Darknet YOLO with ROS 2

Stop Sign Detection - Simulation	Stop Sign Detection - TurtleBot3	Stop Sign Detection - Remote PC

Note: The above demonstrations use this repository for controlling TurtleBot3 to obey the stop sign.

Build:

1. Make a directory ROS2_WS to act as your ROS 2 workspace.

   $ mkdir -p ~/ROS2_WS/src/

2. Clone this repository:

   $ git clone https://github.com/Tinker-Twins/YOLO-ROS-2.git

3. Install OpenCV (or build from source).

   $ sudo apt update
   $ sudo apt install libopencv-dev python3-opencv

4. Build the ROS packages (build in Release mode to maximize performance).

   $ cd ~/ROS2_WS
   $ colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release

5. Source the setup.bash file of your ROS2_WS.

   $ echo "source ~/ROS2_WS/install/setup.bash" >> ~/.bashrc
   $ source ~/.bashrc

Execute:

$ ros2 launch darknet_ros darknet_ros.launch.py

Configure:

• Installation configuration can be managed from CMakeLists.txt.
  • darknet_ros/darknet_ros/darknet/CMakeLists.txt

    # CUDA/cuDNN Settings
    option(ENABLE_CUDA "Enable CUDA support" OFF)
    option(ENABLE_CUDNN "Enable CUDNN" OFF)
    option(ENABLE_CUDNN_HALF "Enable CUDNN Half precision" OFF)

  • darknet_ros/darknet_ros/darknet_ros/CMakeLists.txt

    # CUDA/cuDNN Settings
    set(CUDA_ENABLE OFF)
    set(CUDNN_ENABLE OFF)
    set(FP16_ENABLE OFF)
    
    # YOLO Pre-Trained Model (Weights) Settings
    set(DOWNLOAD_YOLOV2_TINY OFF)
    set(DOWNLOAD_YOLOV3 OFF)
    set(DOWNLOAD_YOLOV4 OFF)
    set(DOWNLOAD_YOLOV4_CSP ON)
    set(DOWNLOAD_YOLOV4_TINY ON)
    set(DOWNLOAD_YOLOV4_MISH OFF)
    set(DOWNLOAD_YOLOV7_TINY ON)

• Names and other parameters of the publishers, subscribers and actions can be modified from darknet_ros/config/ros.yaml.
• Parameters related to YOLO object detection algorithm can be modified from darknet_ros/darknet_ros/darknet_ros/config/yolo.yaml.
• It is recommended to create a copy of the existing configuration file(s) as a template and do necessary modifications.
• Reference the updated configuration file(s) in darknet_ros/darknet_ros/darknet_ros/launch/darknet_ros.launch.py.