Yunpeng Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yunpeng Zhang has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Mengdao Xing
Xidian University
3 shared publications - Jinsong Zhang
School of Information Mechanics and Sensing Engineering, Xidian University, Xi’an 710071, China
2 shared publications Dan Xu
Academy of Advanced Interdisciplinary Research, Xidian University, Xian, China
1 shared publication- Giampaolo Ferraioli
Dipartimento di Scienze e Tecnologie, Universit degli Studi di Napoli Parthenope Naples, Italy
1 shared publication - Guang-Cai Sun
National Key Laboratory of Radar Signal Processing, Xidian University, Xi’an 710071, China
1 shared publication - Jinsong Zhang
Academy of Advanced Interdisciplinary Research, Xidian University, Xian, China
1 shared publication
Latest Publications
- Open Access
- 17 November 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (22) , 3943
Task-Oriented Evaluation of PolSAR Despeckling for Downstream Applications
Polarimetric SAR (PolSAR) records multiple polarization channels to characterize target scattering, but its coherent imaging mechanism produces multiplicative speckle that complicates downstream analysis. Existing despeckling methods pursue different objectives, so visual inspection and conventional image-quality metrics may not reflect their practical utility. In this work, six representative conventional and deep-learning-based despeckling methods and the unfiltered data are evaluated through three downstream tasks: land-cover classification on three scenes, ice-type classification on a GaoFen-3 scene, and CFAR-based ship detection on offshore images of the FPSD dataset. Each task depends on a different property of the filtered covariance data, so the same despeckled product can suit one task and not another: NL-SAR gives the highest land-cover accuracy on two scenes but ranks sixth of seven inputs in ship detection, whereas SDNLM attains the highest calibrated detection F1 but ranks sixth of seven on the Island scene. This task dependence persists under targeted sensitivity analyses of smoothing strength, classifier choice, covariance validity, and detector calibration. Despeckling methods should therefore be evaluated against the information that the intended downstream task requires and selected on that basis, rather than by visual quality or a single task-independent criterion.
- 17 November 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (22) , 3943
Open AccessTask-Oriented Evaluation of PolSAR Despeckling for Downstream Applications
Polarimetric SAR (PolSAR) records multiple polarization channels to characterize target scattering, but its coherent imaging mechanism produces multiplicative speckle that complicates downstream analysis. Existing despeckling methods pursue different objectives, so visual inspection and conventional image-quality metrics may not reflect their practical utility. In this work, six representative conventional and deep-learning-based despeckling methods and the unfiltered data are evaluated through three downstream tasks: land-cover classification on three scenes, ice-type classification on a GaoFen-3 scene, and CFAR-based ship detection on offshore images of the FPSD dataset. Each task depends on a different property of the filtered covariance data, so the same despeckled product can suit one task and not another: NL-SAR gives the highest land-cover accuracy on two scenes but ranks sixth of seven inputs in ship detection, whereas SDNLM attains the highest calibrated detection F1 but ranks sixth of seven on the Island scene. This task dependence persists under targeted sensitivity analyses of smoothing strength, classifier choice, covariance validity, and detector calibration. Despeckling methods should therefore be evaluated against the information that the intended downstream task requires and selected on that basis, rather than by visual quality or a single task-independent criterion.
- Open Access
- 30 April 2025
- Published by MDPI AG in Remote Sensing
- Vol. 17 (9) , 1586
SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
Synthetic aperture radar (SAR) recognition systems often need to collect new data and update the network accordingly. However, the network faces the challenge of catastrophic forgetting, where previously learned knowledge might be lost during the incremental learning of new data. To improve the applicability and sustainability of SAR target classification methods, we propose a multi-stage regularization-based class-incremental learning (CIL) method for SAR targets, called SCF-CIL, which addresses catastrophic forgetting. This method offers three main contributions. First, for the feature extractor, we fuse the convolutional neural network features with the scattering center features using a cross-attention feature fusion structure, ensuring both the plasticity and stability of the extracted features. Next, an overfitting training strategy is applied to provide clustering space for unseen classes with an acceptable trade-off in the accuracy of the current classes. Finally, we analyze the influence of training with imbalanced data on the last fully connected layer and introduce a multi-stage regularization method by dividing the calculation of the fully connected layer into three parts and applying regularization to each. Our experiments on SAR datasets demonstrate the effectiveness of these improvements.
- 30 April 2025
- Published by MDPI AG in Remote Sensing
- Vol. 17 (9) , 1586
Open AccessSCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
Synthetic aperture radar (SAR) recognition systems often need to collect new data and update the network accordingly. However, the network faces the challenge of catastrophic forgetting, where previously learned knowledge might be lost during the incremental learning of new data. To improve the applicability and sustainability of SAR target classification methods, we propose a multi-stage regularization-based class-incremental learning (CIL) method for SAR targets, called SCF-CIL, which addresses catastrophic forgetting. This method offers three main contributions. First, for the feature extractor, we fuse the convolutional neural network features with the scattering center features using a cross-attention feature fusion structure, ensuring both the plasticity and stability of the extracted features. Next, an overfitting training strategy is applied to provide clustering space for unseen classes with an acceptable trade-off in the accuracy of the current classes. Finally, we analyze the influence of training with imbalanced data on the last fully connected layer and introduce a multi-stage regularization method by dividing the calculation of the fully connected layer into three parts and applying regularization to each. Our experiments on SAR datasets demonstrate the effectiveness of these improvements.
- 3 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 20 (1545598X) , 1-5
Robust Multi-Ship Tracker in SAR Imagery by Fusing Feature Matching and Modified KCF
In previous research, most Multi-object tracking (MOT) algorithms focus on the optical image dataset, while the Synthetic Aperture Radar (SAR) image dataset faces the characteristics of few prior samples, high false alarm rate, and various defocusing interference. On the SAR image dataset, a robust MOT algorithm is proposed to fulfill multi-ship tracking in complex imaging conditions. First, the kernelized correlation filters (KCF) algorithm, a single-object tracking algorithm, is modified and applied to reduce the impact of false alarms on tracking performance. After that, different matching strategies are adaptively adapted to associate the targets based on the three intersection patterns between the predictions and the detections, which can reduce the impact of the deviated detections. Finally, the tracker’s time limit with Gaussian distribution is proposed to improve the re-association ability after the tracking interruption caused by the defocusing. The experiment results demonstrate the robust tracking ability of the proposed MOT algorithm.
- 3 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 20 (1545598X) , 1-5
Robust Multi-Ship Tracker in SAR Imagery by Fusing Feature Matching and Modified KCF
In previous research, most Multi-object tracking (MOT) algorithms focus on the optical image dataset, while the Synthetic Aperture Radar (SAR) image dataset faces the characteristics of few prior samples, high false alarm rate, and various defocusing interference. On the SAR image dataset, a robust MOT algorithm is proposed to fulfill multi-ship tracking in complex imaging conditions. First, the kernelized correlation filters (KCF) algorithm, a single-object tracking algorithm, is modified and applied to reduce the impact of false alarms on tracking performance. After that, different matching strategies are adaptively adapted to associate the targets based on the three intersection patterns between the predictions and the detections, which can reduce the impact of the deviated detections. Finally, the tracker’s time limit with Gaussian distribution is proposed to improve the re-association ability after the tracking interruption caused by the defocusing. The experiment results demonstrate the robust tracking ability of the proposed MOT algorithm.