Weldment defect detection and classification in ultrasonic testing using artificial intelligence

Abstract

Ultrasound based inspection techniques are used extensively newlinethroughout industry for detection of flaws in engineering materials The newlineprincipal goal for ultrasonic inspection of engineering materials is the newlinedetection location and classification of internal flaws and defects as quickly newlineand as accurately as possible However this non destructive testing process is newlineoften difficult and time consuming and may well rely significantly on the skill newlineand experience of the tester The combination of the human eye and brain is newlineuniquely capable after training of classifying a wide range and variety of newlinecomplex patterns Performance is however subject to significant variation as a newlineresult of factors such as fatigue and loss of concentration newline newline