AIDERv2 Dataset
Overview
AIDERv2 (Aerial Image Dataset for Emergency Response Applications) is an aerial image classification dataset for disaster recognition. It contains 16,723 UAV and aerial images across four classes:
- Earthquake
- Flood
- Fire
- Normal
Statistics
| Class | Train | Validation | Test | Total |
|---|---|---|---|---|
| Earthquake | 1,927 | 239 | 239 | 2,405 |
| Flood | 4,063 | 505 | 502 | 5,070 |
| Fire | 3,509 | 439 | 436 | 4,384 |
| Normal | 3,900 | 487 | 477 | 4,864 |
| Total | 13,399 | 1,670 | 1,654 | 16,723 |
Image Format
- RGB Images
- Resolution: 224 × 224 pixels
- Dataset split into Train, Validation, and Test sets
Applications
- Disaster Recognition
- Aerial Image Classification
- UAV-Based Emergency Response
- Deep Learning and Computer Vision Research
Source
Original dataset:
https://zenodo.org/records/10891054
Citation
@inproceedings{shianios2023aider,
title={A Benchmark and Investigation of Deep-Learning-Based Techniques for Detecting Natural Disasters in Aerial Images},
author={Shianios, Demetris and Kyrkou, Christos and Kolios, Panayiotis S.},
booktitle={Computer Analysis of Images and Patterns (CAIP)},
year={2023},
publisher={Springer}
}