Skip to content
All writing

May 15, 2024 · Computer Vision · 8 min read

Implementing YOLO for Real-time Object Detection at HAMK

A deep dive into how I implemented YOLO (You Only Look Once) for real-time object detection during my HAMK internship. Learn about the challenges and solutions in computer vision projects.

During my internship at HAMK University of Applied Sciences, I had the opportunity to work on an exciting computer vision project implementing YOLO (You Only Look Once) for real-time object detection.

Project Overview

The project involved developing a real-time object detection system using YOLOv5 and OpenCV. The goal was to create an efficient system that could detect and classify multiple objects in video streams with high accuracy and minimal latency.

Technical Implementation

I used the following technologies and approaches:

  • YOLOv5 for object detection with custom dataset training
  • OpenCV for video processing and frame manipulation
  • Django for creating a web interface to display results
  • Python for backend processing and model integration

Results

The final system achieved 95% accuracy with processing speeds of 30+ FPS, making it suitable for real-time applications.