Beber et al. (2025) Super Resolution of Satellite-Based Land Surface Temperature Through Airborne Thermal Imaging

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2025
• Date: 2025-11-19
• Authors: Raniero Beber, Salim Malek, Fabio Remondino
• DOI: 10.3390/rs17223766

Research Groups

Not explicitly mentioned in the provided text.

Short Summary

This study introduces the Dilated Spatio-Temporal U-Net (DST-UNet), a novel deep learning approach designed to bridge the resolution gap between low-resolution satellite thermal imagery and high-resolution optical data, enabling the generation of detailed, high-frequency urban thermal maps.

Objective

• To develop and validate a deep learning model (DST-UNet) capable of generating high-resolution, airborne-like urban thermal maps from available low-resolution satellite imagery and ancillary data, thereby addressing the existing resolution gap in urban heat island mapping.

Study Configuration

• Spatial Scale: Urban environments, focusing on generating high-resolution (airborne-like) thermal maps from low-resolution satellite imagery, addressing multiscale urban thermal patterns.
• Temporal Scale: Continuous monitoring, enabling the generation of detailed thermal maps with a frequency significantly exceeding that of traditional airborne campaigns.

Methodology and Data

• Models used: Dilated Spatio-Temporal U-Net (DST-UNet), a deep learning model based on a modified U-Net architecture incorporating dilated convolutions.
• Data sources: Low-resolution satellite imagery (e.g., from Landsat missions) and ancillary data.

Main Results

• The DST-UNet model effectively generalizes across different urban environments.
• It enables municipalities to generate detailed, high-resolution urban thermal maps at a frequency far exceeding traditional airborne campaigns.
• The framework leverages open-source data, providing a cost-effective and scalable solution for continuous, high-resolution urban thermal monitoring.

Contributions

• Introduces a novel deep learning architecture (DST-UNet) specifically designed to bridge the resolution gap between low-resolution satellite thermal data and high-resolution optical data for urban heat island mapping.
• Offers a cost-effective and scalable solution for continuous, high-resolution urban thermal monitoring by leveraging open-source satellite data.
• Empowers more effective climate resilience and public health initiatives through enhanced urban thermal mapping capabilities.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Beber2025Super,
  author = {Beber, Raniero and Malek, Salim and Remondino, Fabio},
  title = {Super Resolution of Satellite-Based Land Surface Temperature Through Airborne Thermal Imaging},
  journal = {Remote Sensing},
  year = {2025},
  doi = {10.3390/rs17223766},
  url = {https://doi.org/10.3390/rs17223766}
}