A comprehensive approach for detection and inspection of steel pipeline utilizing variational mode decomposition, clustering, and continuous wavelet transform

• The study introduces an innovative approach combining Variational Mode Decomposition (VMD), Density-Based Spatial Clustering with Noise (DBSCAN), and Continuous Wavelet Transform (CWT) to enhance pipeline inspection accuracy. • VMD is used to decompose magnetic signals into intrinsic mode functions (IMFs), significantly improving the signal-to-noise ratio, essential for accurate analysis in noisy environments. • DBSCAN clustering on the most relevant IMFs successfully identifies pipeline orientation and detects anomalies, providing better insights into the pipe’s condition. • Continuous Wavelet Transform (CWT) is applied to the clustered data, enabling precise defect localization in the time domain, which is critical for pipeline maintenance. • The combination of these techniques offers a robust and reliable methodology for detecting defects and determining pipeline orientation, even under challenging conditions. Pipelines are critical for the long-distance transfer of natural resources, but their integrity is frequently compromised by environmental variables, requiring regular inspections to ensure their safety. This study presents an innovative integrated methodology that incorporates Variational Mode Decomposition (VMD), Density-Based Spatial Clustering with Noise (DBSCAN), and Continuous Wavelet Transform (CWT) to significantly enhance signal-to-noise ratio, precisely determine pipeline orientation, and identify defects. Magnetic data from a 6-meter steel pipe with a known defect were evaluated and supplemented with simulated fracture defects. The proposed method improved SNR by 36 % over EMD and reduced defect localization error by ∼ 25 % when compared to straight CWT. It also avoided mode mixing, which is common in conventional approaches. Furthermore, DBSCAN adaptively finds clusters without requiring manual parameter modification, decreasing human intervention and increasing efficiency. These results show that the integrated VMD-DBSCAN-CWT framework provides a more accurate, resilient, and automated solution for pipeline inspection in noisy and complicated environments.