Date of Award

3-23-2018

Document Type

Thesis

Degree Name

Master of Science in Computer Engineering

Department

Department of Electrical and Computer Engineering

First Advisor

Kenneth M. Hopkinson, PhD.

Abstract

This research explores the use of Convolutional Neural Networks (CNNs) to classify targets of interest within satellite imagery. Methods were specifically devised for the classification of airports within Landsat-8 scenes. A novel automated dataset generation technique was developed to create labeled datasets from satellite imagery using only coordinate metadata. Using this approach a very large dataset of over 132,000 labeled images was created without human input. This dataset was used to evaluate the effects of color and resolution on airport classification accuracy. Two experiments were run with the first experiment classifying large airports with 96.8% accuracy, and the second classifying large and medium airports with 90.2% accuracy. Additionally, a new algorithm was developed which optimizes the selection of multi-spectral color bands in order to best trade-off classification accuracy for the number of spectral bands employed.

AFIT Designator

AFIT-ENG-MS-18-M-043

DTIC Accession Number

AD1056164

Share

COinS