Automatic visual identification of defects by magnetic field imaging using alternating current field measurement technique

Traditional electromagnetic testing methods rely on manual analysis of signals, resulting in low visual clarity and a propensity for human error. Therefore, this paper proposes an automatic defect recognition method based on alternating current field measurement (ACFM) technology, which achieves visual identification of defects by constructing images of distorted magnetic fields caused by the defects. The method integrates magnetic field interpolation imaging technology with image processing algorithms to improve visualization and enable automated defect recognition. First, an ACFM simulation model is established to analyze the relationship between the magnetic induction perpendicular to the specimen surface and the disturbances in the induced current at defect edges, thereby revealing the signal distribution patterns. On this basis, a magnetic field interpolation imaging technique is proposed to transform one-dimensional testing curves into two-dimensional magnetic field distribution images, and visual algorithms are incorporated to design automatic defect recognition rules. Defect quantification is achieved based on hue extremum analysis of the distorted magnetic field regions. Finally, the effectiveness of the method is verified on an ACFM experimental platform. The experimental results demonstrate robust multi-defect detection capability, achieving localization errors below 4 % and maximum recognition errors of less than 6 % for defect length. For depth estimation, the system leverages the Hue component of the magnetic field image, yielding a correlation coefficient exceeding 0.99 between the extracted Hue feature and the actual defect depth. The system also accurately resolves crack orientation and characterizes corrosion morphology, with all recognition errors maintained below 5 %. The proposed magnetic field image-based visual processing rules enable automated and highly accurate quantification of defect parameters, validating the method’s effectiveness in practical scenarios.