roboeye

KO | EN

gitlite — search

#python #java #python3 #arduino #golang #machine-learning #rust #html #flask #javascript #seismology #nodejs

★ 17 Open GitHub ↗

Visual inertial odometry system using Raspberry Pi

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# roboeye
Low cost Visual Inertial Odometry (VIO) system using Raspberry Pi 4/5.

# Hardware
* Raspberry Pi 4B, 4GB RAM
* IMU: MPU6050
* Global shutter monochrome camera (for example, [OV9281](https://www.arducam.com/product/arducam-ov9281-1mp-global-shutter-noir-mono-mipi-camera-with-130deg-m12-mount-for-raspberry-pi/), OR [this](https://www.uctronics.com/arducam-ov9281-monochrome-global-shutter-camera-module-wide-angle.html))

# Setup
**NOTE** This is tested onm Raspberry Pi 4 with 4GB RAM with Bullseye OS. [Raspberry Pi OS (Legacy) with desktop](https://www.raspberrypi.com/software/operating-systems/). The ROS Noetic version works way better than the ROS2 humble version. In fact, openvins estimator tends to diverge quickly using the ROS2 version. However, a setup script is provided to install the ros2 system.

## Recommended Setup
* Use RPi OS 64-bit, [Raspberry Pi OS (Legacy) with desktop](https://www.raspberrypi.com/software/operating-systems/)
* Clone this pkg in your RPi.
    ```sh
    mkdir -p ~/src
    cd $HOME/src && git clone https://github.com/mzahana/roboeye.git
    ```
* Enable environment variables
    ```sh
    BUILD_ROS=true
    BUILD_OPENVINS=true
    # If you are using Arducam camera hat enable the following
    INSTALL_ARDUCAM=true
    ```
* Execute the `setup_noetic_openvins_bullseye.sh` script
    ```sh
    source setup_noetic_openvins_bullseye.sh
    ```
The setup amy take long time!

* Add the following to your `.bashrc`.
    ```sh
    source /home/$USER/ros_noetic/catkin_ws/install_isolated/setup.bash
    source /home/$USER/catkin_ws/devel/setup.bash
    ```
# Run
* After the setup is done, reboot your RPi
* You will need to calibrate your IMU. Follow instructions [here](https://github.com/Brazilian-Institute-of-Robotics/mpu6050_driver)
* You will need to  calibrate your camera-imu setup using [Kalibr tool](https://github.com/ethz-asl/kalibr). It is recommended that you record a rosbag of your camera/IMU topics. You can check [launch/arducam_mpu.launch](launch/arducam_mpu.launch) on how to run camera and IMU nodes, then record a bag.  Then use Kalibr to perfrom the calibration. Do the calibration on your PC. A good calibration is important!
* Transfer the calibration data from Kalibr output to a yaml files that openvins can read. See the examples in [config/openvins](config/openvins)
* You can use [arducam_mpu_openvins.launch](launch/arducam_mpu_openvins.launch) to run the entire system.
* Run Rviz on your PC to visualize the odomery (make sure to set `ROS_MASTER_URI` and `ROS_HOSTNAME` properly).

# System Service
The camera, imu, and openvins can be run upon system startup using the the system service [arducam_stereo_mpu_openvins.service](services/arducam_stereo_mpu_openvins.service) which runs the [arducam_openvins.sh](services/arducam_openvins.sh) script that runs the required nodes. The service assumes that this shell script is located in the `$HOME` directory of the user `vio`. 
* To install the service, copy it to `/etc/systemd/system/`
* Reload `sudo systemctl daemon-reload`
* Enable the service `sudo systemctl enable arducam_stereo_mpu_openvins.service`
* Start the service `sudo systemctl start arducam_stereo_mpu_openvins.service`
* To stop the service `sudo systemctl stop arducam_stereo_mpu_openvins.service`
* To disable the service `sudo systemctl disable arducam_stereo_mpu_openvins.service`

# Useful commands
To add to `.bashrc`

```sh
if [ -f "/home/vio/ros_noetic/catkin_ws/install_isolated/setup.bash" ]; then
    source /home/vio/ros_noetic/catkin_ws/install_isolated/setup.bash
fi

if [ -f "/home/vio/catkin_ws/devel/setup.bash" ]; then
    source /home/vio/catkin_ws/devel/setup.bash
fi

if [ -f "/home/vio/mavros_ws/devel/setup.bash" ]; then
    source /home/vio/mavros_ws/devel/setup.bash
fi

export ROS_MASTER_URI=http://localhost:11311
#export ROS_HOSTNAME=192.168.8.124

alias enable_vio_service='sudo systemctl enable vio_system.service'
alias disable_vio_service='sudo systemctl disable vio_system.service'
alias start_vio_service='sudo systemctl start vio_system.service'
alias stop_vio_service='sudo systemctl stop vio_system.service'
alias restart_vio_service='sudo systemctl restart vio_system.service'
alias vio_service_status='sudo systemctl status vio_system.service'

# Useful scripts
alias print_pose='python /home/vio/scripts/print_pose.py'
alias print_imu='python /home/vio/scripts/print_imu.py'
alias reset_pose='rosservice call /rovio/reset "{}"'

function calibrate_mpu6050 {
    echo "Stopping vio service" && sleep 1
    stop_vio_service
    disable_vio_service
    # Default YAML configuration file path
    local default_yaml_path="$HOME/config/mpu_settings.yaml"
    local yaml_path="${1:-$default_yaml_path}"  # Use the argument if provided, otherwise use the default

# Hardcoded path to the launch file
    local launch_path="$HOME/launch/mpu6050_automated_calibration.launch"

# Inform the user which YAML configuration file is being loaded
    echo "Launching with YAML configuration: $yaml_path"

# Launch the ROS node with the specified YAML configuration
    roslaunch "$launch_path" yaml_file:="$yaml_path"
}
```

roboeye

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README