Theses and Dissertations

ORCID

https://orcid.org/0009-0005-8008-9838

Advisor

Ball, John E.

Committee Member

Gurbuz, Ali C.

Committee Member

Khan, Samee U.

Date of Degree

5-16-2025

Original embargo terms

Immediate Worldwide Access

Document Type

Graduate Thesis - Open Access

Major

Electrical and Computer Engineering

Degree Name

Master of Science (M.S.)

College

James Worth Bagley College of Engineering

Department

Department of Electrical and Computer Engineering

Abstract

Counter-drone systems employ sensors such as radars and cameras to detect, track, and identify drones. These systems can confuse birds with drones, and the advent of bird-like drones exacerbates this problem. This thesis explores radar micro-Doppler, camera imagery, and their fusion for drone and bird classification. First, micro-Doppler models of quadcopters, birds, and bird-like drones were developed to generate a dataset of synthetic spectrograms for training support vector machine, k-nearest neighbors, Naïve Bayes, and convolutional neural network (CNN) classifiers. Second, transfer learning was applied on YOLOv10, a recently developed You Only Look Once object detector model, to detect and classify drones and birds in visual images. Lastly, a radar-camera fusion system that fuses a radar detector and CNN classifier and a camera YOLOv10 detector/classifier at the decision-level is proposed. A dataset of real-world spectrograms and images of quadcopters, bird-like drones, and birds in flight was collected to test this system.

Download

Share

COinS