Wood products identification by internal characteristics reading

Cecilia Fuentealba (1) , Christophe Simon (1) , Denise Choffel (1) , Patrick Charpentier (1) , Daniel Masson (2)

1 CRAN - Centre de Recherche en Automatique de Nancy
2 LERMAB - Laboratoire d’Etude et de Recherche sur le Matériau Bois

The traceability of products is an ever growing demand of quality for many areas. In the manufacturing process, traceability is required to follow-up all the production cycle of the product. Consequently, product traceability can result in a great control of the production process, thus allows improving control of production tools and reacting to their deficiencies. For wood industries, implementation of current identification techniques is not easy, due to the extremely variable nature of the wood and the particular features of the manufacturing process. In this article, we implement a traceability tool within a manufacturing process using the intrinsic characteristics of the wood product. The goal is to achieve an individual identification of products. We use a microwave sensor to obtain a unique signature of the product and then we develop its identification. Thus, the identification process on the basis of a numerical product signature can be considered as a discrimination problem as for the nearest neighbor method. We evaluate the error rate of identification by a Monte Carlo simulation of the process. This work shows the high potential of developing an identification system based on the use of the intrinsic characteristics of product.

Mots clés: wood product; automatic identification; traceability; k-nearest neighbors; pattern recognition

Domaines: Automatique / Robotique

Cecilia Fuentealba, Christophe Simon, Denise Choffel, Patrick Charpentier, Daniel Masson. Wood products identification by internal characteristics reading. IEEE ICIT '04, 2004 IEEE International Conference on Industrial Technology, Dec 2004, Hammamet, Tunisia. pp.763- 768, ⟨10.1109/ICIT.2004.1490171⟩. ⟨hal-00154536⟩