Automated image recognition for wood species identification

Lead Institute HF Institute of Wood Research

Development of automated (digital) image recognition systems for wood species identification using artificial intelligence

Can wood species be identified with absolute certainty by means of machine learning?
We are working on this question in the FNR project financed by the BMEL in collaboration with the Fraunhofer Institute. The aim: automated image recognition systems for wood species identification.

Background and Objective

In order to strengthen the trade with legal raw materials and to protect consumers, the used wood species are determined at the Thünen Centre of Competence on the Origin of Timber. Cut specimens of wood and, in the case of fibrous materials such as paper, individual cells are routinely examined under a light microscope and analyzed by experts at great expense of time. Automated image recognition systems can save an enormous amount of work and time here.

With the introduction of the European Timber Regulation (EUTR) in 2013, the trade is obliged to document, among other things, the type of wood contained in the product in order to ensure its legal origin. Within the framework of the new research project, automated image recognition systems using artificial intelligence (AI) are to be developed in order to be able to check the wood species declaration of the manufacturers on a larger scale.

The Fraunhofer Institute ITWM is contributing its expertise by developing the specific algorithms and training the software using machine learning. The results are to be published scientifically and provided to all testing institutes for the control of internationally traded wood.

Target Group

Authorities, science, timber trade and consumers

Approach

For a selection of relevant hardwoods, samples from the scientific collection and other documented sources are processed into realistic training and test specimens. In the slide scanning microscope specially acquired for this project, 100 slides can be inserted simultaneously. Automated microscopic images are then taken in multiple focal planes over the entire slide area. These images represent the important data basis of the project. In these images, the cells that are essential for the identification of the wood genus by wood anatomists are labeled. Machine learning systems are trained and validated with these at the ITWM.

Microscopic images of fibrous materials can already be analyzed by the first prototype of a graphical user interface. The machine-learning system can currently only identify nine hardwoods with whose references it has been trained. Below you can see how the cells that are important for identification are first detected in the image and then classified. The cells identified by the system can then be checked for validity by wood anatomists.

Our Research Questions

Is it possible to identify wood species with machine learning?

Links and Downloads

Web page Fraunhofer ITWM

Online-Seminar „Tree stem and wood species identification using AI“ on 11 January, 2024

Literature about the project

0
Helmling S, Nieradzik L, Sieburg-Rockel IJ, Weibel T, Wrage S, Gospodnetic P, Keuper J, Stephani H, Rauhut M, Olbrich A (2024) Automated wood species identification in microscopic images of fibrous materials using machine learning / AI. In: Forests & society towards 2050 : 26th IUFRO World Congress, Stockholm, Sweden, 23-29 June 2024 ; Book of abstracts. p 3523
https://literatur.thuenen.de/digbib_extern/dn068687.pdf
1
Nieradzik L, Sieburg-Rockel IJ, Helmling S, Keuper J, Weibel T, Olbrich A, Stephani H (2024) Automating wood species detection and classification in microscopic images of fibrous materials with deep learning. Microsc Microanal 30(3):508-520, DOI:10.1093/mam/ozae038
2
Nieradzik L, Stephani H, Sieburg-Rockel IJ, Helmling S, Olbrich A, Keuper J (2024) Challenging the black box: A comprehensive evaluation of attribution maps of CNN applications in agriculture and forestry. In: Radeva P, Furnari A, Bouatouch K, Sousa AA (eds) Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2024). Vol. 2. pp 483-492, DOI:10.5220/0012363400003660
3
Nieradzik L, Stephani H, Sieburg-Rockel IJ, Helmling S, Olbrich A, Wrage S, Keuper J (2024) WoodYOLO: A novel object detector for wood species detection in microscopic images. Forests 15(11):1910, DOI:10.3390/f15111910
https://literatur.thuenen.de/digbib_extern/dn069063.pdf
4
Helmling S, Sieburg-Rockel IJ, Wrage S, Olbrich A, Nieradzik L, Stephani H, Weibel T, Gospodnetic P, Rauhut M (2023) Automatisierte Holzartenidentifizierung in mikroskopischen Bildern von Fasermaterialien mit Hilfe von maschinellem Lernen / KI. In: KIDA-Fachtagung, 27. - 28. September 2023, Quedlinburg : Abstractbuch. Braunschweig: Geschäftsstelle Think Tank Digitalisierung, Johann Heinrich von Thünen-Institut, pp 21-22