Defects detection of gas pipeline near the welds based on self quotient image and discrete cosine transform

Abstract

The magnetic flux leakage (MFL) inspection has been widely used in the inline inspection for the evaluation of steel pipelines and plates. In this paper, a defect detection algorithm based on the MFL inspection is proposed for detecting defects near the welds. The defect in this paper means some deformations and deterioration of steel pipes because of corrosions and cracks by humidity and pressure after gas pipes were buried and it doesn’t mean bad welding. The MFL signal of the defects near the welds is worse than that of the far-away from the welds and has low signal-to-noise ratio (SNR). In this paper, the MFL signal of the defects near the welds is enhanced by using the Self Quotient Image (SQI) in this paper and the position of the defects is detected after applying the Discrete Cosine Transform (DCT). Experiments are conducted to demonstrate the accuracy of the proposed defect detection algorithm for the artificial defects carved on the pipes at the pipeline simulation facility (PSF) and the results show that the proposed algorithm can successfully detect the position of the defects on the pipes near the welds.