Fusion of Multimodal Nde Data for Improved Corrosion Detection

Corrosion is a significant maintenance problem, especially for aging aircrafts, with the Air Force recently estimating corrosion maintenance costs exceeding $800M per year. Nondestructive evaluation (NDE) techniques can be used to detect corrosion under paint and hidden within layered structures. However, it is difficult to capture a comprehensive picture of the corrosion environment using any single NDE technique. This research presents a data fusion approach for combining raw multimodal NDE data for detecting surface and subsurface corrosion, which facilitates limiting false corrosion detections. Experimental results are presented using NDE techniques, including microwave, eddy current, ultrasound, and radiography for a lap-joint mimic.