There is a newer version of the record available.

Published October 6, 2023 | Version v2

SpectralGPT: The first remote sensing foundation model customized for spectral data

  • 1. Aerospace Information Research Institute, Chinese Academy of Sciences
  • 2. Southeast University
  • 3. The University of Tokyo
  • 4. Technical University of Munich
  • 5. ROR icon Helmholtz-Zentrum Dresden-Rossendorf
  • 6. University of New South Wales
  • 7. University of Extremadura
  • 8. University of Pavia
  • 9. University of Iceland
  • 10. Univ. Grenoble Alpes

Description

SpectralGPT is the first purpose-built foundation model designed explicitly for spectral RS data. It considers unique characteristics of spectral data, i.e., spatial-spectral coupling and spectral sequentiality, in the MAE framework with a simple yet effective 3D GPT network.

We will gradually release the trained models (SpectralGPT, SpectralGPT+), the new benchmark dataset (SegMunich) for the downstream task of semantic segmentation, original code, and implementation instructions.

Files

SegMunich.zip

Files (3.8 GB)

Name Size
md5:59aeb5d227a320ec148900a4bf7833e7
1.4 GB Preview Download
md5:aeaa08388f0937e3f07ecbd7757f20dc
1.2 GB Download
md5:0464b11758bc309b02ee6616b58f0bed
1.2 GB Download