README.md

---
license: cc-by-4.0
task_categories:
- robotics
- feature-extraction
language:
- en
tags:
- Space
- Structure-from-Motion
- SfM
- SLAM
- Asteroid
- Stereophotoclinometry
pretty_name: Photoclinometry-from-Motion (PhoMo)
size_categories:
- n<1K
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: image
    dtype: image
  - name: site
    dtype: string
  - name: body
    dtype: string
  - name: fits_path
    dtype: string
  - name: npy_path
    dtype: string
  - name: id
    dtype: string
  splits:
  - name: train
    num_bytes: 92236737
    num_examples: 99
  download_size: 92237643
  dataset_size: 92236737
---

<div align="center">
<img src="assets/phomo.png" alt="logo" width="400">
<h1>Photoclinometry-from-Motion (PhoMo)</h1>

<a href="https://huggingface.co/datasets/travisdriver/phomo-data"><img src="https://img.shields.io/badge/🤗-Hugging%20Face-yellow.svg" alt="HuggingFace"></a>
<a href="https://arxiv.org/abs/2504.08252"><img src="https://img.shields.io/badge/arXiv-2504.08252-b31b1b" alt="arXiv"></a>

[Travis Driver](https://travisdriver.github.io/), [Andrew Vaughan](https://www.linkedin.com/in/andrewtvaughan/), [Yang Cheng](https://www-robotics.jpl.nasa.gov/who-we-are/people/yang_cheng/), [Adnan Ansar](https://www-robotics.jpl.nasa.gov/who-we-are/people/adnan_ansar/), [John Christian](https://ae.gatech.edu/directory/person/john-christian), [Panagiotis Tsiotras](https://ae.gatech.edu/directory/person/panagiotis-tsiotras)
</div>

#### This is the official repository for [Stereophotoclinometry Revisited](https://arxiv.org/abs/2504.08252), which is currently under review for publication to AIAA's [Journal of Guidance, Control, and Dynamics (JGCD)](https://arc.aiaa.org/loi/jgcd)

**Photoclinometry-from-Motion (PhoMo)** is a framework for _autonomous_ image-based surface reconstruction and characterization of small celestial bodies. PhoMo integrates photoclinometry into a structure-from-motion (SfM) pipeline that leverages deep learning-based keypoint extraction and matching (i.e., [RoMa](https://github.com/Parskatt/RoMa)) to enable _simultaneous_ optimization of the spacecraft pose, landmark positions, Sun vectors, and surface normals and albedos.

If you find our datasets or results useful for your research, please use the following citation:

```bibtex
@article{driver2025phomo,
  title={Stereophotoclinometry Revisited},
  author={Driver, Travis and Vaughan, Andrew and Cheng, Yang, and Ansar, Adnan and Christian, John and Tsiotras, Panagiotis},
  journal={arXiv:2504.08252},
  year={2025},
  pages={1--45}
}
```

<!---

## Dataset Structure

Each directory contains a subdirectory for each of the sites used in the paper, i.e., Cornelia (`cornelia/`), Ahuna Mons (`ahunamons/`), and Ikapati (`ikapati/`).

`images/`: Contains the input images for each site (i.e., Cornelia, Ahuna Mons, and Ikapati) in multiple formats.

- `*_calib.npy`: Radiometrically calibrated to units of reflectance (L/F).
- `*_uncalib.npy`:
- `*.png`:
--->