# New Zealand Wildlife Thermal Imaging (Cacophony Project)

## Facts

- Provider: The Cacophony Project
- License: CDLA-Permissive-1.0 (https://lila.science/datasets/new-zealand-wildlife-thermal-imaging/)
- Format: HDF5 (in sampled files)
- Stored at: Azure Blob Storage, https://lilawildlife.blob.core.windows.net/lila-wildlife/nz-thermal
- Access: public, readable without an account
- Category: Environmental Science
- Source: https://lila.science/datasets/new-zealand-wildlife-thermal-imaging/
- Page: https://scigantic.com/archive/2ab2661a-07b9-5955-9b63-f9915afa085f
- Keywords: Environmental Science, HDF5, Azure Blob Storage

## About

121,190 thermal videos from The Cacophony Project, captured mostly at night across New Zealand to monitor native wildlife and invasive predators such as possums, rodents, cats, hedgehogs and mustelids. Capture is triggered by a change in incident heat; about 24,000 videos are labeled false positives, and labels cover 45 categories (most common bird, rodent, possum). Annotations are machine-generated tracks with human-reviewed tags in JSON: one main metadata file plus a per-clip JSON with track coordinates; a location-based train/test split is provided. Each clip comes as a background-subtracted MP4, a normalised MP4 and an HDF5 file of raw thermal frames (frame rate 9). Layout (measured 2026-11-13): unzipped videos/ (MP4) and hdf/ (.hdf5) folders and individual-metadata/, plus a 35.9 GB zip of metadata and MP4s and small counts, metadata and splits files. The MP4 and HDF5 files exist as individual files, so the read-only mount can stream them with h5py or OpenCV without extracting the zip. Released by LILA BC under the Community Data License Agreement (permissive variant), CDLA-Permissive-1.0; no citation is required.

## Load the data

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

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

## Data structure

File types in sampled files: HDF5

Example file: `hdf/1000015.hdf5`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/2ab2661a-07b9-5955-9b63-f9915afa085f
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/2ab2661a-07b9-5955-9b63-f9915afa085f
- Over MCP: `get_schema_card` with id `2ab2661a-07b9-5955-9b63-f9915afa085f` at https://api.scigantic.com/mcp

## Related topics

- [Environmental Science datasets](https://scigantic.com/topics/environmental-science)
- [HDF5 datasets](https://scigantic.com/topics/hdf5)
