Characterization of Magneto-Optic Imaging Data for Aircraft Inspection

The magneto-optic imager (MOI) is widely used in detecting surface and subsurface cracks and corrosion in aircraft skins. The instrument provides analog images of the anomalies based on eddy current induction techniques and a magneto-optic (MO) sensor using the Faraday rotation effect. The merits of the MOI that make it attractive include rapid and large-area inspection, insensitivity to liftoff variations, and easy interpretation in contrast to the complex impedance data of conventional eddy current inspections. The current MOI system lacks the capability of providing a quantitative measure of the defects. In addition, the presence of noise due to the underlying domain structures in the MO sensor can lead to inconsistent accept or reject decisions by the inspector. This paper presents an image processing and automated classification algorithm for MO image analysis and also provides a quantitative basis for characterizing these images