Wishing you all a Happy New Year from the IMOS Lab – EPFL! ✨
For this new year we are excited to share with you our new paper: “Non-contact sensing for anomaly detection in wind turbine blades: A focus-SVDD with complex-valued auto-encoder approach”
In this work we propose a new manufacturing defects detection pipeline for monolithic materials using a non-invasive inspection method: the Frequency Modulated Continuous Wave (FMCW) radar. We enhance the quality assurance utilising FMCW radar by disentangling material-specific and round-trip delay information from the received wave. A focus Support Vector Data Description (focus-SVDD) is used to remove healthy data features, thereby focusing on the attributes of anomalies. We employs a complex-valued autoencoder and introduces a new activation function called Exponential Amplitude Decay (EAD) to improve the healthy features extraction of the complex-valued signal input.
This work comes from an international collaboration between Gaetan Frusque and Olga Fink from Ecole polytechnique fédérale de Lausanne and Daniel Mitchell, Jamie Blanche and David Flynn from University of Glasgow!
The paper is available here:
https://lnkd.in/e2Gy6f37
