Self-Supervised Super-Resolution for Sentinel-5P Hyperspectral Images

2026-04-19

AI Summary

This work addresses the challenge of enhancing the spatial resolution of Sentinel-5P satellite imagery, which is inherently limited by its coarse resolution and further hindered by the absence of real high-resolution ground truth for training. To overcome the reliance on synthetic data in existing super-resolution methods, the authors propose a self-supervised framework that uniquely integrates Stein’s Unbiased Risk Estimator (SURE) with equivariant imaging constraints. By explicitly modeling the sensor-specific degradation process and incorporating signal-to-noise ratio metadata, the method operates without access to true high-resolution labels. It employs a depthwise-separable convolutional U-Net architecture designed for spectral fidelity and computational efficiency. Evaluated on both synthetic and real-world data, the approach achieves performance comparable to supervised methods, substantially outperforms bicubic interpolation in spatial detail, and demonstrates physically plausible reconstructions validated against EMIT hyperspectral observations.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Sentinel-5P (S5P) plays a critical role in atmospheric monitoring; however, its spatial resolution limits fine-scale analysis. Existing super-resolution (SR) approaches rely on supervised learning with synthetic low-resolution (LR) data, since true high-resolution (HR) data do not exist, limiting their applicability to real observations. We propose a self-supervised hyperspectral SR framework for S5P that enables training without HR ground truth. The method combines Stein's Unbiased Risk Estimator (SURE) with an equivariant imaging constraint, incorporating the S5P degradation operator and noise statistics derived from signal-to-noise ratio (SNR) metadata. We also introduce depthwise separable convolution U-Net architectures designed for efficiency and spectral fidelity. The framework is evaluated in two settings: (i) LR-HR, where synthetic LR data are used for direct comparison with supervised learning, and (ii) GT-SHR, where super-resolved images surpass the native spatial resolution without HR reference. Results across multiple bands show that self-supervised models achieve performance comparable to supervised methods while maintaining strong consistency. Qualitative analysis shows improved spatial detail over bicubic interpolation, and validation with EMIT data confirms that reconstructed structures are physically meaningful. Code is available at https://github.com/hyamomar/Sentinel-5P-Super-Resolution/tree/main/self_supervised

Problem

• Super-Resolution
• Sentinel-5P
• Hyperspectral Images
• Self-Supervised Learning
• Spatial Resolution

Innovation

• self-supervised learning
• hyperspectral super-resolution
• Stein's Unbiased Risk Estimator
• equivariant imaging
• depthwise separable convolution