Comprehensive Surface Analysis Using PCA and Gradient Vector Flow Snakes for Defect Localization in Robotics Automation

Abstract

Because pipeline flaws provide such serious concerns, thorough and effective inspection methods are required. Although magnetic flux leakage (MFL) sensing yields useful information, it is frequently high-dimensional and noisy. Combining Gradient Vector Flow (GVF) Snakes for accurate edge detection with Principal Component Analysis (PCA) for dimensionality reduction in order to improve fault identification and localization accuracy. PCA is used to pre-process the MFL data, and then GVF Snakes is used to localize the defect boundaries and quantify the defect dimensions. With a 99.1% fault detection accuracy and quantification errors of less than 9%, the suggested approach demonstrated exceptional performance. Outperforming current methods, our scalable and interpretable methodology enhances problem localization and shows industrial relevance for semi-autonomous inspections.