Furg Fire
2D Box
Action/Event Detection
License: CC0 1.0


A dataset to compare the performance of different non-stationary video based fire detectors. This is a new dataset of fire videos containing frame by frame annotations which may be used for non-stationary fire detection algorithms evaluation. The dataset includes hand-held, robot attached and drone attached footages and aims to boost the development of fully autonomous fire fighter robots. The presented ground truth and metrics may adapt to any state-of-the-art technique and provide a reliable and unbiased solution to compare them.

Data Annotation

Each video file has an associated XML annotation file (in the weird OpenCv 2.4.9 XML format) with regular rectangles around each flame.
This XML was generated using OpenCv 2. 4.9 and hasn't (probably won't) been updated since.
Rect(int x, int y, int width, int height) x and y are the top left corner. For more info, have a look at the documentation: OpenCv Rect


  title={An unconstrained dataset for non-stationary video based fire detection},
  author={Steffens, Cristiano Rafael and Rodrigues, Ricardo Nagel and da Costa Botelho, Silvia Silva},
  booktitle={Robotics Symposium (LARS) and 2015 3rd Brazilian Symposium on Robotics (LARS-SBR), 2015 12th Latin American},


CC0 1.0

Data Summary
Provided by
Cristiano Rafael Steffens
Software Engineer and PhD Candidate in Florianópolis, State of Santa Catarina, Brazil.
Start Building AI Now