KAIST Divergent Path (KAIST-DP) Dataset

---

Publish Paper

If you use KAIST-DP dataset for your academic research, please cite our paper.

• Sungjae Shin, Yeeun Kim, Byeongho Yu, Eungchang Mason Lee, and Hyun Myung, "PanoNetVLAD: Visual Loop Closure Detection in Continuous Space Represented With Panoramic View Using Multiple Cameras," in Proc International Conference on Control, Automation and Systems (ICCAS), 2023, DOI:10.23919/ICCAS59377.2023.10316996.

---

Description

• Intel Realsense D455 x3
• Xsens MTI-300 IMU
• Velodyne LiDAR VLP 16

Message Topic Name

• Image Topic (sensor_msgs/Image):

  • Front Camera: "/front/camera/infra1/image_rect_raw" & "/front/camera/infra2/image_rect_raw"
  • Left Camera: "/left/camera/infra1/image_rect_raw" & "/left/camera/infra2/image_rect_raw"
  • Right Camera: "/right/camera/infra1/image_rect_raw" & "/right/camera/infra2/image_rect_raw"

• IMU Topic (sensor_msgs/Imu):

  • "/imu/data"

• LiDAR Topic (velodyne_msgs/VelodyneScan):

  • "/velodyne_packets"

1. Prerequisites

1.1 Ubuntu and ROS

Ubuntu 64-bit 16.04 or 18.04. ROS Kinetic or Melodic. ROS Installation

1.2. Realsense SDK

Follow Realsense SDK

1.3 Xsense Driver

Follow Xsens Driver

2. Build KAIST-DP dataset

Clone the repository and catkin_make:

    cd ~/catkin_ws/src
    git clone https://github.com/SungJaeShin/KAIST-DP.git
    cd ../../
    catkin config -DCMAKE_BUILD_TYPE=Release
    catkin build KAIST-DP
    source ~/catkin_ws/devel/setup.bash 

(if you fail in this step, try to find another computer with clean system or reinstall Ubuntu and ROS)

3. Get KAIST-DP dataset

Open two terminals, run multi camera and IMU and get the bag file respectively:

    roslaunch KAIST-DP three_cams.launch 
    
    roscd KAIST-DP
    cd sh_files/
    sh get_kaist_dp_bag.sh

(then, bag file save at bag folder !!)

4. Provided Dataset (ROSBAG Data)

• KAIST-DP Outdoor Dataset (Library):

  • E9 Building & E9 Building Tilt
  

• KAIST-DP EE-Building:

  • EE-Building
  

• KAIST-DP Indoor Dataset (EE):

  • E3 Building
  

5. Example of sensor data applying LIO-SAM and VINS-Fusion

• For LiDAR data applied LIO-SAM

• For Camera data applied VINS-Mono

6. Acknowledgements

These cameras and IMU are provided URL of KAIST.

Thank you for Byeongho Yu who made zig to get the KAIST-DP dataset.

For commercial inquiries, please contact Sungjae Shin (pootti@kaist.ac.kr)