ITU ARIS Lab. Acoustic Drone Dataset
Authors/Creators
Description
An acoustic dataset for drone detection.
The recordings were made outdoors in a single continuous session, with several drone models flown through a range of manoeuvres at distances from a few metres out to roughly two kilometres. Weather, wind and background traffic were whatever the day provided; nothing was staged.
The audio is split into one-second segments and labelled in two classes, drone and background. Ready-made train/validation/test splits are included, along with a file mapping each segment back to the part of the session it came from. Use that mapping rather than splitting the files at random — neighbouring segments overlap, so a random split puts the same audio on both sides. The included README explains the layout and the known limitations of the set.
Collected and released by the ARIS Laboratory, Department of Electronics and Communication Engineering, Istanbul Technical University.
Files
itu-aris-lab-acoustic-drone-dataset-v1.0.0.zip
Additional details
Additional titles
- Translated title (Turkish)
- İTÜ ARIS Lab. Akustik Dron Tespiti
Related works
- Is supplement to
- Conference paper: 10.1109/SIU71813.2026.11636719 (DOI)
Dates
- Accepted
-
2026Accepted for IEEE Conference.
Software
- Repository URL
- https://github.com/imm-hub/acoustic-drone-dcl
- Programming language
- Python
- Development Status
- Active