ROS_YOLO

This repository is to operacionalize YOLO model encapsulated into ROS2 middleware.

Project Structure

• train.py: File responsible for training YOLOv8 with the dataset using a .yaml configuration.
• test.py: File responsible for performing tests to verify if the integration between the camera and OpenCV is functioning correctly.
• yolo_opencv.py: File responsible for performing inference with the chosen dataset using images from the camera.
• keypoints.py: File responsible for testing YOLOv8 Pose keypoints for the desired angle.
• objects.py: File responsible for detecting pointed objects.
• multiperson_objects.py: File responsible for detecting pointed objects in multi-person scenarios.
• dataset/: Folder responsible for storing the pose dataset.
• runs/: Folder responsible for storing the trained model with the dataset and the obtained metrics.
• runs/weights/best.pt: File responsible for performing inferences (trained model).
• ros2_ws/: Folder containing files responsible for performing pose classification and pointed object detection in ROS 2.