Automated System for Visual Non-Destructive Testing

an intelligent defect recognition system based on artificial neural network (ANN) is proposed. A convolutional neural network (CNN) has been selected based on the analysis. To achieve this goal, the following has been carried out: creation of an image database for network training; network architecture development; implementation of a program in MATLAB. The study has been carried out for various objects of non-destructive testing. The result of the testing system work was assigning to one of the defects classes and determining whether there is a defect or not. A technique for automated visual non-destructive testing based on artificial neural network has been developed.