea ice and open water using fully polarimetric spaceborne C-band RADARSAT-2 synthetic aperture radar (SAR) imagery. The approach is based on a random forest (RF) model that utilizes HH and HV polarized radar backscatter, incidence angle, and optimal polarimetric features as inputs. Through a physics-based unsupervised scheme, well-annotated sea ice and open water samples are generated, reducing labeling errors. The RF model is trained and validated using over one million labeled samples, achieving a sea ice and water classification accuracy of 99.94% and a Kappa coefficient of 0.999. Optimal polarimetric features improve ice-water discrimination accuracy by about 4%-10%, with polarization difference being the foremost parameter for distinguishing sea ice from open water. High-resolution ice-water classification maps demonstrate clear separation of sea ice leads and their surroundings.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Yiru Lu, Biao Zhang, William Perrie, Alexis Mouche, Guosheng Zhang
Summary: The study validates the performance of the new HH-polarized geophysical model function CMODH for C-band synthetic aperture radar ocean surface wind speed retrieval. The wind speeds estimated using CMODH are consistent with buoy measurements and show good accuracy and consistency in the validation with in-situ data. Compared to other models, CMODH achieves the smallest root mean square error for wind speed retrieval.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2021)
