ROS2 YOLOv4 with USB Camera Integration

This README provides a comprehensive guide on how to integrate YOLOv4 for object detection in ROS2 using a USB camera. It covers the setup of a new ROS2 workspace, building the YOLOv4 package, and launching nodes for capturing video feed and performing object detection.

Prerequisites

Ensure you have the following installed on your system before proceeding:

• ROS2
• A USB camera compatible with ROS2

Setup Instructions

1. Install Packages Packages that may need to be installed.

   sudo apt update
   sudo apt install ros-humble-cv-bridge ros-humble-image-transport
   pip install opencv-python

Creating the ROS2 Workspace

2. Create a New Workspace:

   Navigate to your desired workspace directory. If you are using Docker, this should be at the same level as your Docker container. Create a new workspace named yolov4_ws.

   mkdir -p ~/yolov4_ws
   cd ~/yolov4_ws

3. Clone the YOLOv4 ROS2 Package:

   Clone the YOLOv4 package from the provided GitHub repository into the src directory of your workspace.

   git clone https://github.com/TKUwengkunduo/ROS2_YOLOv4_usbCamera.git src

4. Build the Workspace:

   Use colcon to build the workspace. The --symlink-install flag creates symbolic links instead of copying files, saving disk space.

   colcon build --symlink-install

Launching the Nodes

To launch the necessary nodes for capturing video and performing YOLOv4 object detection, follow these steps:

Open the Camera Node

To capture video from the USB camera, you'll need to launch the camera node.

source install/setup.bash
ros2 run camera publish_usb_cam

Start the YOLOv4 Detection Node

In a new terminal window, source the ROS2 environment again and then start the YOLOv4 detection node. This node subscribes to the video feed, performs object detection, and publishes the results.

source install/setup.bash
ros2 run yolov4 yolo_detection

View Detection Results

To see the detection results published by the YOLOv4 node, open another terminal window, source the ROS2 environment, and use the echo command to listen to the detection results topic.

source install/setup.bash
ros2 topic echo /yolo_detection/detections