Logistic-object-detection

This project contains a Gazebo simulation with all the LOCO dataset and the chance to perform object detection using YOLOv8 n with this data.

Logistic simulation

1. Clone repository

cd ros2_ws/src
git clone https://github.com/vidaldani/Logistic-object-detection.git

2. Compile workspace

cd ..
colcon build
source install/setup.bash 

3. Export Turtlebot model

gedit ~/.bashrc
# copy this instruction at the end: export TURTLEBOT3_MODEL=waffle
source ~/.bashrc

4. Copy the models inside /turtlebot3_gazebo/models/logistic_objects to ~/.gazebo/models folder

cp -r ~/ros2_ws/src/Logistic-object-detection/src/turtlebot3_gazebo/models/logistic_objects/* ~/.gazebo/models/

6. Launch logistic simulation:

ros2 launch turtlebot3_gazebo versuchshalle.launch.py

Launch object detection

1. Install ultralytics (if it is already installed jump to step 2)

   pip3 install ultralytics==8.3.40

2. On a different terminal launch logistic object detection

   cd ros2_ws/
   source install/setup.bash 
   ros2 launch turtlebot3_recognition launch_yolov8.launch.py

3. Open rviz to visualize the inference results. The configuration file can be found in the folder /turtlebot3_recognition/rviz

   rviz2 -d $(pwd)/src/turtlebot3_ws/src/turtlebot3_recognition/rviz/config.rviz

4. In a new terminal launch the keyboard teleoperation to move the robot around and perform object detection

   ros2 run teleop_twist_keyboard teleop_twist_keyboard

3D object detection

1. Launch 3D object detection

   ros2 launch turtlebot3_recognition launch_3d_node.launch.py