Corrosion defect extraction in ferromagnetic materials based on array pulsed eddy current

Pulsed eddy current testing (PECT) is an efficient non-destructive electromagnetic testing method that is widely used in industrial fields for defect identification and metal corrosion detection. However, conventional methods often lack an accurate defect extraction strategy after detection imaging. In this study, the relationship between the slope of the attenuation phase (SAP) of the semilogarithmic transformed detection voltage and the specimen thickness is established using a single probe. Subsequently, an array sensor is employed to scan ferromagnetic specimens with circular hole defects. A 4 × 17 sampling matrix is constructed and visualised using pseudo-colour imaging for intuitive defect visualisation. To improve resolution, bicubic interpolation is applied. Three segmentation algorithms, Otsu, Gaussian mixture model (GMM) and double-threshold region growing (DTRG), are applied for defect extraction. The experimental results demonstrate that the proposed method enables accurate defect imaging, with the DTRG algorithm achieving a detection error below 5%. This approach offers an effective solution for corrosion detection in cladding pipes and pressure vessels, showing strong potential for industrial applications.