An implementation of Isaac VIO and SLAM using the OakD stereo camera and imu
Note: This project is still under development. Optimizations for camera calibration, and IMU integration are still being developed. Provided here is a short video demo of how the project is progressing.
This implementation is intended to be run using the ROS docker system Nvidia provides here.
To build and run the docker images, the provided script can be used. This script adds OakD support to the existing docker system provided by Nvidia.
./run_docker.sh
Plug in the camera via USB3. Inside the docker container, launch the applications to read sensor data and process Elbrus SLAM
ros2 launch oakd_s2 oakd_s2.launch.py
To visualize output with Rviz, open a new termal in the container with the run_docker.sh
script, the run
./rviz.sh