Land Vehicle Speed Estimation in an Augmented Reality Space using Computer Vision

Samuel Sciberras

Supervised by Dr Vanessa Camilleri
Co-supervised by Mr Dylan Seychell

Department of Artificial Intelligence
Faculty of ICT
University of Malta

A dissertation submitted in partial fulfilment of the requirements for the degree of M.Sc. in Artificial Intelligence.

May 2021

---

Table of Contents

1. Built With
2. Prerequisites
3. Installation
4. How To Use
5. Contact

Built With

• Python
• TensorFlow

Prerequisites

1. Download and install Anaconda distribution for Python 3.8
2. (Optional*) Download and install CUDA Toolkit 10.2
3. (Optional*) Download and install cuDNN compatible with version CUDA 10.2

* Required for real-time processing
* Requires Nvidia CUDA compatible GPU

Installation

1. Clone repository

   git clone https://github.com/samuelsciberras/msc-ai

2. Create conda environment

   conda env create --file yv4-dsort-speedest.yml

3. Download YOLOv4 trained models here and place them in directory /System/checkpoints/

How To Use

1. Activate conda environment

   conda activate yv4-dsort-speedest

2. Run

   cd System/
   python main.py

Contact

Samuel Sciberras - [email protected]