# Conservation Drones BIRDSAI (aerial thermal infrared)

## Facts

- Provider: Elizabeth Bondi et al. (BIRDSAI)
- License: CDLA-Permissive-1.0 (https://lila.science/datasets/conservationdrones)
- Format: CSV (in sampled files)
- Stored at: Azure Blob Storage, https://lilawildlife.blob.core.windows.net/lila-wildlife/conservationdrones
- Access: public, readable without an account
- Category: Environmental Science
- Source: https://lila.science/datasets/conservationdrones
- Page: https://scigantic.com/archive/e55f120e-edce-5386-86a9-9c7328442b58
- Keywords: Environmental Science, CSV, Azure Blob Storage

## About

BIRDSAI (Benchmarking IR Dataset for Surveillance with Aerial Intelligence) is a long-wave thermal infrared (TIR) dataset of night-time aerial video of animals and humans in Southern Africa, built for anti-poaching detection and tracking. LILA states 48 real TIR videos (62k images) and 124 synthetic AirSim videos (100k images), with per-object tracks labelled animal or human, species where possible (elephant, lion, giraffe and others) and occlusion and noise flags. Annotations follow the MOT CSV format (frame, object id, x, y, w, h, class, species, occlusion, noise); read with pandas. Layout (measured 2026-11-13): unzipped folders under v01/ (TrainReal/, TrainSimulation/, conservationdrones-testset/), each with annotations and JPEG frame folders, alongside zips of the same data (train_real 2.3 GB, train_simulation 42.2 GB, test_real 1.8 GB) and a top-level test-images.zip. The unzipped frames read directly from the read-only mount. Used for the ICVGIP 2020 Visual Data Challenge; helper code at github.com/exb7900/BIRDSAI. Released by LILA BC under the Community Data License Agreement (permissive variant), CDLA-Permissive-1.0.

## Load the data

```python
import adlfs
fs = adlfs.AzureBlobFileSystem(account_name='lilawildlife', anon=True)
fs.ls('lila-wildlife/conservationdrones/')[:10]
```

Lists the files with anonymous access. Runs on any machine, with no cloud account.

## Data structure

File types in sampled files: CSV

Example file: `v01/TrainReal/annotations/0000000010_0000000000.csv`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/e55f120e-edce-5386-86a9-9c7328442b58
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/e55f120e-edce-5386-86a9-9c7328442b58
- Over MCP: `get_schema_card` with id `e55f120e-edce-5386-86a9-9c7328442b58` at https://api.scigantic.com/mcp

## Related topics

- [CSV datasets](https://scigantic.com/topics/csv)
- [Environmental Science datasets](https://scigantic.com/topics/environmental-science)
