Real-Time Vehicle Detection, Tracking, and Counting

This repository contains code for real-time vehicle detection, logging, and tracking . The system utilizes YOLOv8 and DeepSORT algorithms implemented using the OpenCV library in Python.

Overview

The project aims to perform real-time vehicle logging system , making it suitable for various applications. Specifically, it was developed for our internship project at NITK.

Features

• Real-time Processing: The system performs vehicle detection, counting, and tracking in real-time, enabling its use in dynamic environments.
• Modified Dataset: The dataset used in the project is derived from Roboflow. Initially consisting of 20 classes, it was narrowed down to 6 classes to focus on specific vehicle types based on categories, ensuring a more focused and accurate model.
• Tesseract OCR Integration: The system integrates Tesseract OCR for automated time extraction from captured frames. This feature enables efficient logging of vehicle passage times.

Object Classes

The following 6 categories are used for vehicle classification:

1. Car/Jeep/Van/MV
2. LCV/LGV/Mini Bus
3. 2 axle
4. 3 axle Commercial
5. 4 to 6 axle
6. Oversized (7 axle)

Model Architecture

The vehicle detection model is based on YOLOv8, which is trained on the modified dataset. This architecture provides efficient and accurate detection of vehicles in real-time scenarios.

Acknowledgments

The project was made possible by leveraging the Roboflow platform for dataset management and annotation.

Dataset Source

The dataset used in this project can be found at: Roboflow Vehicle Detection Dataset