Window Feature-Based Two-Stage Defect Identification Using Magnetic Flux Leakage Measurements

Abstract

Magnetic flux leakage (MFL) testing, one of the nondestructive testing methods, is widely adapted by approximately 90% of in-service pipelines. It is very important to identify defects in MFL testing. This paper presents a method to detect defects and determine precise defect regions using window features generated from MFL measurements. The main contributions are: 1) four novel window features of defects, namely, saliency, contrast, center point (CP), and fingerprint with parameters, are presented, which provide sufficient information for defect identification. In particular, a Bayesian method is developed to estimate the near-optimal parameters of these features and 2) a novel two-stage identification process is proposed, which can not only detect defects, but also segment the complicated defect regions precisely. The performance of the proposed method is demonstrated by real MFL measurements collected from experimental pipelines and in-service pipelines, respectively. The results show that the proposed method has satisfied accuracy for defect identification of experimental and engineering application.