An Approach to the Detection of Post-seismic Structural Damage Based on Image Segmentation Methods

Crack detection is critical in ensuring basic structural security, however manual identification of cracks is time-consuming and is subject to the judgments of reviewers. This research presents a crack detection technique based on image processing. The digital image processing is divided into different phases and each of them follow techniques that improve the quality of the images. In the segmentation phase, images traits need to be highlighted. This document portrays the image segmentation of a set of digital photographs of cracks and crevices of the different structures of the buildings of the faculties of the University of Guayaquil. In this study, a function is developed using the computational tool, Matlab, to obtain results by submitting the images to the different segmentation techniques applied during the investigation, for which methods are proposed such as: The Canny transform, The Sobel Operator and the Prewitt Transform. With the obtained results, crack measurement is applied based on the manual selection of pixels in order to generate damage assessment.