Skip to main content Skip to navigation
Washington State University National Center for Transportation Infrastructure Durability & Life-Extension (TriDurLE)

Project

Automated Detection of Characterization of Cracks Using Structure-From-Motion Based Photogrammetry: A Feasibililty Study

Team

Dr. Xiong Zhang
Missouri University of Science & Technology

Description

In infrastructure such as pavement, bridges and tunnels, crack widths and patterns on surfaces are two of the most important signs used to estimate durability. Conventional techniques suffer from challenges such as tediousness, subjectivity, and high cost. A new measurement technique that overcomes these challenges while measuring crack displacement with high accuracy and low cost in aging structures is needed. The research will develop a Structure-from-Motion Based photogrammetry technique for measuring crack widths and patterns using videos taken by commercially available low cost digital cameras. Software will be developed to analyze the videos by combining deep-learning techniques and modern close-range photogrammetry. 3D models of the pavement and bridge structures with high accuracy will be constructed using the videos and will be compared and validated using the results generated from high accuracy LiDAR system. Post-processing algorithms will be developed to automatically calculate the real lengths as well as the real width and depth of a crack at any arbitrary locations. This method for 3D crack mapping will provide us a high accuracy, low cost, and easy-to-operate tool for pavement and bridge management.

More Detail

Visual