Using Deep Machine Learning to detect Cracks in Concrete

Using different machine learning strategies to detect cracks in images of concrete

From rather blurry pictures of the fundament of the Öresund Bridge, the project started with finding ways to detect which images represented cracks. The blurry pictures introduces challenges that may be hard to overcome. Different kinds of techniches are used here, such as, augmentation, supplementary datasets, and Generative adversarial networks.

Three conference presentations and publications so far, all in ieee-based conferences.
We are expecting further similar results from the new, and future, studies on Generative AI.