Intelligent Measurement of Length and Width of Long Concrete Cracks on Bridges Using 3D Reconstruction of Binocular Vision

Accurate measurement of bridge cracks is essential for assessing structural safety and durability. In recent years, binocular vision technology has been widely used in the size measurement of concrete cracks due to its noncontact and highly efficient advantages. However, current crack size measurement has shortcomings in dense point cloud modeling and long crack splicing, resulting in inaccurate measurement. Hence, this paper proposed an intelligent measurement of concrete crack length and width using three-dimensional (3D) reconstruction of binocular vision. For length measurement, a cumulative summation method of 3D interpoint distances in the crack centerline is proposed, as well as considering the splicing of a long crack that cannot be completely captured in a single image; for width measurement, a method of calculating the shortest distance from the crack centerline to crack boundary lines in the 3D point cloud is proposed. During the cost-matching phase in the 3D reconstruction of concrete cracks, the census algorithm is selected as the optimal cost calculation method. During the splicing period of a long crack for iterative closest point (ICP) splicing, the SURF algorithm is selected as the optimal feature matching algorithm, and the relative-distance-square-normalization weight is selected as the optimal calculation function for ICP weight parameters. In the field test, the relative error of the whole crack length is 4.2%, the average error between the calculated and measured crack widths is 0.15 mm, and the relative average error is 4.28%, which verifies high accuracy of the intelligent measurement of concrete crack length and width using 3D reconstruction of binocular vision, providing robust support for subsequent bridge management and maintenance.