Crack detection by a climbing robot using image analysis

Abstract

Cracking on surfaces as walls or roofs of a building results in a rapid deterioration of such structures. Access to these places is sometimes very difficult for people. One approach to solve such inconvenient is the use of climbing robots provided with sensing devices. In this paper, we propose a computer vision system using image analysis for inspection. A stereo camera is mounted on the thorax of a hexapod robot, named Hex-piderix. The images are processed using a method described in this document. It is a method invariant to illumination to detect cracks in the surface. Image is improved through the estimation and removal of lighting pattern, and then a thresholding is applied using Otsu method. Finally morphological operations are applied to extract crack information.