Research on Early Corrosion Evaluation of Carbon Steel Based on Eddy Current Pulsed Thermography and Multiple Feature Characterization Methods
摘要
Early corrosion of coated steel poses a significant threat to the performance of steel structures. This study investigates the use of Eddy Current Pulsed Thermography (ECPT) for detecting corrosion in coated steel. We propose a method utilizing Bidimensional Empirical Mode Decomposition (BEMD) to enhance the infrared thermal image sequences, improving the quality of the raw images by reducing noise and preserving essential details. Fast Fourier Transform (FFT), Principal Component Analysis (PCA), and skewness feature extraction methods were employed for the detection and evaluation of early-stage corrosion. The differential skewness feature demonstrated the best detection sensitivity. Additionally, the phase image at 2.07 Hz and the reconstructed image using the second principal component effectively highlighted the corrosion details. ECPT technique is proved as a smart sensor system for paint-coated corrosion detection and characterization.