JSON format icon JSON RDF format icon RDF ISO19115-3 XML

Lightweight Neural Network for Spatiotemporal Filling of Data Gaps in Sea Surface Temperature Images

ARCHIVED

Created 20/01/2025

Updated 20/01/2025

JSON format icon JSON RDF format icon RDF ISO19115-3 XML

Cloud coverage remains a key issue for researchers working with satellite data. Accurate reconstruction of measurements obstructed by cloud can enhance the usefulness of satellite databases for identifying trends and changes in various environments. In this work, we develop, train and test a bidirectional long short-term memory (BiLSTM) model with a custom temporal penalty layer for filling gaps in sea surface temperature (SST) images acquired by the Himawari- 8 satellite. The proposed model showed strong performance, achieving a per-image MAE of 0.1193◦C and per-image RMSE of 0.0985◦C. Our model is also shown to outperform previous state-of-the-art literature. Overall, this work shows that our BiLSTM algorithm is an effective tool for gapfilling cloud-affected SST data. Citation: S. Baker, Z. Huang and B. Philippa, "Lightweight Neural Network for Spatiotemporal Filling of Data Gaps in Sea Surface Temperature Images," in IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1-10, 2023, Art no. 4204310, doi: 10.1109/TGRS.2023.3273575

Files and APIs

Tags

Additional Info

Field Value
Title Lightweight Neural Network for Spatiotemporal Filling of Data Gaps in Sea Surface Temperature Images
Language eng
Licence Not Specified
Landing Page https://devweb.dga.links.com.au/data/dataset/7287c291-3e8c-4d70-a5fb-196550c90ea9
Contact Point
Geoscience Australia
clientservices@ga.gov.au
Reference Period 08/04/2019
Geospatial Coverage
Map data © OpenStreetMap contributors
{
  "coordinates": [
    [
      [
        112.0,
        -44.0
      ],
      [
        154.0,
        -44.0
      ],
      [
        154.0,
        -9.0
      ],
      [
        112.0,
        -9.0
      ],
      [
        112.0,
        -44.0
      ]
    ]
  ],
  "type": "Polygon"
}
Data Portal Data.gov.au