Defect detection on air bearing surface with gray level co-occurrence matrix

Abstract

Air bearing surface (ABS) is the part of magnetic read/write head flying height controller. It is very important part in magnetic disk (hard disk drive), defected on ABS lead to crash between read/write head and disk surface, therefore its verifying is necessary. The best way to verify defect on ABS is machine vision. Main problem of machine vision in real world is variation of luminance intensity that affects image acquisition. This research proposes method for detecting defect on ABS which has variance luminance intensity, the Co-Occurrence Matrix is used to avoid the variance intensity of ABS image then feature parameter is defined by four identification features and defected detect by threshold that selected from Euclidean distance of each identification. The experimental results show very low error of defect detection on ABS by Co-Occurrence matrix and their identification feature.