Horizon detection in the visible spectrum
Loading...
Authors
Crisman, Donald M.
Subjects
attitude determination
machine learning
image classification
earth horizon sensor
computer vision
line detection
visible horizon
visible spectrum imagery
horizon detection
machine learning
image classification
earth horizon sensor
computer vision
line detection
visible horizon
visible spectrum imagery
horizon detection
Advisors
Kolsch, Mathias
Date of Issue
2016-09
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
In the last few decades, machine learning and computer vision techniques have enabled precise and repeatable image recognition. Computer vision techniques can also recognize star patterns in star trackers for satellite attitude determination. Horizon detection in the visible spectrum was largely discarded for attitude determination in favor of thermal imagery, due to the greater consistency of the earth's thermal radiation. This thesis examines computer vision and machine learning techniques to develop a horizon detection algorithm for the visible spectrum. By examining different features of visual imagery, machine learning techniques were evaluated on the ability to detect a visible horizon and determine its orientation. An empirical analysis of visual imagery from low-earth orbit was conducted to develop a horizon brightness transition model, which allows for consistent and adjustable determination of the horizons location. The final result is a horizon detection and orientation determination algorithm that successfully indicates if a horizon is present in an image with 96% precision and 92% recall. The brightness model correctly identifies the location of the horizon in 85% of the tested image set.
Type
Thesis
Description
Series/Report No
Department
Organization
Identifiers
NPS Report Number
Sponsors
Funding
Format
Citation
Distribution Statement
Approved for public release; distribution is unlimited.
Rights
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.