# HLS-GPT: A Generative Pretrained Transformer (GPT) for Continental-Scale NASA Harmonized Landsat and Sentinel-2 (HLS) Reflectance Reconstruction Across All Bands on Arbitrary Dates

arXiv: https://arxiv.org/abs/2606.18115 · HTML: https://mlsanity.com/paper/2606.18115/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 15 of its 115 references that are in the map.
- First posted on arXiv 16 Jun 2026.

## Builds on (most cited first, 15 of 15)

- [SatMAE: Pre-training Transformers for Temporal and Multi-Spectral Satellite Imagery](https://mlsanity.com/paper/2207.08051/) (2022)
- [Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning](https://mlsanity.com/paper/2212.14532/) (2022)
- [CROMA: Remote Sensing Representations with Contrastive Radar-Optical Masked Autoencoders](https://mlsanity.com/paper/2311.00566/) (2023)
- [Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications](https://mlsanity.com/paper/2412.02732/) (2024)
- [Towards a Unified Copernicus Foundation Model for Earth Vision](https://mlsanity.com/paper/2503.11849/) (2025)
- [EarthNet2021: A large-scale dataset and challenge for Earth surface forecasting as a guided video prediction task](https://mlsanity.com/paper/2104.10066/) (2021)
- [GLF-CR: SAR-Enhanced Cloud Removal with Global-Local Fusion](https://mlsanity.com/paper/2206.02850/) (2022)
- [Multi-modal learning for geospatial vegetation forecasting](https://mlsanity.com/paper/2303.16198/) (2023)
- [Seeing Through Clouds in Satellite Images](https://mlsanity.com/paper/2106.08408/) (2021)
- [Cloud removal in remote sensing images using generative adversarial networks and SAR-to-optical image translation](https://mlsanity.com/paper/2012.12180/) (2020)
- [AgriFM: A Multi-source Temporal Remote Sensing Foundation Model for Agriculture Mapping](https://mlsanity.com/paper/2505.21357/) (2025)
- [U-TILISE: A Sequence-to-sequence Model for Cloud Removal in Optical Satellite Time Series](https://mlsanity.com/paper/2305.13277/) (2023)
- [Long time-series NDVI reconstruction in cloud-prone regions via spatio-temporal tensor completion](https://mlsanity.com/paper/2102.02603/) (2021)
- [Paving the way toward foundation models for irregular and unaligned Satellite Image Time Series](https://mlsanity.com/paper/2407.08448/) (2024)
- [OBSUM: An object-based spatial unmixing model for spatiotemporal fusion of remote sensing images](https://mlsanity.com/paper/2310.09517/) (2023)

## Abstract (arXiv)

Recent deep learning methods for Landsat and Sentinel-2 reflectance time series reconstruction remain limited by restricted spectral coverage, limited geographic scalability, or patch-based designs with short temporal contexts. We present HLS-GPT, a large-scale generative pretrained Transformer model for reconstructing NASA Harmonized Landsat Sentinel-2 30 m surface reflectance for all bands, any date, and any pixel location. HLS-GPT uses a hierarchical Transformer architecture to handle the different spectral band configurations of Landsat and Sentinel-2 and operates on single-pixel 12-month time series. To capture geographic and seasonal variability, the model was trained with nine years of HLS time series from more than 0.25 million training pixels across the conterminous United States. A random cropping and masking strategy extracts 12-month periods with varying start dates across epochs, masks 50% of valid observations, and trains the model to reconstruct the masked reflectance values from the remaining observations. Evaluation using more than 62,000 independent test pixels shows robust reconstruction under diverse land surface conditions, including complex crop phenology and sparse, irregular observations. Leave-one-observation-out evaluation achieved reconstruction RMSE below 0.026 for all HLS spectral bands, with relative RMSE below 35% for visible bands and below 13% for other bands. Red-edge band errors were comparable to red and near-infrared errors despite the absence of red-edge bands on Landsat. Sensitivity analyses that randomly masked 10% to 90% of test observations showed only modest degradation when 10% to 50% of observations were masked, with all-band RMSE below 0.028. Image reconstruction over nine independent 109 by 109 km CONUS HLS tiles further demonstrates that HLS-GPT outperforms two conventional methods and the NASA-IBM Prithvi model.