Find results with:

Add another field

Search by date

to

Searching collections:

Dartmouth Dissertations

Add or remove collections

Home Dartmouth Dissertations Page 1

Reference URL

To link to the entire object, paste this link in email, IM or document

To embed the entire object, paste this HTML in website

To link to this page, paste this link in email, IM or document

To embed this page, paste this HTML in website

Crevasse detection in ice sheets using ground penetrating radar and machine learning / by Rebecca M. Williams.

you wish to report:

Your comment:

Your Name:

Submit

Cancel

...

Back to top

Powered by CONTENTdm®

Author	Williams, Rebecca M.
Title	Crevasse detection in ice sheets using ground penetrating radar and machine learning / by Rebecca M. Williams.
Date	2013
Abstract	The purpose of this thesis is to describe methods for detecting hidden crevasses on glaciers in Polar regions. Crevasses form in ice sheets as a result of forces imposed upon the brittle ice. The crevasses are hidden from view by bridges of soft snow accumulated over the voids. Humans, snow machines, and tractors can easily break these bridges when attempting to cross. The broader goal of this project is to aid the endeavors of scientists attempting to study hazardous areas. By reducing the significant risks posed to personnel and equipment working in extreme environments, more resources will be available to support scientific endeavors. The platform used is a 73 kg four-wheeled robot, named Yeti, who tows a 400 MHz ground penetrating radar system on his backside. This thesis presents methods to automatically classify ground penetrating radar (GPR) images of crevasses on ice sheets using a combination of support vector machines (SVM) and hidden Markov models (HMM) with appropriate un-biased processing that is suitable for real-time analysis and detection. Three processing schemes were tested and evaluated on 96 examples of Antarctic GPR imagery of crevasses from 2010 and 129 examples of Greenland imagery from 2011, collected by Yeti and a Pisten Bully tractor. The Antarctic and Greenland data were collected in the shear zone near McMurdo Station and between Thule Air Base, and Summit Station respectively. Using a modified cross validation technique, 86 of the Antarctic examples and 90 of the Greenland examples were correctly classified when trained and evaluated on down-sampled GPR images of crevasses as compared to 60% classification rate using raw data. However, the false alarm rates for SVM are high, on the order of 70 false positives per crevasse crossing. In order to reduce the false positives, HMM was used to pre-screen the data and mark locations in the GPR files to evaluate with the SVM. The combined HMM-SVM method retains all the correct classifications by the SVM to give 83% true positive rate, and re
Description	xi, 182 leaves : ill. (some col.) 28 cm.
Format	application/pdf
Advisor	Laura E. Ray, James H. Lever, Steven A. Arcone, Robert L. Hawley
College	Thayer School of Engineering.
Issuing Body	Dartmouth Dissertations are made available through the Dartmouth Digital Library.
Degree	Thesis (Ph.D.)--Dartmouth College.
Copyright	All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means without permission in writing from the author.
DOI	http://dx.doi.org/10.1349/ddlp.1013

Find results with:

Searching collections:

Home Dartmouth Dissertations Page 1

Crevasse detection in ice sheets using ground penetrating radar and machine learning / by Rebecca M. Williams.

Object Description

Description

Crevasse detection in ice sheets using ground penetrating...

Crevasse detection in ice sheets using ground penetrating...

	All
	Pages Enter page numbers or page ranges separated by commas. For example: 1,3,5-12