Automatic classification of magneto-optic image data for aircraft rivet inspection

Abstract

Magneto-optic imaging is a recently developed nondestructive evaluation technique that is widely used for detecting cracks and corrosion in aircraft skins. This paper investigates multiple definitions of skewness functions of magneto-optic image data for quantifying and classifying defects originating at rivet sites and thereby substantiating its usefulness in developing automatic defect classification system.