Characterization of pipeline defect in guided-waves based inspection through matching pursuit with the optimized dictionary

An optimized dictionary of matching pursuit (MP) is designed and combined with the ultrasonic guided wave for characterizing pipeline defect, which consequently achieves good performance in estimating the axial length of a defect. In the research field of guided wave based pipeline defect inspection, quantitative characterization on the detected defect remains as an unsolved research task and is seldom reported in the related publications. This paper reports a new method of MP equipped with an optimized dictionary through analyzing the interference between the reflection components embedded in a defect reflection signal in order to enable efficient extraction of defect information. The proposed method not only enhances signal-to-noise ratio of reflection signal, but also characterizes the axial length of pipeline defect directly and accurately. The method has been verified by simulated data, artificial defects in real pipes as well as real corrosion in a servicing pipe. © 2012 Elsevier Ltd.