Special Issue: Intelligent Processing and Analysis of Aerospace Remote Sensing Images

Diffusion super-resolution of SAR images integrating wavelet guidance and structural enhancement

  • Feiming WANG ,
  • Menglin LI ,
  • Yang QU ,
  • Bin PAN
Expand
  • School of Statistics and Data Science,Nankai University,Tianjin 300071,China

Received date: 2025-09-19

  Revised date: 2025-11-06

  Accepted date: 2025-11-28

  Online published: 2025-12-15

Supported by

National Key Research and Development Program of China(2022YFA1003800);National Natural Science Foundation of China(62571273);Natural Science Foundation of Tianjin(25JCLMJC01090)

Abstract

Influenced by scattering characteristics and imaging geometry, single-polarization SAR remote sensing images suffer from severe speckle noise and significant degradation, leading to texture loss and structural distortion in SAR image super-resolution tasks. To address this issue, this paper proposes a diffusion model-based super-resolution method that integrates frequency-domain processing and structural perception. The method adopts a latent diffusion model as its backbone and introduces a wavelet-guided module and a direction-aware enhancement module to improve performance. The wavelet-guided module performs multi-scale modulation of high-frequency sub-bands through wavelet decomposition and spatially adaptive normalization, thereby dynamically enhancing texture representation and high-frequency reconstruction capability according to the diffusion timestep. The direction-aware enhancement module incorporates multiple adaptive convolutional kernels, embedded with channel attention into the deep residual structures of the encoder to increase the sensitivity of compressed feature maps to structural information. In experiments, a realistic degradation model combining speckle noise and blur kernels is established to closely approximate the actual imaging pipeline. Results demonstrate that the proposed method significantly outperforms existing frameworks across multiple datasets, achieving an average improvement of 8.12% over the best baseline, verifying its effectiveness and advancement in SAR image super-resolution tasks.

Cite this article

Feiming WANG , Menglin LI , Yang QU , Bin PAN . Diffusion super-resolution of SAR images integrating wavelet guidance and structural enhancement[J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2026 , 47(10) : 532804 -532804 . DOI: 10.7527/S1000-6893.2025.32804

References

[1] ZHENG X, CUI H, XU C, et al. Dual Teacher: A semi-supervised co-training framework for cross-domain ship detection[J]. IEEE Transactions on Geoscience and Remote Sensing202361: 1-12.
[2] 连文慧. 模型驱动神经网络SAR成像特征重建[D]. 天津: 中国民航大学, 2024.
  LIAN W H. Model-driven neural network for SAR imagery feature reconstruction[D]. Tianjin: Civil Aviation University of China, 2024 (in Chinese).
[3] XIE J, GAO F, ZHOU X, et al. Wavelet-based bi-dimensional aggregation network for SAR image change detection[J]. IEEE Geoscience and Remote Sensing Letters202421: 1-5.
[4] 张冰玉, 潘志刚, 姚锴, 等. 基于生成对抗网络的SAR解压缩图像重建算法[J]. 中国科学院大学学报(中英文)202542(5): 666-676.
  ZHANG B Y, PAN Z G, YAO K, et al. SAR decompression image reconstruction algorithm based on generative adversarial network[J]. Journal of University of Chinese Academy of Sciences202542(5): 666-676 (in Chinese).
[5] LI X, DING M, GU Y, et al. An end-to-end framework for joint denoising and classification of hyperspectral images[J]. IEEE Transactions on Neural Networks and Learning Systems202334(7): 3269-3283.
[6] WANG L, ZHENG M, DU W, et al. Super-resolution SAR image reconstruction via generative adversarial network[C]∥12th International Symposium on Antennas, Propagation and EM Theory (ISAPE 2018). Piscataway: IEEE Press, 2018: 1-4.
[7] KANAKARAJ S, NAIR M S, KALADY S. Adaptive importance sampling unscented Kalman filter based SAR image super resolution[J]. Computers & Geosciences2019133: 104310.
[8] 刘涛, 钱锋, 张葆. 遥感图像的MAP超分辨重建[J]. 液晶与显示201833(10): 884-892.
  LIU T, QIAN F, ZHANG B. MAP super-resolution reconstruction of remote sensing image[J]. Chinese Journal of Liquid Crystals and Displays201833(10): 884-892 (in Chinese).
[9] 杨磊, 连文慧, 陈思佳, 等. 稀疏正则驱动的超分辨SAR图像重建[J]. 电讯技术202464(9):1361-1369.
  YANG L, LIAN W H, CHEN S J, et al. Super-resolution SAR image reconstruction based on sparse regularization driven mechanism[J]. Telecommunication Engineering202464(9): 1361-1369 (in Chinese).
[10] 史振威, 雷森. 图像超分辨重建算法综述[J]. 数据采集与处理202035(1): 1-20.
  SHI Z W, LEI S. Review of image super-resolution reconstruction[J]. Journal of Data Acquisition and Processing202035(1): 1-20 (in Chinese).
[11] LUO Z, YU J, LIU Z. The super-resolution reconstruction of SAR image based on the improved FSRCNN[J]. The Journal of Engineering2019, 2019(19): 5975-5978.
[12] 罗宇轩, 吴高昌, 高明. 自适应卷积和轻量化Transformer的遥感图像超分辨网络[J]. 计算机工程与应用202561(9): 263-276.
  LUO Y X, WU G C, GAO M. Remote sensing image super-resolution network with adaptive convolution and lightweight transformer[J]. Computer Engineering and Applications202561(9): 263-276 (in Chinese).
[13] DONG W, XU Y, QU J, et al. Learning multi-modal cross-scale deformable transformer network for unregistered hyperspectral image super-resolution[C]∥Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2024: 1573-1581.
[14] 李永军, 陈锦智敏, 李孟军, 等. 基于生成对抗网络的图像超分辨重建算法研究[J]. 河南大学学报 (自然科学版)202454(4): 436-442.
  LI Y J, CHEN J Z M, LI M J, et al. Research on super-resolution image reconstruction method based on generative adversarial networks[J]. Journal of Henan University (Natural Science)202454(4): 436-442 (in Chinese).
[15] 刘艳芳, 李春升, 杨威. 基于深度学习的SAR图像质量提升方法研究[J]. 上海航天(中英文)202239(3): 91-99.
  LIU Y F, LI C S, YANG W. Research on SAR image quality improvement based on deep learning[J]. Aerospace Shanghai (Chinese & English)202239(3): 91-99 (in Chinese).
[16] HO J, JAIN A, ABBEEL P. Denoising diffusion probabilistic models[C]∥Proceedings of the 34th International Conference on Neural Information Processing Systems (NeurIPS 2020). Red Hook: Curran Associates Inc., 2020: 6840-6851.
[17] 郭龙飞. 基于扩散模型的图像超分辨率综述[J]. 电视技术202549(7): 222-225.
  GUO L F. An overview of image super-resolution based on diffusion model[J]. Video Engineering202549(7): 222-225 (in Chinese).
[18] SAHARIA C, HO J, CHAN W, et al. Image super-resolution via iterative refinement[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence202345(4): 4713-4726.
[19] LI H, YANG Y, CHANG M, et al. SRDiff: Single image super-resolution with diffusion probabilistic models[J]. Neurocomputing2022479: 47-59.
[20] ROMBACH R, BLATTMANN A, LORENZ D, et al. High-resolution image synthesis with latent diffusion models[C]∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE Press, 2022: 10684-10695.
[21] XIAO Y, YUAN Q, JIANG K, et al. EDiffSR: An efficient diffusion probabilistic model for remote sensing image super-resolution[J]. IEEE Transactions on Geoscience and Remote Sensing202462: 1-14.
[22] 邹承权, 黄江成, 孙正宝, 等. 遥感图像超分辨率重建方法: 综述与实验[J]. 遥感技术与应用202540(4): 969-989.
  ZOU C Q, HUANG J C, SUN Z B, et al. Super-resolution reconstruction methods for remote sensing images: A review and experiments[J]. Remote Sensing Technology and Application202540(4): 969-989 (in Chinese).
[23] CHEN Z, ZHANG S, XIONG B. Super-resolving SAR images with diffusion models: A dual evaluation of metrics and applications[C]∥2024 IEEE 17th International Conference on Signal Processing (ICSP). Piscataway: IEEE Press, 2024: 340-346.
[24] HUANG S, ZENG H, CHEN H, et al. Spatial and cluster structural prior-guided subspace clustering for hyperspectral image[J]. IEEE Transactions on Geoscience and Remote Sensing202462: 1-15.
[25] QIAN Y, CAI Q, PAN Y, et al. Boosting diffusion models with moving average sampling in frequency domain[C]∥2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE Press, 2024: 8911-8920.
[26] PENG L, CAO Y, PEI R J, et al. Efficient real-world image super-resolution via adaptive directional gradient convolution[DB/OL]. arXiv preprint: 2405.07023, 2024.
[27] WANG J, YUE Z, ZHOU S, et al. Exploiting diffusion prior for real-world image super-resolution[J]. International Journal of Computer Vision2024132(12): 5929-5949.
[28] 王晓曼, 陈艳平, 杨采薇, 等. 多方向梯度特征提取的嵌套命名实体识别方法[J]. 计算机应用202545(11): 3547-3554.
  WANG X M, CHEN Y P, YANG C W, et al. Nested named entity recognition method for multi-directional gradient feature extraction[J]. Journal of Computer Applications202545(11): 3547-3554 (in Chinese).
[29] 袁姮, 霍欣燃, 姜文涛. 全息梯度差分卷积的图像分类网络[J]. 自动化学报202551(9): 2106-2130.
  YUAN H, HUO X R, JIANG W T. Image classification network of holographic gradient differential convolution[J]. Acta Automatica Sinica202551(9): 2106-2130 (in Chinese).
[30] 孙巍, 王乾宙, 陈雪凌, 等. 真实场景下图像超分辨技术现状与趋势[J]. 中国图象图形学报202530(6): 1576-1592.
  SUN W, WANG Q Z, CHEN X L, et al. Development of real-world image super-resolution[J]. Journal of Image and Graphics202530(6): 1576-1592 (in Chinese).
[31] 闫河, 巫茜. 劣质SAR图像退化模型研究[J]. 计算机工程与设计201031(24): 5351-5354.
  YAN H, WU Q. Research on low-quality SAR image degradation model[J]. Computer Engineering and Design201031(24): 5351-5354 (in Chinese).
[32] SONG J, MENG C, ERMON S. Denoising diffusion implicit models[DB/OL]. arXiv preprint2010.02502, 2022.
[33] REN B, MA S, HOU B, et al. A dual-stream high resolution network: Deep fusion of GF-2 and GF-3 data for land cover classification[J]. International Journal of Applied Earth Observation and Geoinformation2022112: 102896.
[34] ICEYE. ICEYE synthetic aperture radar (SAR) sample data[EB/OL]. [2025-07-15]. .
[35] LI X, ZHANG G, CUI H, et al. MCANet: A joint semantic segmentation framework of optical and SAR images for land use classification[J]. International Journal of Applied Earth Observation and Geoinformation2022106: 102638.
[36] LI Y, LI X, LI W, et al. SAR Det-100K: Towards open-source benchmark and tool kit for large-scale SAR object detection[C]∥The Thirty-eighth Annual Conference on Neural Information Processing Systems (NeurIPS). Red Hook: Curran Associates Inc., 2024.
[37] 吴樊, 张红, 王超, 等. SARBuD 1.0: 面向深度学习的GF-3精细模式SAR建筑数据集[J]. 遥感学报202226(4): 620-631.
  WU F, ZHANG H, WANG C, et al. SARBuD 1.0: A SAR building dataset based on GF-3 FSII imageries for built-up area extraction with deep learning method[J]. National Remote Sensing Bulletin202226(4): 620-631 (in Chinese).
[38] BLAU Y, MICHAELI T. The perception-distortion trade-off[C]∥2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2018: 6228-6237.
[39] ZHANG R, ISOLA P, EFROS A A, et al. The unreasonable effectiveness of deep features as a perceptual metric[C]∥2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2018: 586-595.
[40] HEUSEL M, RAMSAUER H, UNTERTHINER T, et al. GANs trained by a two time-scale update rule converge to a local Nash equilibrium[C]∥Advances in Neural Information Processing Systems. Red Hook: Curran Associates, Inc., 2017: 6626-6637.
[41] XU Z, FENG X, TIAN S, et al. Edge preserved low-rank SAR image despeckling via hierarchical prior knowledge regulation[J]. IEEE Transactions on Geoscience and Remote Sensing202361: 5205317.
[42] YANG J, WRIGHT J, HUANG T, et al. Image super-resolution as sparse representation of raw image patches [C]∥2008 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2008: 1-8.
[43] WANG X, XIE L, DONG C, et al. Real-ESRGAN: Training real-world blind super-resolution with pure synthetic data[C]∥2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). Piscataway: IEEE Press, 2021: 1905-1914.
[44] LIM B, SON S, KIM H, et al. Enhanced deep residual networks for single image super-resolution[DB/OL]. arXiv preprint: 1707.02921, 2017.
[45] LEPCHA D C, GOYAL B, DOGRA A, et al. Image super-resolution: A comprehensive review, recent trends, challenges and applications[J]. Information Fusion202391: 230-260.
Outlines

/