In partnership with DIAM Bouchage, CEA Tech's CEA List institute has developed a technology based on artificial intelligence analysis of X-ray images to detect faulty corks quickly and accurately.
Although cork stoppers are often considered ideal for long-term wine conservation, sometimes a wine turns out to be corked or oxidized when a bottle is opened... To perform its function correctly, a cork must let in just enough oxygen to both allow the wine to mature and guarantee its taste quality over time. However, it is impossible to assess the tightness of a cork with the naked eye.
Diam Bouchage called on the Carnot CEA List institute to develop an automated, reliable sorting method for cork stoppers produced by tubing cork oak bark. Selection is based on X-ray tomography images, which allow the cork to be viewed in its entire volume. To analyze these images, the researchers implemented an automatic learning algorithm adapted to classification problems. By comparing the characteristics obtained by imaging (number and distribution of lenticels, growth lines, cork density, etc.) with long-term data on the rate of oxygen transfer, they have set up a model that automatically assesses cork tightness, a guarantee of cork quality.
The researchers succeeded in significantly improving cork classification, with 75 % corks correctly classified in just a few seconds. Requiring just two images per cork, this method could be used to deploy automated in-line non-destructive testing of natural corks. Work is continuing with a view to achieving 100% of correct classification, a mandatory criterion for the corking of fine wines.
Visit the site:
www-list.cea.fr