Ultrasonic data compression and analysis during material inspection

Abstract

Ultrasonic inspection is nowadays widely used in material inspection to detect and locate internal defects. However, in case of large structures inspection, such as pipelines, vessel reactors etc., the amount of signals to acquire and interpret may be very important, especially in modern techniques that use image representations. In this article, a proposed algorithm to ultrasonic data selection and processing is described. Images will be replaced by 2 arrays that contain only the data regarding a detected defect. This allows avoiding the storage of useless received ultrasonic data, and simple pattern recognition techniques may be exploited to defect characterization.