SHAPE CLASSIFICATION OF FLAW INDICATIONS IN 3-DIMENSIONAL ULTRASONIC IMAGES

The rapid evolution of computing hardware technology now allows sophis ticated software techniques to be employed which will aid the NDT data interpreter in the process of defect detection and classification. Th e paper describes an investigation into the area of three-dimensional ultrasonic image evaluation and, more specifically, the problem of cha racterising the shape of suspect flaw regions. A backpropagation neura l network is used as the classifier for a series of four three-dimensi onal feature extraction methods which are individually assessed on two particular recognition problems. The optimum technique was determined for inclusion in an evaluation environment called the NDT Workbench, which has been designed for the processing of real data. Two acquired ultrasonic data sets are assessed using the best-performing classifica tion method.