NORMA eResearch @NCI Library

Military and Non-Military Vehicle Detection by Faster R-CNN and SSD300 Models using Transfer Leaning

Nandimandalam, Venkata Devaraju (2020) Military and Non-Military Vehicle Detection by Faster R-CNN and SSD300 Models using Transfer Leaning. Masters thesis, Dublin, National College of Ireland.

[img]
Preview
PDF (Master of Science)
Download (971kB) | Preview
[img]
Preview
PDF (Configuration manual)
Download (2MB) | Preview

Abstract

The need for using deep learning in military applications is increasing every day and this research is helpful for surveillance, target detections, and taking precautionary measures using Unmanned Aerial Vehicles (UAV) and strategic warning systems by detecting military vehicles. Detecting small military vehicles from Aerial images and segregating military and non-military vehicles is a challenging task. Solving image detection problems involve deep learning algorithms like Faster R-CNN and SSD300 that are proposed and implemented with the help of transfer learning. For evaluating the model’s performance twelve different metrics are calculated. As opposed to the performance of SSD300 model Faster R-CNN with FPN detected objects with high mAP value around 82 percent. The customized dataset is used for training the models and it contains five military and two non-military vehicle classes.
Keywords: Faster R-CNN, SSD300, transfer learning, military and non-military vehicle detection, aerial images.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science

Q Science > QA Mathematics > Computer software
T Technology > T Technology (General) > Information Technology > Computer software
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Dan English
Date Deposited: 20 Jan 2021 17:53
Last Modified: 20 Jan 2021 17:53
URI: http://norma.ncirl.ie/id/eprint/4410

Actions (login required)

View Item View Item