# Sweet Corals: Coral Reef 3D Photogrammetry Surveys (Indonesia)

## Facts

- Provider: wildflow.ai and University of Derby (Michael Sweet)
- License: CC-BY-4.0 (https://huggingface.co/datasets/wildflow/sweet-corals)
- Size: 96,952 files
- Format: JPEG (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/wildflow/sweet-corals
- Access: public, readable without an account
- Category: Environmental Science
- Source: https://wildflow.ai/protocol
- Page: https://scigantic.com/archive/bc96d7a3-e5c3-569c-8818-c4c33f4d6b02
- Keywords: Environmental Science, image segmentation, JPEG, Hugging Face

## About

Underwater photogrammetry surveys of coral reefs in Indonesia, collected by wildflow.ai with the University of Derby following a published 3D mapping protocol. The release currently provides raw GoPro images (the card counts 90,289 images, about 352 GB) and some colour-corrected images, with camera poses, point clouds, meshes, orthomosaics, annotations and Gaussian splatting models announced for later. Survey folders follow a {country}_{site}_{patch}_{date} pattern: Tabuhan, Pemuteran and Watudodol reef patches from February 2025 and a Banyuwangi coral nursery farm whose tables each hold a single coral species, collected for species classification rather than reconstruction. Each folder holds raw/ images and a low-resolution orthomosaic; all sites except the farm are georeferenced and scaled. The repository uses about 340 GB (measured 2026-11-14). Intended uses are coral species classification, 3D reconstruction and semantic segmentation of reefs, disease detection and reef monitoring for conservation. Images are JPEG; process them with COLMAP, Metashape or other structure-from-motion tools, or load them with PIL and Hugging Face datasets. Licensed CC BY 4.0 with attribution to wildflow.

## Load the data

```python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="wildflow/sweet-corals", filename="_indonesia_tabuhan_p1_20250210/colmap/cameras.bin", repo_type="dataset")
print(path)
```

## Data structure

File types in sampled files: JPEG

Example file: `_indonesia_tabuhan_p1_20250210/colmap/cameras.bin`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/bc96d7a3-e5c3-569c-8818-c4c33f4d6b02
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/bc96d7a3-e5c3-569c-8818-c4c33f4d6b02
- Over MCP: `get_schema_card` with id `bc96d7a3-e5c3-569c-8818-c4c33f4d6b02` at https://api.scigantic.com/mcp

## Related topics

- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Environmental Science datasets](https://scigantic.com/topics/environmental-science)
- [JPEG datasets](https://scigantic.com/topics/jpeg)
