NORMA eResearch @NCI Library

Comparative Analysis of YOLO Variants for Object Detection in Thermal Images

Eramplackal, Shajo Varghese (2023) Comparative Analysis of YOLO Variants for Object Detection in Thermal Images. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
Preview
PDF (Master of Science)
Download (1MB) | Preview
[thumbnail of Configuration Manual]
Preview
PDF (Configuration Manual)
Download (784kB) | Preview

Abstract

This study delves into the application of the state-of-the-art object detection models inside the YOLO (You Only Look Once) family, with a special focus on YOLOv5, YOLOv7, and YOLOv8, in the context of thermal imaging for autonomous vehicles and which can also be extended to security applications. Focusing on the limitations of traditional vision systems under challenging and extreme conditions such as low visibility and extreme weather conditions, it advocates for thermal imaging as a viable alternative This approach seeks to overcome the weaknesses of Camera sensors,RADAR (Radio Detection and Ranging) and LIDAR(Light Detection and Ranging). The study involves the acquisition and preprocessing of a diverse set of thermal images (TIR - Thermal Infrared), consisting of various real-world scenarios to ensure compatibility with the YOLO architecture. A detailed comparative analysis of YOLOv5, YOLOv7, and YOLOv8 is conducted, focusing on evaluation metrics like Precision, Recall, and mAP (Mean Average Precision). This analysis aims to capture the model’s performance in the thermal imaging domain for detecting car and humans under real-world conditions. YOLOv5 and YOLOv8 had similar mAP, but the latter had slightly higher precision and recall, making it the best performing model in this study. The findings from this research will contribute significantly to autonomous driving systems and security surveillance, offering valuable insights for the selection of appropriate object detection model considering application-specific requirements.

Item Type: Thesis (Masters)
Supervisors:
Name
Email
Agarwal, Bharath
UNSPECIFIED
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
H Social Sciences > HD Industries. Land use. Labor > Specific Industries > Motor Industry
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Ciara O'Brien
Date Deposited: 08 May 2025 09:54
Last Modified: 08 May 2025 09:54
URI: https://norma.ncirl.ie/id/eprint/7510

Actions (login required)

View Item View Item