AstroSmartphoneDataset, is a collection of astronomical images captured with smartphones with Google Camera (GCam), a Google mobile application allowing to take long-exposure astronomical images with a large field of view. GCam’s astrophotography mode captures and produces long-exposure frames to reduce noise, enhance details, improve dynamic range, while preserving non-astronomical elements like buildings and landscapes, ensuring a natural look.
Data have been collected between 2021 and 2025 from different locations and with 4 smartphones having different sensors and capabilities: Google Pixel 4a, Google Pixel 6, Google Pixel 8Pro, Google Pixel 8a. Data acquisition took place over a long period and required the use of often ingenious techniques to orientate and stabilise the phones during capture. Images were taken under different weather conditions (clear sky, foggy sky, almost overcast sky), in order to obtain an heterogeneous set.
AstroSmartphoneDataset contains 4635 medium-res and 197 high-res RGB images in JPEG format (minimal compression), and each RGB image has a minimal resolution of 1440× 1920 pixels: they were obtained after registration, stacking and post processing of one-second shots, giving final images corresponding to a cumulative exposure time of 4 minutes.
The dataset was used to train a YOLO models -- allowing to obtain the following results.
AI-powered Detection and Annotation of streaks in a cloudy view of Pegasus constellation, with a Samsung S25 Ultra
AI-powered detection in hyperlapse and widefield view, Samsung S25 Ultra (12/10/2025)
The dataset can be found here:
- [2025] AstroSmartphoneDataset: a collection of astronomical images captured with smartphones – DOI Link
The following paper describes the dataset:
- [2025] Toward a Public Dataset of Wide-Field Astronomical Images Captured with Smartphones – DOI Link
If you use this dataset in a publication, please notify us to include it in this list.
If you use AstroSmartphoneDataset in your work, please cite it as follows:
@conference{data25,
author={Olivier Parisot and Diogo Fernandes},
title={Toward a Public Dataset of Wide-Field Astronomical Images Captured with Smartphones},
booktitle={Proceedings of the 14th International Conference on Data Science, Technology and Applications - Volume 1: DATA},
year={2025},
pages={644-650},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013642300003967},
isbn={978-989-758-758-0},
}This dataset is released under the CC Attribution-NonCommercial-NoDerivatives 4.0 International.
See LICENSE for details.
For questions or collaborations, please contact Olivier Parisot at olivier.parisot@list.lu or open an issue on GitHub.
Copyright 2021-2025 Luxembourg Institute of Science and Technology (LIST - http://www.list.lu/).