TY - JOUR
T1 - Semi-automatic crack width measurement using an OrthoBoundary algorithm
AU - Li, Zhe
AU - Miao, Yi
AU - Eskandari Torbaghan, Mehran
AU - Zhang, Hongfei
AU - Zhang, Jiupeng
N1 - Acknowledgments:
This study was supported by the National Natural Science Foundation of China (No. 51978068), Natural Science Foundation of Shaanxi Province (No. 2020JM-217). The authors gratefully acknowledge their financial support. The first author would also like to acknowledge the China Scholarship Council (CSC) for supporting his visit to University of Birmingham, UK (No. 202106560034).
PY - 2024/2
Y1 - 2024/2
N2 - Evaluation of pavements' crack severity levels currently relies heavily on width measurement, which necessitates the development of a rapid, and high-accurate, automatic measurement approach for complex pavement cracks. This paper presents an OrthoBoundary algorithm that leverages the crack boundary and skeleton directions to determine crack propagation. Comparative analysis has been conducted between OrthoBoundary and Area-Length, Skeleton Shortest Distance (SSD), Edge Shortest Distance (ESD), and Orthogonal Projection (OP) methods. Results indicate that the OrthoBoundary algorithm achieves an average accuracy of 90.10%, outperforming the Area-Length (86.60%), SSD (76.01%), ESD (87.24%), and OP (88.07%) methods. Notably, the OrthoBoundary algorithm also exhibits processing speeds approximately 120 times faster than other considered methods while demonstrating improved robustness and user-friendliness. It has significant potential to quantify and assess the severity of pavement cracks, as well as to facilitate maintenance decision-making processes in road infrastructure management systems.
AB - Evaluation of pavements' crack severity levels currently relies heavily on width measurement, which necessitates the development of a rapid, and high-accurate, automatic measurement approach for complex pavement cracks. This paper presents an OrthoBoundary algorithm that leverages the crack boundary and skeleton directions to determine crack propagation. Comparative analysis has been conducted between OrthoBoundary and Area-Length, Skeleton Shortest Distance (SSD), Edge Shortest Distance (ESD), and Orthogonal Projection (OP) methods. Results indicate that the OrthoBoundary algorithm achieves an average accuracy of 90.10%, outperforming the Area-Length (86.60%), SSD (76.01%), ESD (87.24%), and OP (88.07%) methods. Notably, the OrthoBoundary algorithm also exhibits processing speeds approximately 120 times faster than other considered methods while demonstrating improved robustness and user-friendliness. It has significant potential to quantify and assess the severity of pavement cracks, as well as to facilitate maintenance decision-making processes in road infrastructure management systems.
KW - Pavement crack
KW - Width measurement
KW - Orthogonal projection
KW - Principal component analysis
KW - Skeleton
U2 - 10.1016/j.autcon.2023.105251
DO - 10.1016/j.autcon.2023.105251
M3 - Article
SN - 0926-5805
VL - 158
JO - Automation in Construction
JF - Automation in Construction
M1 - 105251
ER -