Unmanned Aerial Vehicle Based Pavement Crack Identification Method Integrated with Geographic Information Systems

Building of an efficient pavement crack identification system is the challenge of developing a modern pavement management system (PMS). The aim of a PMS is to have autonomous inspections that work fast, accurate and robust when inspecting the in-service pavements compared to currently used ones which are labor intensive, biased and time consuming. Recently, the use of unmanned aerial vehicles (UAVs) increased in monitoring of civil infrastructures. In this study, UAVs were used to capture images of the pavement surface, from which pavement cracks were identified using certain features and a machine learning algorithm called Support Vector Machines (SVM). Preliminary operations were performed to make the images uniformly illuminated and noise free for segmentation. Comparatively darker features were obtained using automated thresholding algorithm. Image operations yield regions for potential cracks, categorized into four classes: transverse, longitudinal, alligator and non-crack. Crack candidate regions are classified according to their features using a SVM model. Geographic Information System (GIS) integration fulfills the last step of the proposed system by positioning classified crack images into Google Maps. A case study was performed to measure the performance of the method where 181 images were taken along a highly damaged pavement. The prediction results were quite successful. This system has a few benefits such as being cheap and computationally efficient when identifying multiple cracks in one image. Further work is on its way for identification of other types of distresses or implementing other tools into the system such as UAV mounted laser scanners.
Citation Formats
A. B. Ersöz, O. Pekcan, and T. Teke, “Unmanned Aerial Vehicle Based Pavement Crack Identification Method Integrated with Geographic Information Systems,” presented at the Transportation Research Board 96th Annual Meeting, (8 - 12 Ocak 2017), Washington, DC. USA, 2017, Accessed: 00, 2021. [Online]. Available: https://trid.trb.org/view.aspx?id=1438958.