Main content

Digital Elevation Model Generation and Fusion

Show simple item record

dc.contributor.advisor Lindsay, John
dc.contributor.advisor Berg, Aaron
dc.contributor.author Fuss, Colleen E.
dc.date.accessioned 2013-09-25T15:00:42Z
dc.date.available 2013-09-25T15:00:42Z
dc.date.copyright 2013-09
dc.date.created 2013-09-10
dc.date.issued 2013-09-25
dc.identifier.uri http://hdl.handle.net/10214/7571
dc.description.abstract Digital elevation models (DEMs) are a necessary dataset in modelling the Earth’s surface and the many physical processed that interact with it. There are several ways to acquire elevation data and generate DEMs, and while each method has advantages and disadvantages all DEMs contain error. DEM fusion techniques with the aim of reducing DEM error have been proposed and tested in published literature with several successful results. These techniques have not however, utilized a clustering algorithm on multiple DEMs to exploit consistency in the estimates as an indication of accuracy and precision. This research developed and tested a new DEM fusion algorithm on multiple, overlapping DEMs generated from RADARSAT-2 imagery using stereo-radargrammetric methods. The main steps of the algorithm include slope and elevation thresholding followed by k-means clustering of the elevation estimates, as well as filtering and smoothing of the fusion product. Corroboration of the input DEMs, as well as products of each main step of the fusion algorithm, with a higher accuracy reference DEM by landuse class within the study area enabled a detailed analysis of the effectiveness of the DEM generation and the fusion algorithm. The generated DEMs contained systematic errors, large blunders, and regional offsets that varied according to landuse type, as well as the differences in scene acquisition date and sensor parameters. The main findings of the research were: the k-means clustering of the elevations improved the global accuracy of the estimates but reduced the precision; the number of final cluster members and the standard deviation of elevations before clustering both had a strong relationship to the error in the k-means estimates. It is therefore recommended that further research be conducted to investigate the relationship between elevation clustering error and the distribution of elevations before clustering, especially for specific landuse classes such as agricultural fields. en_US
dc.description.sponsorship OMAFRA, CSA, AAFC, GEOIDE, ORF en_US
dc.language.iso en en_US
dc.publisher University of Guelph en_US
dc.rights Attribution-NonCommercial-NoDerivs 2.5 Canada *
dc.rights.uri http://creativecommons.org/licenses/by-nc-nd/2.5/ca/ *
dc.subject DEM en_US
dc.subject digital elevation model en_US
dc.subject fusion en_US
dc.subject radargrammetry en_US
dc.subject generation en_US
dc.title Digital Elevation Model Generation and Fusion en_US
dc.type Thesis en_US
dc.degree.programme Geography en_US
dc.degree.name Master of Science en_US
dc.degree.department Department of Geography en_US
dc.degree.grantor University of Guelph en_US


Files in this item

Files Size Format View Description
Fuss_Colleen_201309_Msc.pdf 3.712Mb PDF View/Open CFuss MSc Thesis

This item appears in the following Collection(s)

Show simple item record

Attribution-NonCommercial-NoDerivs 2.5 Canada Except where otherwise noted, this item's license is described as Attribution-NonCommercial-NoDerivs 2.5 Canada