EUVIMEDEUVIMED entdecken
Crossref

BLCNet: A Blind-Spot-Guided Local–Nonlocal Collaborative Self-Supervised Denoising Network for SAR Images

Jiajie Ma, Lijun Zhao, Lianzhi Huo, Chunning Meng, Zhiqing Zhang

IEEE Geoscience and Remote Sensing Letters · 2026 · 23 · 1-5

DOI: 10.1109/lgrs.2026.3672326 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Geoscience and Remote Sensing Letters, 2026. Verfasst von Jiajie Ma und 4 weiteren Autor:innen. Die Arbeit nennt 23 Literaturangaben. Gefördert von: Fundamental Research Funds for the Central Universities, Shenzhen Medical Research Fund, Civil Aerospace Technology Pre-Research Project of China’s 14th Five-Year Plan.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Synthetic aperture radar (SAR) images, due to their coherent imaging mechanism, are inevitably affected by severe speckle noise, which significantly limits subsequent target recognition and image interpretation. Existing self-supervised denoising methods struggle to balance the local continuity and nonlocal correlation characteristics of SAR images, thus constraining their denoising performance. To address this issue, this letter proposes a blind-spot-guided local–nonlocal collaborative denoising network (BLCNet), which achieves high-quality SAR image denoising without requiring clean images for supervision. The network employs a blind-spot masking mechanism to prevent direct reliance on target pixels. The local structure modeling (LSM) branch extracts contextual information from neighboring regions via dilated convolutions, while the nonlocal relation modeling (NRM) branch leverages a sparse Transformer to capture long-range yet semantically related dependencies. Furthermore, a texture enhancement module (TEM) is introduced to reinforce fine structural details through bidirectional pooling and an attention mechanism. Extensive experimental results demonstrate that BLCNet outperforms current state-of-the-art self-supervised denoising methods on both real and synthetic SAR image datasets.

Förderung

  • Fundamental Research Funds for the Central Universities
  • Shenzhen Medical Research Fund
  • Civil Aerospace Technology Pre-Research Project of China’s 14th Five-Year Plan
  • Open Project of Tianjin Key Laboratory of Retinal Functions and Diseases
  • 100 Young Talent Scholar Program of Nankai University

Zitieren

APA

Ma, J., Zhao, L., Huo, L., Meng, C., & Zhang, Z. (2026). BLCNet: A Blind-Spot-Guided Local–Nonlocal Collaborative Self-Supervised Denoising Network for SAR Images. IEEE Geoscience and Remote Sensing Letters, 23, 1-5. https://doi.org/10.1109/lgrs.2026.3672326

Vancouver

Ma J, Zhao L, Huo L, Meng C, Zhang Z. BLCNet: A Blind-Spot-Guided Local–Nonlocal Collaborative Self-Supervised Denoising Network for SAR Images. IEEE Geoscience and Remote Sensing Letters. 2026;23:1-5. doi:10.1109/lgrs.2026.3672326

BibTeX
@article{Ma2026,
  author = {Ma, Jiajie and Zhao, Lijun and Huo, Lianzhi and Meng, Chunning and Zhang, Zhiqing},
  title = {BLCNet: A Blind-Spot-Guided Local–Nonlocal Collaborative Self-Supervised Denoising Network for SAR Images},
  journal = {IEEE Geoscience and Remote Sensing Letters},
  year = {2026},
  volume = {23},
  pages = {1--5},
  doi = {10.1109/lgrs.2026.3672326}
}
Kostenlos starten →