BAMBI - Re-Identification Dataset
Authors/Creators
-
1.
Hogeschool PXL
- 2. University of Applied Sciences Upper Austria
- 3. University of Veterinary Medicine Vienna
-
4.
Management Center Innsbruck
- 5. FH Oberösterreich Fakultät für Informatik Kommunikation Medien: Hagenberg, Oberösterreich, AT
- 6. FH Oberösterreich Fakultät für Informatik Kommunikation Medien
- 7. Büro für Wildökologie und Forstwirtschaft e.U.
Description
This dataset contains thermal infrared drone footage of 1,724 individually annotated wild animals, captured for the purpose of advancing re-identification (Re-ID) algorithms in wildlife monitoring. The dataset includes three species commonly found in Central European habitats:
-
Red deer (Cervus elaphus)
-
Roe deer (Capreolus capreolus)
-
Wild boar (Sus scrofa)
Each individual animal is assigned a unique identifier across thermal video sequences, enabling animal Re-ID research under challenging real-world conditions. The dataset captures natural behaviors, varying poses, occlusions, group dynamics, and different environmental settings (e.g., forest edges, open fields) recorded at various times of day and year. All data were collected using drone-mounted thermal cameras under ethical guidelines and applicable legal regulations.
Files
reid.zip
Files
(75.4 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:e0b485e735b4eaca2a175d45df39ab82
|
75.4 MB | Preview Download |
Additional details
Dates
- Created
-
2025-05-28