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
Holistic-Subjective Automatic Grading of Sawn Timber: Sensitivity to Systematic Changes
Luleå University of Technology, Department of Engineering Sciences and Mathematics, Wood Science and Engineering.
Luleå University of Technology, Department of Engineering Sciences and Mathematics, Wood Science and Engineering.
Luleå University of Technology, Department of Engineering Sciences and Mathematics, Wood Science and Engineering.ORCID iD: 0000-0002-4526-9391
2019 (English)In: International Wood Machining Seminar, 2019Conference paper, Published paper (Other academic)
Abstract [en]

Holistic-subjective automatic grading of sawn timber by a partial least squares (PLS) regression model required training of the model. This study tests the sensitivity towards systematic changes of a specialized PLS model, trained on a selected type of material suitable for a specific paneling product, when used to grade sawn timber systematically different than the material it was trained on. A sawmills automatic scanning system used cameras to measure knot and bark features on 900 planks. Each plank was split into three boards, and each board was shaped into an indoor paneling product and manually graded as desirable or undesirable at a planing mill. The plank grade was decided as the majority of the board-grade outcome. The knot and bark measurements were used to create a large set of feature variables for each plank that was correlated to the plank’s grade by PLS regression. Of the 900 available planks, 434 planks sawn from top logs were used as a class-balanced specialistic training set, with half of the planks resulting in a majority of desirable boards. The regression model trained on the class-balanced specialistic training set was used to grade a test set of 282 planks, containing 64 planks that by manual classification of automatically captured images were determined to be sawn from butt logs and were systematically different from the training material. The PLS model’s grading accuracy of the planks sawn from top logs was 76%, compared to 70% for the plank sawn from butt logs. The grading outcome resulted in a higher proportion of both delivered planks from the sawmill and received desirable planks by the planing mill when grading planks from top logs as compared to planks from butt logs. The results indicated that a specialistic PLS model should not be used for a generalistic use-case.

Place, publisher, year, edition, pages
2019.
Keywords [en]
sawmilling, wood, automatic grading, PLS-regression
National Category
Wood Science
Research subject
Wood Science and Engineering
Identifiers
URN: urn:nbn:se:ltu:diva-74170OAI: oai:DiVA.org:ltu-74170DiVA, id: diva2:1320130
Conference
International Wood Machining Seminar
Available from: 2019-06-04 Created: 2019-06-04 Last updated: 2019-08-15

Open Access in DiVA

fulltext(407 kB)9 downloads
File information
File name FULLTEXT02.pdfFile size 407 kBChecksum SHA-512
950249d6c916ad2a547919daf7d14869e518cd528aee34060571ec7abd3f608b3d114ee3c335a9e2206ae216cb377245d44db892ea09f124814da51313cac2da
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Olofsson, LinusBroman, OlofSandberg, Dick
By organisation
Wood Science and Engineering
Wood Science

Search outside of DiVA

GoogleGoogle Scholar
Total: 9 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: 38 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