RealSense Camera YOLO Tools

This repository contains Python scripts for working with Intel RealSense cameras, designed to streamline image capture and object detection workflows for use with tools like Roboflow.

Scripts

1. image_capture.py

A simple GUI for capturing images from a RealSense camera.

2. detection_gui.py

A GUI for running a trained YOLOv8 .pt model on live RealSense camera input.

Installation

1. Clone the repository: git clone https://github.com/Galaxia5987/realsense-tools

2. Make a virtual enviroment and install dependencies:

cd realsense-tools
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

3. Run one of the scripts: python3 image_capture.py

Typical Workflow

1. Capture Images:
   Use image_capture.py to collect and save images for training.

2. Annotate:
   Annotate images using Roboflow.

3. Train Model: Train the model using this Kaggle notebook.

4. Run Detection:
   Use detection_gui.py to load your trained .pt model and perform real-time detection from a Realsense camera.

Notes

• Make sure your RealSense camera is connected before running the scripts.