ros2_yolo_integration

yolo_pkg

Usage

1. Run the provided activation script to start the container and prepare the environment

./yolo_activate.sh

2. Do colcon build and source ./install/setup.bash

r

3. Run yolo node

ros2 run yolo_pkg yolo_detection_node

Mode

The YOLO node can operate in different modes depending on your use case:

• Mode 1: Draw bounding boxes without screenshot
  • Displays YOLO-detected bounding boxes on the output topic /yolo/detection/compressed.
• Mode 2: Draw bounding boxes with screenshot
  • Captures a screenshot of each detected object and saves it.
• Mode 3: 5 fps screenshot.
  • Takes a screenshot from the camera feed at a rate of 5 frames per second.
• Mode 4: segmentation
  • Enables segmentation mode, displaying segmentation masks on the output topic /yolo/detection/compressed.

Function

• draw_bounding_boxes
  • draw_crosshair(bool) :
    • If set to True, the function will draw a crosshair at the center of the image.
  • screenshot(bool) :
    • If True, the function will capture a screenshot of the image after drawing the bounding boxes.
  • segmentation_status(bool) :
    • Controls whether the segmentation overlay is displayed on the output image.
  • bounding_status(bool) :
    • Controls whether bounding boxes are displayed on the output image.
• save_fps_screenshot
  • Captures screenshots at a fixed rate of 5 frames per second (FPS).

class diagram

yolo_example_pkg

This is a ROS 2 project for integrating YOLO with ROS 2, providing functionality for real-time object detection and bounding box visualization.

Features

• Receive compressed images from the /yolo/detection/compressed topic.
• Process images to convert them into OpenCV format.
• Draw bounding boxes around detected objects on the images.
• Publish processed images back to the /yolo/detection/compressed topic, which includes the drawn bounding boxes.

Usage

1. Run the provided activation script to start the container and prepare the environment

./yolo_activate.sh

2. Do colcon build and source ./install/setup.bash

r

3. Run yolo node

ros2 run yolo_example_pkg yolo_node

How to put your yolo model

To use the YOLO model in this package, follow these steps:

1. Place the Model File

   Download or obtain the YOLO .pt file (e.g., yolov8n.pt) and place it inside the models directory within the pkg folder:

   <your_ros2_workspace>/
   ├── src/
   │   ├── yolo_example_pkg/
   │   │   ├── models/
   │   │   │   ├── yolov8n.pt
               ├── yolov8n-seg.pt
   

   You can update the model used in the following files inside yolo_pkg:

   • yolo_segmentation_model.py
   • yolo_detect_model.py

2. Model Path Configuration

   The script dynamically loads the model from the package's shared directory using the following code:

   import os
   from ament_index_python.packages import get_package_share_directory
   
   model_path = os.path.join(
       get_package_share_directory("yolo_example_pkg"), "models", "yolov8n.pt"
   )