Catacombs_Visual-Inertial_Data

This dataset contains visual inertial data for an experiment at the Catacombs section of an underwater cave in Devil's Eye Cave System, FL, USA.

The data are located at:

https://huggingface.co/datasets/afrl-uw/Catacombs_Visual-Inertial_Data

The Software Pipeline described at:

https://github.com/AutonomousFieldRoboticsLab/Catacombs_SW_Pipeline

One dive was performed at the Catacombs section of the Devil's Eye Cave System with durations of 55 (2024), while data were recorded for 39 minutes. Calibration targets (4-by-2 and 5-by-4) were deployed in two different locations and stayed fixed for the duration of the dive.

The directory structure is as follows:

• Configurations
  -- april_3x5_10x10cm.yaml
  -- april_4x5_4x4cm.yaml
  -- GoProCamIMU.yaml
• Ros2Bags
  -- 2024_Center/
  -- 2024_Left/
  -- 2024_Right/
• COLMAP_SVIn2_MNemo_Trajectories
  -- 2024_Center_Colmap.txt
  -- 2024_Left_Colmap.txt
  -- 2024_Right_Colmap.txt
  -- 2024_Center_SVIn2.txt
  -- 2024_Left_SVIn2.txt
  -- 2024_Right_SVIn2.txt
  -- CatacombsLC.txt
• Videos
  -- 2024_Center/
  -- 2024_Left/
  -- 2024_Right/
• WaterDepthProfiles
  -- 2024_Perdix.csv/

Citation

If find this dataset useful, please cite: