​Natural cork stoppers are considered the gold standard by wine enthusiasts. So, why do some wines end up corked or oxidized? For wine to be conserved properly, the cork must let exactly the right amount of oxygen into the bottle for the the wine to mature and for its flavor to continue to develop over time. However, there is no way to tell how much oxygen a cork will let in in simply by looking at it.

Cork manufacturer Diam Bouchage partnered with CEA-List to come up with a reliable automated sorting solution for its corks manufactured by punching natural cork oak bark. The process begins with X-ray tomography images of the entire volume of the cork. The researchers applied a machine learning algorithm suitable for classification problems to the images obtained. The cork characteristics (number and distribution of lenticels, growth lines, cork density, etc.) in the X-ray tomography images were cross-referenced with existing long-term oxygen transfer data, to create a model that can automatically estimate how airtight a cork will be, providing a reliable indicator of cork quality.

The solution significantly improved cork sorting, correctly classifying 75% of the corks inspected in just seconds. The method requires only two images per cork, making it suitable for automated in-line non-destructive testing of natural corks. The researchers are now setting their sights on a 100% correct classification rate, a prerequisite for use with fine wines.