Article - AdheScan − adhesive failure surface inspection system

Adhesive failure surface inspection based on machine learning of
high-resolution stereo images
by Luca Kuhlmann, André Lamott, Armin Thomsen, Heiko Mühlenfeld-Keßler und Anja Knigge

AdheScan finally provides valuable and reproducible data for bonding professionals as a basis for further research and development. Based on high-resolution images and height information, an expert can mark the identifiable fracture patterns. An algorithm is then trained using these annotations and further inspection is quick and easy. A prediction mode was developed for laboratory use when dealingwith very frequently changing substrate/adhesive samples. The large number of samples required in this industry can now be evaluated in an efficient, reproducible and accurate manner. [more]