ROS-Humble Object Detection

This repository contains a ROS-Humble based object detection project. It allows you to record videos, replay them, and run object detection on the recorded data using ROS 2 and an object detection node.

Table of Contents

• About the Project
• Usage Instructions
  • 1. Record Video in .mcap Format
  • 2. Play Recorded .mcap File
  • 3. Run Object Detection Node
• Setup and Installation
• ROS Commands
• Contributing
• License

---

About the Project

This project is a ROS-Humble based object detection application that uses a pre-trained object detection model. The system supports recording videos in .mcap format, replaying them, and running an object detection node that processes the frames and identifies objects in real-time.

Usage Instructions

1. Record Video in .mcap Format

To record a video in .mcap format, use a ROS 2-compatible recording tool. Ensure that your ROS setup is running and configured to capture the desired video stream.

2. Play Recorded .mcap File

Once you have recorded a video, you can replay it using the following command:

ros2 bag play -s mcap rec1_0.mcap

3. Run Object Detection Node

colcon build

4. Source the environment to make ROS packages available:

source ./install/setup.bash

5. Run the object detection node with the following command:

ros2 run object_det object_det_node --ros-args -p image_subscription:="/oakd/rgb/preview/image_raw"