Vehicle Detection and Counting using YOLOv5

This project is a vehicle detection and counting system using the YOLOv5 model. It uses a pre-trained model to detect various types of vehicles in a video file and stores the detected counts for analysis. The script is implemented in a Jupyter Notebook and uses the YOLOv5 model to detect objects in a given video. The results are saved, and video processing, vehicle counting, and result visualization are implemented in this project.

Features

• Vehicle detection in a video using a YOLOv5 model.
• Counting detected vehicles of different classes.
• Visualization of detection results.
• Saving processed video with vehicle detections.

Prerequisites

• Python 3.9+
• CUDA-compatible GPU (for running YOLOv5 on GPU)
• Jupyter Notebook (if you prefer running the script interactively)

Python Packages

The following Python packages are required:

• torch
• opencv-python
• yolov5 (installable from the official YOLOv5 GitHub repository)
• numpy
• ipython

Running

• 1. Clone the repository
• 2. Prepare the Jupyter Notebook
• 3. Run all jupyter cells