Diffusion-Driven Super-Resolution for Remote Sensing: Dual- Decoder Attention Models for High-Fidelity Earth Observation

Chibueze Favour Aririguzo; Gladys Chinyere Olumba; Daniel Agbonifo; Richard; Iherorochi Nneji; Grace Amarachi Megwa; Nzubechi Augustine Oriaku; Uchechi Joyce; Nneji; Edwin Sunday Umana

doi:10.56201/ijemt.vol.12.no1.2026.pg395.435

Diffusion-Driven Super-Resolution for Remote Sensing: Dual- Decoder Attention Models for High-Fidelity Earth Observation

Chibueze Favour Aririguzo, Gladys Chinyere Olumba, Daniel Agbonifo, Richard, Iherorochi Nneji, Grace Amarachi Megwa, Nzubechi Augustine Oriaku, Uchechi Joyce, Nneji, Edwin Sunday Umana

Abstract

Remote sensing images have played a significant role in various areas in recent years, such as urban infrastructure planning, disaster detection, and environmental detection. They provide helpful information about pollution levels, climate change, and farming conditions which helps to take timely actions that enhance efficiency and reduce waste. Also, farmers can rapidly evaluate estimable actions to enhance efficiency while reducing the wastage of resources. Nevertheless, atmospheric interference, sensor limitations, and spatial resolution gaps can often compromise the quality of these images. Advanced technologies like deep learning techniques can, fortunately, solve this problem. This study applies a diffusion model for denoising and uses a U-Net and Diffusion Transformer decoder to capture a high level of detail and project each step’s noise. Furthermore, the model has an attention mechanism incorporated to capture global information and longer dependencies of the image input. This approach greatly boosts the quality of the image while keeping the textual information intact. To enhance the restoration further, the use of various losses enables the model to focus on the details of the image. In addition to conventional loss functions such as the L1 and L2, frequency loss and edge loss are also used. Together, these loss functions manage to keep the image’s texture, details, structures, and pixel-level accuracy. To assess its performance, the model is trained on the OLI2MSI dataset. Accordingly, the evaluation metrics demonstrate that the model can effectively reconstruct low-resolution images into high-resolution ones. The model achieved an FID (Fréchet Inception Distance) score of 11.47945, a PSNR (Peak Signal-to-Noise Ratio) score of 40.28188, an SSIM (Structural Similarity Index) score of 0.88802, and an LPIPS (Learned Perceptual Image Patch Similarity) score of 0.06426. The results confirm that the model outperforms the state of art techniques, in terms of nu

Keywords

Remote Sensing ImagesSuper ResolutionDenoising DiffusionDeep LearningAttention Mechanism. 1 Introduction

References

Arefin, M. R., et al. (2020). Multi-image super-resolution for remote sensing using deep recurrent networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (pp. 816–825). https://doi.org/10.1109/CVPRW50498.2020.00111 Chen, X., Li, D., Liu, M., & Jia, J. (2023). CNN and transformer fusion for remote sensing image semantic segmentation. Remote Sensing, 15(18), 4455. https://doi.org/10.3390/rs15184455 Deng, H., Cheng, K., & Li, Y. (2024). PESiT: Progressive joint enhancement and blind super- resolution for low-light and low-resolution images under total variation constraints. IEEE Access, 12, 78354–78365. https://doi.org/10.1109/ACCESS.2024.3408248 Dong, C., Loy, C. C., He, K., & Tang, X. (2016). Image super-resolution using deep convolutional networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(2), 295–307. https://doi.org/10.1109/TPAMI.2015.2439281 Harb, M. M., & Dell’Acqua, F. (2017). Remote sensing in multirisk assessment: Improving disaster preparedness. IEEE Geoscience and Remote Sensing Magazine, 5(1), 53–65. https://doi.org/10.1109/MGRS.2016.2625100 Hou, Z., & Kung, S.-Y. (2020). Efficient image super resolution via channel discriminative deep neural network pruning. In Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (pp. 3647–3651). https://doi.org/10.1109/ICASSP40776.2020.9054019 Hu, T., et al. (2024). Global sparse attention network for remote sensing image super-resolution. Knowledge-Based Systems, 304, 112448. https://doi.org/10.1016/j.knosys.2024.112448 Li, R., Zheng, S., Duan, C., Wang, L., & Zhang, C. (2022). Land cover classification from remote sensing images based on multi-scale fully convolutional network. Geo-Spatial Information Science, 25(2), 278–294. https://doi.org/10.1080/10095020.2021.2017237 Li, X., Wang, Y., Monday, H. N., & Nneji, G. U. (2025). A novel residual learning of multi- scale feature extraction model for the classification of rice grain varieties. Computers and Electronics in Agriculture, 237, 110491. Liang, G., U, K., Yin, H., Liu, J., & Luo, H. (2023). Multi-scale hybrid attention graph convolution neural network for remote sensing image super-resolution. Signal Processing, 207, 108954. https://doi.org/10.1016/j.sigpro.2023.108954 Lin, C., Mao, X., Qiu, C., & Zou, L. (2024). DTCNet: Transformer-CNN distillation for super- resolution of remote sensing image. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 17, 11117–11133. https://doi.org/10.1109/JSTARS.2024.3409808 Monday, H. N., et al. (2022a). A wavelet convolutional capsule network with modified super- resolution GAN for fault diagnosis and classification. Complex & Intelligent Systems, 8(1), 1–15. https://doi.org/10.1007/s40747-022-00733-6 Monday, H. N., et al. (2022b). WMR-DepthwiseNet: A wavelet multi-resolution depthwise separable CNN for COVID-19 diagnosis. Diagnostics, 12(3), 765. https://doi.org/10.3390/diagnostics12030765 Monday, H. N., et al. (2022c). COVID-19 diagnosis from chest X-ray images using robust multi- resolution analysis Siamese neural network. Diagnostics, 12(3), 741. https://doi.org/10.3390/diagnostics12030741 Monday, H. N., et al. (2025a). Enhancing ECG classification in cardiac diagnostics using adaptive focal cross-entropy loss function. IEEE Journal of Biomedical and Health Informatics. Nneji, G. U., et al. (2022a). Fine-tuned Siamese network with modified enhanced SRGAN plus for COVID-19 identification. Diagnostics, 12(3). https://doi.org/10.3390/diagnostics12030717 Nneji, G. U., et al. (2022b). COVID-19 identification from low-quality CT using modified enhanced SRGAN plus and Siamese capsule network. Healthcare, 10(2), 403. https://doi.org/10.3390/healthcare10020403 Nneji, G. U., et al. (2022c). Identification of diabetic retinopathy using weighted fusion deep learning. Diagnostics, 12(2), 540. https://doi.org/10.3390/diagnostics12020540 Nneji, G. U., et al. (2022d). Multi-channel based image processing scheme for pneumonia identification. Diagnostics, 12(2), 325. https://doi.org/10.3390/diagnostics12020325 Nneji, G. U., et al. (2025). FFS-IML: Fusion-based statistical feature selection for machine learning-driven interpretability of chronic kidney disease. International Journal of Machine Learning and Cybernetics, 1–34. Qiu, Z., Shen, H., Yue, L., & Zheng, G. (2023). Cross-sensor remote sensing imagery super- resolution via an edge-guided attention-based network. ISPRS Journal of Photogrammetry and Remote Sensing, 199, 226–241. https://doi.org/10.1016/j.isprsjprs.2023.04.016 Tao, H. (2020). Super-resolution of remote sensing images based on a deep plug-and-play framework. In Proceedings of IGARSS 2020 (pp. 625–628). https://doi.org/10.1109/IGARSS39084.2020.9324647 Teixeira, I., Morais, R., Sousa, J. J., & Cunha, A. (2023). Deep learning models for the classification of crops in aerial imagery: A review. Agriculture, 13(5), 965. https://doi.org/10.3390/agriculture13050965 Wang, C., et al. (2024). MSWAGAN: Multispectral remote sensing image super-resolution based on multiscale window attention transformer. IEEE Transactions on Geoscience and Remote Sensing, 62, 1–15. https://doi.org/10.1109/TGRS.2024.3385752 Wang, J., Wang, B., Wang, X., Zhao, Y., & Long, T. (2023a). Hybrid attention-based U-shaped network for remote sensing image super-resolution. IEEE Transactions on Geoscience and Remote Sensing, 61, 1–15. https://doi.org/10.1109/TGRS.2023.3283769 Wang, J., Chen, Z., Ma, A., & Zhong, Y. (2022). Capformer: Pure transformer for remote sensing image caption. In Proceedings of IGARSS 2022 (pp. 7996–7999). https://doi.org/10.1109/IGARSS46834.2022.9883199 Wang, P., & Sertel, E. (2023). Multi-frame super-resolution of remote sensing images using attention-based GAN models. Knowledge-Based Systems, 266, 110387. https://doi.org/10.1016/j.knosys.2023.110387 Wang, P., Zhang, H., Zhou, F., & Jiang, Z. (2019). Unsupervised remote sensing image super- resolution using cycle CNN. In Proceedings of IGARSS 2019 (pp. 3117–3120). https://doi.org/10.1109/IGARSS.2019.8898648 Wang, Y., Shao, Z., Lu, T., Wu, C., & Wang, J. (2023b). Remote sensing image super-resolution via multiscale enhancement network. IEEE Geoscience and Remote Sensing Letters, 20, 1–5. https://doi.org/10.1109/LGRS.2023.3248069 Zhang, S., Shao, Z., Huang, X., Bai, L., & Wang, J. (2021). An internal-external optimized CNN for arbitrary oriented object detection from optical remote sensing images. Geo-Spatial Information Science, 24(4), 654–665. https://doi.org/10.1080/10095020.2021.1972772 Zheng, Y., et al. (2024). A heterogeneous remote sensing image matching method for urban areas with complex terrain based on 3D spatial relationship constraints. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 17, 6791–6804. https://doi.org/10.1109/JSTARS.2024.3374327

Diffusion-Driven Super-Resolution for Remote Sensing: Dual- Decoder Attention Models for High-Fidelity Earth Observation

Abstract

Keywords

References

More Articles from INTERNATIONAL JOURNAL OF ENGINEERING AND MODERN TECHNOLOGY