CRACKVIEW AI enables automated detection of MT indications with the help of artificial intelligence
The CRACKVIEW AI system digitises magnetic particle testing. The surfaces of the test parts prepared with magnetic particle are photographed using industrial cameras. The images of the surfaces are then analysed by a neural network (artificial intelligence / AI) trained by KARL DEUTSCH. It carries out a semantic segmentation of the images and determines the shape, position and number of possible crack indications with the associated crack probability. Thus, the components can be sorted to OK / NOT OK without a human inspector having to view them. The system significantly reduces inspection costs and delivers consistent inspection quality 24/7, all year round. In addition, the individual test results of the components can be documented, archived and traced for the first time.