A neural network-based system for tracking sea-ice floes
Climate modelling and high-latitude marine navigation require improved information on sea-ice floe extents and dynamics. New satellite sensors provide raw data of this nature but the volume of information makes human analysis impractical. To address this problem, a software system for automatic trac...
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis |
Language: | English |
Published: |
McGill University
1996
|
Subjects: | |
Online Access: | http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=24014 |
id |
ftcanadathes:oai:collectionscanada.gc.ca:QMM.24014 |
---|---|
record_format |
openpolar |
spelling |
ftcanadathes:oai:collectionscanada.gc.ca:QMM.24014 2023-05-15T18:16:28+02:00 A neural network-based system for tracking sea-ice floes James, Zachary D. Lewis, John E. (advisor) Master of Science (Department of Geography.) 1996 application/pdf http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=24014 en eng McGill University alephsysno: 001538947 proquestno: MM19823 Theses scanned by UMI/ProQuest. http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=24014 All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. Sea ice -- Remote sensing Sea ice -- Computer simulation Neural networks (Computer science) Electronic Thesis or Dissertation 1996 ftcanadathes 2014-02-16T00:56:53Z Climate modelling and high-latitude marine navigation require improved information on sea-ice floe extents and dynamics. New satellite sensors provide raw data of this nature but the volume of information makes human analysis impractical. To address this problem, a software system for automatic tracking of sea-ice floes in satellite imagery has been designed and evaluated. Using a recurrent neural network model, experiments were conducted to discover suitable design parameters. Simulated imagery time-sequences of increasing complexity were produced to train successive models. The networks produced were evaluated based on performance and reliability. A small-scale working system, able to map multiple input features in image sequences to Cartesian coordinates, was produced. Results show that a recurrent neural network is suitable for the tracking task and has advantages in robustness and speed over other approaches. Recurrency (feedback) was found to be crucial in achieving good performance. Thesis Sea ice Theses Canada/Thèses Canada (Library and Archives Canada) |
institution |
Open Polar |
collection |
Theses Canada/Thèses Canada (Library and Archives Canada) |
op_collection_id |
ftcanadathes |
language |
English |
topic |
Sea ice -- Remote sensing Sea ice -- Computer simulation Neural networks (Computer science) |
spellingShingle |
Sea ice -- Remote sensing Sea ice -- Computer simulation Neural networks (Computer science) James, Zachary D. A neural network-based system for tracking sea-ice floes |
topic_facet |
Sea ice -- Remote sensing Sea ice -- Computer simulation Neural networks (Computer science) |
description |
Climate modelling and high-latitude marine navigation require improved information on sea-ice floe extents and dynamics. New satellite sensors provide raw data of this nature but the volume of information makes human analysis impractical. To address this problem, a software system for automatic tracking of sea-ice floes in satellite imagery has been designed and evaluated. Using a recurrent neural network model, experiments were conducted to discover suitable design parameters. Simulated imagery time-sequences of increasing complexity were produced to train successive models. The networks produced were evaluated based on performance and reliability. A small-scale working system, able to map multiple input features in image sequences to Cartesian coordinates, was produced. Results show that a recurrent neural network is suitable for the tracking task and has advantages in robustness and speed over other approaches. Recurrency (feedback) was found to be crucial in achieving good performance. |
author2 |
Lewis, John E. (advisor) |
format |
Thesis |
author |
James, Zachary D. |
author_facet |
James, Zachary D. |
author_sort |
James, Zachary D. |
title |
A neural network-based system for tracking sea-ice floes |
title_short |
A neural network-based system for tracking sea-ice floes |
title_full |
A neural network-based system for tracking sea-ice floes |
title_fullStr |
A neural network-based system for tracking sea-ice floes |
title_full_unstemmed |
A neural network-based system for tracking sea-ice floes |
title_sort |
neural network-based system for tracking sea-ice floes |
publisher |
McGill University |
publishDate |
1996 |
url |
http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=24014 |
op_coverage |
Master of Science (Department of Geography.) |
genre |
Sea ice |
genre_facet |
Sea ice |
op_relation |
alephsysno: 001538947 proquestno: MM19823 Theses scanned by UMI/ProQuest. http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=24014 |
op_rights |
All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
_version_ |
1766190087836532736 |