New developed testing system of defect in cementitious material with Neural Network

A developed testing system of defect detection in mortar is introduced. Base on this development, more information of the defection could be obtained. Machine Learning algorithm, named Artificial Neural Network (ANN)with pattern recognition, is used to verify the results. As a result, the medium with and without defects could be identified. More over, the number of defects could be indicated. This study shows the methodology for implementing the system, factors to be considered, and verifying the results. Classification examples are given to demonstrate the performance of the network by using the data set from different mediums. The result shows the effectiveness of this classifier and possibility of further development.