Yujie Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yujie Zhang has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Xueru Bai
Xidian University
7 shared publications - Feng Zhou
The School of Aerospace Science and Technology, Xidian University, Xi’an 710071, China
4 shared publications - Jinqi Wang
Xidian University
1 shared publication Shihao Liu
National Key Laboratory of Radar Signal Processing, Xidian University, Xian, China
1 shared publicationSiqi Liu
National Laboratory of Radar Signal Processing, Xidian University, Xian, China
1 shared publication- Xiaoyong Li
Hangzhou Institute of Technology, Xidian University, Hangzhou, China
1 shared publication
Latest Publications
- 23 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 62 (00189251) , 3609-3622
Robust High-Resolution ISAR Imaging Based on Unfolded Diffusion Model
Achieving robust imaging under complex and varying observation conditions is crucial for the practical application of ISAR. However, existing iterative sparse signal reconstruction methods require manual parameter tuning, while the unfolded networks necessitate retraining under varying conditions, both lacking capability for robust imaging. To address these issues, this paper proposes a robust ISAR imaging method based on an unfolded diffusion model (UDM). Firstly, an iterative algorithm for ISAR imaging is proposed within the Bayesian framework, where a deep denoising network is incorporated to learn the image prior, enhancing the model representation flexibility. Then, the iterative algorithm is unfolded into a deep network and integrated with a diffusion model, where the signal-to-noise ratio (SNR) and data missing rate (DMR) information are effectively integrated into the image reconstruction process through a lightweight attention module, thereby achieving robust imaging under varying SNR and DMR conditions. Finally, experimental results under different observation conditions have demonstrated the superiority and effectiveness of the proposed method.
- 23 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 62 (00189251) , 3609-3622
Robust High-Resolution ISAR Imaging Based on Unfolded Diffusion Model
Achieving robust imaging under complex and varying observation conditions is crucial for the practical application of ISAR. However, existing iterative sparse signal reconstruction methods require manual parameter tuning, while the unfolded networks necessitate retraining under varying conditions, both lacking capability for robust imaging. To address these issues, this paper proposes a robust ISAR imaging method based on an unfolded diffusion model (UDM). Firstly, an iterative algorithm for ISAR imaging is proposed within the Bayesian framework, where a deep denoising network is incorporated to learn the image prior, enhancing the model representation flexibility. Then, the iterative algorithm is unfolded into a deep network and integrated with a diffusion model, where the signal-to-noise ratio (SNR) and data missing rate (DMR) information are effectively integrated into the image reconstruction process through a lightweight attention module, thereby achieving robust imaging under varying SNR and DMR conditions. Finally, experimental results under different observation conditions have demonstrated the superiority and effectiveness of the proposed method.
- 14 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-16
FSD3: Few-Shot SAR Object Detection With Dynamic Perception RPN and Category Knowledge Decoder
Limited by observational constraints and non-cooperative targets, SAR images often suffer from extreme scarcity of training samples for certain classes, which severely degrades the detection performance and leads to the challenging issue of few-shot SAR object detection. To address this, we propose a Few-Shot SAR object Detection model with Dynamic perception RPN and category knowledge Decoder, dubbed as FSD3. Specifically, a dynamic perception RPN is employed to infuse global semantics of support classes into local features of query feature maps, thereby enhancing the recall rate of novel-class targets under few-shot conditions. Subsequently, a category knowledge decoder is designed, which automatically decodes aggregated features by utilizing query features as prompts and effectively mitigates the feature-label mismatch phenomenon. Finally, a dual-constrained detection head is proposed, which leverages the large margin cosine loss and center calibration loss to explicitly achieve inter-class separation and intra-class compactness, thereby achieving high-precision and robust classification in few-shot scenarios. Experimental results on the public SAR-AIRCraft-1.0 dataset and our custom MSTAR-Fewshot dataset have demonstrated that the proposed model achieves superior few-shot object detection performance across various experimental configurations.
- 14 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-16
FSD3: Few-Shot SAR Object Detection With Dynamic Perception RPN and Category Knowledge Decoder
Limited by observational constraints and non-cooperative targets, SAR images often suffer from extreme scarcity of training samples for certain classes, which severely degrades the detection performance and leads to the challenging issue of few-shot SAR object detection. To address this, we propose a Few-Shot SAR object Detection model with Dynamic perception RPN and category knowledge Decoder, dubbed as FSD3. Specifically, a dynamic perception RPN is employed to infuse global semantics of support classes into local features of query feature maps, thereby enhancing the recall rate of novel-class targets under few-shot conditions. Subsequently, a category knowledge decoder is designed, which automatically decodes aggregated features by utilizing query features as prompts and effectively mitigates the feature-label mismatch phenomenon. Finally, a dual-constrained detection head is proposed, which leverages the large margin cosine loss and center calibration loss to explicitly achieve inter-class separation and intra-class compactness, thereby achieving high-precision and robust classification in few-shot scenarios. Experimental results on the public SAR-AIRCraft-1.0 dataset and our custom MSTAR-Fewshot dataset have demonstrated that the proposed model achieves superior few-shot object detection performance across various experimental configurations.
- 8 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Radar Systems
- Vol. 3, 995-1005
High-Resolution ISAR Imaging of Maneuvering Targets Based on 2-D Complex Fast Sparse Bayesian Learning
The maneuvering of the targets will induce time-varying Doppler during observation, posing great challenges for well-focused ISAR imaging. Furthermore, ISAR may encounter complex observation conditions such as incomplete data and low signal-to-noise ratio (SNR), which render the conventional maneuvering targets imaging methods invalid. To address these issues, this paper proposes a novel high-resolution ISAR imaging method of maneuvering targets. Firstly, the sparse imaging model of maneuvering targets is constructed by incorporating the rotation parameters into the observation dictionary. Then, the gamma-complex Gaussian prior is assigned to the ISAR image to exploit its sparse nature. On this basis, to circumvent the matrix inversion embedded in the traditional SBL method, the model lower bound is relaxed and a novel algorithm is proposed for efficient ISAR image reconstruction, dubbed 2D complex fast SBL (2D-CFSBL). Furthermore, the maximum likelihood estimation is utilized to estimate the rotation parameters accurately. Finally, ISAR image reconstruction and rotation parameters estimation are performed iteratively to obtain well-focused image. Experimental results have validated the effectiveness and superiority of the proposed method under incomplete data and low SNR conditions.
- 8 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Radar Systems
- Vol. 3, 995-1005
High-Resolution ISAR Imaging of Maneuvering Targets Based on 2-D Complex Fast Sparse Bayesian Learning
The maneuvering of the targets will induce time-varying Doppler during observation, posing great challenges for well-focused ISAR imaging. Furthermore, ISAR may encounter complex observation conditions such as incomplete data and low signal-to-noise ratio (SNR), which render the conventional maneuvering targets imaging methods invalid. To address these issues, this paper proposes a novel high-resolution ISAR imaging method of maneuvering targets. Firstly, the sparse imaging model of maneuvering targets is constructed by incorporating the rotation parameters into the observation dictionary. Then, the gamma-complex Gaussian prior is assigned to the ISAR image to exploit its sparse nature. On this basis, to circumvent the matrix inversion embedded in the traditional SBL method, the model lower bound is relaxed and a novel algorithm is proposed for efficient ISAR image reconstruction, dubbed 2D complex fast SBL (2D-CFSBL). Furthermore, the maximum likelihood estimation is utilized to estimate the rotation parameters accurately. Finally, ISAR image reconstruction and rotation parameters estimation are performed iteratively to obtain well-focused image. Experimental results have validated the effectiveness and superiority of the proposed method under incomplete data and low SNR conditions.
- 10 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Imaging
- Vol. 11 (25730436) , 1115-1127
Joint Translational Motion Compensation and High-Resolution ISAR Imaging Based on Sparse Bayesian Learning
Accurate translational motion compensation is crucial to high-resolution inverse synthetic aperture radar (ISAR) imaging, while complex observation environments generally render conventional translational motion compensation and imaging methods invalid. To tackle this issue, a novel method for joint translational motion compensation and high-resolution ISAR imaging is proposed within the Bayesian framework. Firstly, the statistical model is established, in which the gamma process-Laplace (GaPL) nonparametric prior is proposed to attain flexible sparse representation. Then, to address the issue of matrix inversion in conventional Bayesian inference, a 2D inverse-free Bayesian imaging method with the GaPL prior, namely 2D-IFGaPL, is proposed, which significantly improves the computational efficiency. After that, a method for joint range offsets and phase errors estimation is proposed to achieve accurate translational motion compensation. Finally, sparse Bayesian imaging and translational motion compensation are alternatively implemented until convergence to generate well-focused image. Experimental results under complex observation environments, e.g. low signal-to-noise ratio (SNR) and incomplete data, have demonstrated the effectiveness of the proposed method.
- 10 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Imaging
- Vol. 11 (25730436) , 1115-1127
Joint Translational Motion Compensation and High-Resolution ISAR Imaging Based on Sparse Bayesian Learning
Accurate translational motion compensation is crucial to high-resolution inverse synthetic aperture radar (ISAR) imaging, while complex observation environments generally render conventional translational motion compensation and imaging methods invalid. To tackle this issue, a novel method for joint translational motion compensation and high-resolution ISAR imaging is proposed within the Bayesian framework. Firstly, the statistical model is established, in which the gamma process-Laplace (GaPL) nonparametric prior is proposed to attain flexible sparse representation. Then, to address the issue of matrix inversion in conventional Bayesian inference, a 2D inverse-free Bayesian imaging method with the GaPL prior, namely 2D-IFGaPL, is proposed, which significantly improves the computational efficiency. After that, a method for joint range offsets and phase errors estimation is proposed to achieve accurate translational motion compensation. Finally, sparse Bayesian imaging and translational motion compensation are alternatively implemented until convergence to generate well-focused image. Experimental results under complex observation environments, e.g. low signal-to-noise ratio (SNR) and incomplete data, have demonstrated the effectiveness of the proposed method.
- 3 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-12
High-Resolution ISAR Imaging With SSFCS Based on Nonparametric Bayesian Learning and Genetic Algorithm
For inverse synthetic aperture radar (ISAR) adopting the sparse stepped frequency chirp signal (SSFCS), the target echoes are sparse in the fast time domain with unknown phase errors induced by translational motion, bringing great challenges to well-focused imaging. To tackle this issue, an efficient method for joint motion compensation and high-resolution imaging is proposed under low signal-to-noise ratio (SNR) scenario. Firstly, the signal model is constructed, which is then converted to a probabilistic model with the nonparametric Gamma Process-Complex Gaussian prior. Then, 2D-fast image reconstruction, i.e., 2D-FIR, is proposed for efficient image inference, which avoids matrix inversion by relaxing the lower bound and has high computational efficiency. Finally, a cost function is designed and genetic algorithm is utilized to jointly estimate the translational motion and the 2D image. Experimental results on simulated and measured data have verified the effectiveness of the proposed method.
- 3 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-12
High-Resolution ISAR Imaging With SSFCS Based on Nonparametric Bayesian Learning and Genetic Algorithm
For inverse synthetic aperture radar (ISAR) adopting the sparse stepped frequency chirp signal (SSFCS), the target echoes are sparse in the fast time domain with unknown phase errors induced by translational motion, bringing great challenges to well-focused imaging. To tackle this issue, an efficient method for joint motion compensation and high-resolution imaging is proposed under low signal-to-noise ratio (SNR) scenario. Firstly, the signal model is constructed, which is then converted to a probabilistic model with the nonparametric Gamma Process-Complex Gaussian prior. Then, 2D-fast image reconstruction, i.e., 2D-FIR, is proposed for efficient image inference, which avoids matrix inversion by relaxing the lower bound and has high computational efficiency. Finally, a cost function is designed and genetic algorithm is utilized to jointly estimate the translational motion and the 2D image. Experimental results on simulated and measured data have verified the effectiveness of the proposed method.
- 21 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-11
High-Resolution Radar Imaging of Off-Grid Maneuvering Targets Based on Parametric Sparse Bayesian Learning
In high-resolution radar imaging, the time-varying Doppler induced by maneuvering targets generally invalidates traditional methods. Although sparse signal reconstruction methods can be applied to achieve better performance, the off-grid problem embedded in these methods still prohibits well-focused imaging. In this article, we propose to perform high-resolution radar imaging of maneuvering targets with off-grid scattering centers in a Bayesian framework. First, the statistical model with a parametric dictionary is established, in which the Doppler frequencies of scattering centers are treated as unknown model parameters instead of being discretized into the grid. To be consistent with the physical characteristics, the posterior of the Doppler is approximated by the von Mises distribution. Then, the model parameters and rotation parameters are estimated iteratively by variational inference (VI) and the Newton method. Experiments have demonstrated that the proposed method provides an effective way for high-resolution and well-focused radar imaging of maneuvering targets with off-grid scattering centers in complex scenarios, such as incomplete data and low signal-to-noise ratio.
- 21 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-11
High-Resolution Radar Imaging of Off-Grid Maneuvering Targets Based on Parametric Sparse Bayesian Learning
In high-resolution radar imaging, the time-varying Doppler induced by maneuvering targets generally invalidates traditional methods. Although sparse signal reconstruction methods can be applied to achieve better performance, the off-grid problem embedded in these methods still prohibits well-focused imaging. In this article, we propose to perform high-resolution radar imaging of maneuvering targets with off-grid scattering centers in a Bayesian framework. First, the statistical model with a parametric dictionary is established, in which the Doppler frequencies of scattering centers are treated as unknown model parameters instead of being discretized into the grid. To be consistent with the physical characteristics, the posterior of the Doppler is approximated by the von Mises distribution. Then, the model parameters and rotation parameters are estimated iteratively by variational inference (VI) and the Newton method. Experiments have demonstrated that the proposed method provides an effective way for high-resolution and well-focused radar imaging of maneuvering targets with off-grid scattering centers in complex scenarios, such as incomplete data and low signal-to-noise ratio.
- Open Access
- 14 February 2022
- Published by MDPI AG in Remote Sensing
- Vol. 14 (4) , 901
High-Resolution ISAR Imaging Based on Plug-and-Play 2D ADMM-Net
We propose a deep learning architecture, dubbed Plug-and-play 2D ADMM-Net (PAN), by combining model-driven deep networks and data-driven deep networks for effective high-resolution 2D inverse synthetic aperture radar (ISAR) imaging with various signal-to-noise ratios (SNR) and incomplete data scenarios. First, a sparse observation model of 2D ISAR imaging is established, and a 2D ADMM algorithm is presented. On this basis, using the plug and play (PnP) technique, PnP 2D ADMM is proposed, by combining the 2D ADMM algorithm and the deep denoising network DnCNN. Then, we unroll and generalize the PnP 2D ADMM to the PAN architecture, in which all adjustable parameters in the reconstruction layers, denoising layers, and multiplier update layers are learned by end-to-end training through back-propagation. Experimental results showed that the PAN with a single parameter set can achieve noise-robust ISAR imaging with superior reconstruction performance on incomplete simulated and measured data under different SNRs.
- 14 February 2022
- Published by MDPI AG in Remote Sensing
- Vol. 14 (4) , 901
Open AccessHigh-Resolution ISAR Imaging Based on Plug-and-Play 2D ADMM-Net
We propose a deep learning architecture, dubbed Plug-and-play 2D ADMM-Net (PAN), by combining model-driven deep networks and data-driven deep networks for effective high-resolution 2D inverse synthetic aperture radar (ISAR) imaging with various signal-to-noise ratios (SNR) and incomplete data scenarios. First, a sparse observation model of 2D ISAR imaging is established, and a 2D ADMM algorithm is presented. On this basis, using the plug and play (PnP) technique, PnP 2D ADMM is proposed, by combining the 2D ADMM algorithm and the deep denoising network DnCNN. Then, we unroll and generalize the PnP 2D ADMM to the PAN architecture, in which all adjustable parameters in the reconstruction layers, denoising layers, and multiplier update layers are learned by end-to-end training through back-propagation. Experimental results showed that the PAN with a single parameter set can achieve noise-robust ISAR imaging with superior reconstruction performance on incomplete simulated and measured data under different SNRs.