↓Skip to main content

Data Overview

Table of Contents

File Structure #

M3ED/
├── input/
│   ├── raw_bags/
│   │   ├── car_forest_camera_calib.bag
│   │   ├── car_imu_calib.bag
│   │   └── car_forest_data_1.bag
│   └── lidar_calibrations/
│       └── car_forest_data_1.npz
└── processed/
    ├── car_forest_data_1/
    │   ├── car_forest_data_1.h5
    │   ├── car_forest_data_1.pcd
    │   ├── car_forest_data_1_stats.yaml
    │   ├── imgs.avi
    │   └── rgb.avi
    ├── car_forest_camera_calib/
    │   ├── car_forest_camera_calib.h5
    │   ├── camchain.yaml
    │   ├── report-cam.pdf
    │   ├── results-cam.txt
    │   ├── imgs.avi
    │   └── rgb.avi
    └── car_imu_calib/
        └── camchain.yaml

Input Files #

  • raw_bags stores the unprocessed data obtained from the sensor tower. These files are not time corrected and events are stored in binary format. Please refer to the m3ed repo for more information.
  • lidar_calibrations stores the manually tuned LiDAR calibrations for each bag.

Processed Files #

Each sequence is recorded individually in the processed folder.

  • data.h5 time synchronized and event decoded. Use these files to access the event camera data.
  • gt.h5 ground truth depth and poses from FasterLIO. Use these files to access ground truth poses, as well as to access the depth GT.
  • stats.yaml statistics for the file.
  • *.mp4 recordings showing the full data.
  • semantics.h5 InternImage semantic outputs.

For calibration bags, results from Kalibr are provided.