Bridges, Tunnels and other Transportation Structures
Full Session with Abstracts
Surface distresses, usually caused by heavy vehicle traffic and severe weather conditions, are commonly observed in structural surfaces, such as pavement and bridge decks. Among the various types of surface distresses, much attention has been paid to cracking because of its particular harms to transportation safety and efficiency. Cracking not only leads to the deterioration of structural servicability, and also expedite further corrosion process of structural members by providing routes for corrosive agents (e.g., water).
Accurate crack detection and feature extraction are essential to ensure strcutural integratiry and funtionality. Manual inspection is the traditional survey method for crack detection, which is extremely time consuming, lacking of objectivity, prone to human errors, and risky to the inspection personnel due to traffic at high speed. Currently, an accurate and efficient crack surveying method is largely missing. Recently, novel techniques based on nondestructive testing and sensing methods, signal processing, and computer vision have greatly advanced frontiers in crack detection of civil infrastructure.
In this study, a vision based crack detection framework is presented. This framework is applied to the image data collected from typical bridge deck surfaces. It contains data collection, image interpretation and preprocession, and crack identification. The goal of the study is to offer a way to detect the cracks on structural surfaces automatically under image noises and undesired disturbances, and present the cracks in a form that can facilitate the extraction of essential information about the geometry of cracks. Through the interpretation and comparison of the results, the efficacy and accuracy of the detection results are demonstrated.