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
Automated Surface Inspection of Cross Laminated Timber
University West, Department of Engineering Science, Division of Industrial Engineering and Management, Electrical- and Mechanical Engineering.
2020 (English)Independent thesis Advanced level (degree of Master (One Year)), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

The use of wood, such as Cross Laminated Timber (CLT), has increased over the last years in the construction of buildings. CLT is primarily considered a structural component inside walls, floors, and ceilings. But, when the design specifies an exposed surface of the CLT, where the wood surface will be visible in the finished building, there are requirements on the visual appearance of the CLT. To fulfil the requirements visual inspections of wood defects of the surface are conducted, before the CLT-panel is shipped to the construction site. Human inspectors are not consistent evaluators of products, and their accuracy depends on factors such as expertise, motivation, and fatigue. This study proposes a solution to automate the inspection, where captured images are analysed by an object detection model, based on deep learning. The model Faster R-CNN has proven to be successful in medical image processing and for object detection. A challenge with deep learning models, such as Faster R-CNN, is the number of test images needed to train the model. Also, there are a limited number of suitable test images of wood defects available. By using a pre-trained Faster R-CNN model and adapt the model's skill to detect wood defects, instead of the model's intended objects, the model can be trained with a relatively small number of test images. This study relies on images of planks, similar to the ones used in CLT, and the number of test images in the study is very limited. To improve the results, the model needs to be trained with more test images of CLT-panels. Nevertheless, the model can locate the defects in the CLT-panel and identify the type of defect to a high degree. The results indicate that the proposed model can be used to automate the inspection of wood defects on the CLT surface.

Place, publisher, year, edition, pages
2020. , p. 26
Keywords [en]
Wood defect detection, Visual inspection, Object detection, Image processing, Deep learning
National Category
Robotics
Identifiers
URN: urn:nbn:se:hv:diva-15519Local ID: EXM810OAI: oai:DiVA.org:hv-15519DiVA, id: diva2:1455150
Subject / course
Robotics
Educational program
Master i robotik och automation
Supervisors
Examiners
Available from: 2020-07-23 Created: 2020-07-22 Last updated: 2020-07-23Bibliographically approved

Open Access in DiVA

No full text in DiVA

By organisation
Division of Industrial Engineering and Management, Electrical- and Mechanical Engineering
Robotics

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 548 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