Depth estimation of steel cracks using laser and image processing techniques

Abstract

Crack detection is needed to maintain safety and its automation is a must when human-based inspection cannot be made due to space limitations and/or hazards. In this study, an automatic crack depth measurement method using image processing and laser methods is developed. Measurement of maximum actual depths is done using Keyence (VK-X100) laser microscope. Microscope capture crack image segments using 1/3 in. (8.5 mm) sensor color charge-coupled device (CCD) camera with high resolution and 10x constant magnification. Depths are also calculated using the updated Make3D toolbox. Measured and calculated depths are compared for 11 cracked specimens with 105 segments. The comparison showed that the minimum and overall average error between measured and calculated depths are about 6.13% and 28.22% respectively.