Flooded Road Environments Dataset (FRED)

This autonomous vehicle dataset has been developed to enable research into the detection of flooded roads during on-road deployment. The dataset was collected using a Renault Zoe with custom modifications to enable autonomy, including front and rear Blackfly cameras, an Ouster OS1 LiDAR, and a GNSS-corrected IMU. Data has been collected using the vehicle's sensor stack from 5 separate locations around Brisbane, Australia, both during and after flooding events. Semantic labels are provided for images to enable the development of detection methods, and corresponding position information from the GNSS-corrected IMU has been provided across sequences to additionally enable localization research for these scenarios.

Preprint Paper: https://arxiv.org/abs/2605.22018
Github python-based development kit: https://github.com/AVR3-Training-Centre/python-FRED
Huggingface space for dataset and SDK testing: https://huggingface.co/spaces/CMalone-Jupiter/python-FRED
Our Research Group: https://avr3.org.au/

Cite This Work

If you use this dataset in your work, you can cite our preprint article using:

@article{malone2026fred,
      title={FRED: A Multi-Modal Autonomous Driving Dataset for Flooded Road Environments}, 
      author={Connor Malone and Sebastien Demmel and Sebastien Glaser},
      year={2026},
      journal={Arxiv},
      url={https://arxiv.org/abs/2605.22018},
}

Dataset Structure

We adopt the following structure for FRED to include a KITTI-style format for the dataset and the native recording format using RTmaps.

  ├── flooded                             # Location sequences captured during flooding events
  │   ├── KITTI-style                     # Sequences in a KITTI-style format
  │   |   ├── Cambogan_20250811_113017    # Sequence by location
  │   |   |   ├── back-imgs
  │   |   |   |   └── <timestamp>.png     # Images in 'png' format
  │   |   |   ├── back-labels
  │   |   |   |   └── <timestamp>.png     # Semantic labels in 'png' format
  │   |   |   ├── front-imgs
  │   |   |   |   └── <timestamp>.png     # Images in 'png' format
  │   |   |   ├── front-labels
  │   |   |   |   └── <timestamp>.png     # Semantic labels in 'png' format
  │   |   |   ├── imu
  │   |   |   |   └── <timestamp>.txt     # IMU data formatted as a 'txt' file
  │   |   |   ├── ouster
  │   |   |   |   └── <timestamp>.bin     # Point clouds formatted as a binary file
  │   |   |   └── utm
  │   |   |       └── <timestamp>.txt     # UTM locations formatted as a 'txt' file
  │   |   ├── ...
  │   |   └── ...
  │   └── native-RTmaps                   # Sequences in native recording format
  │       ├── Cambogan_20250811_113017    # Sequence by location
  │       |   ├── Camera_Rec              # Recording files for image playback
  │       |   ├── IMU_Info_Rec            # Recording files for IMU playback
  │       |   └── Ouster_Rec              # Recording files for LiDAR playback
  │       ├── ...
  │       └── ...
  │
  └── dry                             # Location sequences captured while 'dry'
      ├── KITTI-style                     
      └── native-RTmaps              

Data Formats

Image Format

Images are stored in PNG format.

Point Cloud Format

Point clouds are stored in binary format (.bin), with each point containing x, y, z positions, as well as reflectivity values. Reflectivity values are surface normalized signal intensity measurements that range from 0 to 255. 3D coordinates are captured in the right-hand coordinate frame with the positive x-axis in the vehicle's direction of travel.

UTM Format

UTM data is stored in text file format (.txt), with UTM x and y values being stored as space separated values in the file.

IMU Format

Additional IMU information is also stored in text file format (.txt). A space delimiter is again used to separate values. The additional IMU data is stored in the following order:

[ Latitude, Longitude, Altitude,  
Roll, Pitch, Yaw,  
North Velocity, East Velocity,  
x Velocity, y Velocity, z Velocity,
x Angular Velocity, y Angular Velocity, z Angular Velocity,  
x Angular Velocity, y Angular Velocity, z Angular Velocity,  
x Angular Accel, y Angular Accel, z Angular Accel,  
x Angular Accel, y Angular Accel, z Angular Accel,  
Position Accuracy, Velocity Accuracy,  
Navstate Value, Numstat Value,  
Position Mode, Velocity Mode, Orientation Mode ]