Date of Award

2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Electrical and Computer Engineering

Committee Chair

Laurie L. Joiner

Committee Member

Reza Adhami

Committee Member

Yuri Shtessel

Committee Member

Shangbing Ai

Committee Member

David Pan

Subject(s)

Image processing, Three-dimensional imaging

Abstract

This dissertation details a novel class of descriptors called Oriented Feature Constellations (OFC). This new class of descriptors is used to relocate, or match, features in image sequences taken from a monocular camera. It produces faster and more precise results than a highly cited comparison method. Faster, when both are configured to produce the same number of points and more precise, by producing more correct matches when all other variables are held constant. This result directly corresponds to an increase in the precision of the 3D point clouds that can be reconstructed from an image sequence. This dissertation also provides analysis of the OFC descriptor class in order to demonstrate the effects of its definition and parameterization on performance. Through a constructed simulation that isolated different types of image distortions, various descriptor definitions are compared against each other. Parameters that further define descriptor variations were also analyzed to determine their effect. From this, an optimized definition of the OFC descriptor class was created. Testing of the OFC descriptor class was performed through the creation of a specific descriptor instantiation using a modification of an existing image corner detector as the feature extraction method. Named after the underlying corner detection method, this descriptor is called OFC-FAST. The performance of OFC-FAST is used in analysis against the comparative method. Tests were performed both in feature matching and in their overall reconstruction capability.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.