This project is an on-going research based project on the the classification of cracks in the concrete walls. There are four types of cracks identified by sophisticated tools. These tools use acoustic signals reflected from the concrete surface to detect the damage. These hidden cracks are normally not visible and are identified through equipments.
In this research effort, we train the Artificial Neural Networks to classify such cracks based on acoustic characteristics. Level I cracks are the smallest ones whereas the level IV cracks are the largest. Each type of cracks have their unique acoustic crack signatures which are identified from the reflected signals. Hence level I cracks have minimum acoustic signatures.
Accompanied Journal research article (https://github.com/jadoonengr/Civil-Concrete-Damage/blob/master/Acoustic%20Emissions%20Relevant%20Paper.pdf) highlights the original effort. The research article related to this neural network based approach will soon be published.