Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Tree Species ClassicationUsing Terrestrial Photogrammetry
Umeå University, Faculty of Science and Technology, Department of Computing Science.
2013 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

This thesis investigates how texture classification can be used to automatically classify tree species from image of bark texture. The texture analysis methods evaluated in the thesis are, grey level co-occurrence matrix (GLCM), two different wavelet texture analysis methods and finally the scale-invariant feature transform. To evaluate the methods two classifiers, a linear support vector machine (SVM) and a kernel based import vector machine (IVM) was used. The tree species that were classified were Scotch Pine and Norwegian Spruce and the auxiliary class ground. Three experiments were conducted to test the methods. The experiments used subimages of bark extracted from terrestrial photogrammetry images. For each sub-image, the X ,Y and Z coordinates were available. Thefirst experiment compared the methods by classifying each sub-image individually based on image data alone. In the second experiment the spatial data was added. Additionally feature selection was performed in both experiments to determine the most discriminating features. In the final experiment individual trees were classified by clustering all data from each tree.For sub-image classification, the addition of spatial data increased the overall accuracy for the best method from 75.7% to 94.9% The best method was IVM on GLCM textural features. The most discriminating textural feature was homogeneity in the horizontal direction. The best methods to classify individual trees were SVM with GLCM with an overall accuracy of 88%.In summary, the methods was found to be promising for tree bark classification. However, the individual tree results were based on a low number of trees. To establish the methods' true usefulness, testing on a larger number of trees is necessary.

Place, publisher, year, edition, pages
2013.
Series
UMNAD, 847
National Category
Engineering and Technology
Identifiers
URN: urn:nbn:se:umu:diva-80792OAI: oai:DiVA.org:umu-80792DiVA: diva2:651502
Educational program
Master of Science Programme in Computing Science and Engineering
Supervisors
Examiners
Available from: 2013-09-26 Created: 2013-09-26 Last updated: 2013-09-26Bibliographically approved

Open Access in DiVA

fulltext(2916 kB)379 downloads
File information
File name FULLTEXT01.pdfFile size 2916 kBChecksum SHA-512
d909cbf98b3ec6b41632cc9245444a16ed7f3dddfc9116519c5b3b2af2fb26ff1be64413e51f6632faa8bdb4e75c470719569978a71cd9d473e253dd6a2c936f
Type fulltextMimetype application/pdf

By organisation
Department of Computing Science
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 379 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 720 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf