JAUE2024-081: Measurement of Concrete Cracks in Structures Lacking Reference Objects and Feature Points

Abstract

Information about the location of cracks in concrete structures is a crucial factor that enables appropriate maintenance or reinforcement measures. Most concrete crack-related research is limited to crack detection and identification, with a lack of studies related to crack location information. The originality of this study lies in developing application technology related to crack localization by proposing a methodology that can estimate the location of concrete cracks using unmanned aerial vehicles and image processing technology, even in cases where there is a lack of reference objects or feature points. To develop and validate the proposed method, aerial photography and image acquisition were performed by mounting a laser pointer model on an unmanned aerial vehicle. To construct analysis data, image distortion correction and feature point extraction were performed using homography matrices and scale-invariant feature transform algorithms. Spatial information was constructed using point cloud techniques and image stitching techniques, and crack location was estimated by generating crack representation data through hierarchical merging. The proposed method was validated through comparison with field measurement data. In the future, the proposed methodology can be used to complement and improve existing methods for visual inspection of infrastructure and facilities.