Construction Quality Monitoring of High-Pier Bridges Based on Highly Accurate Three-Dimensional Reconstruction

Abstract

This study proposes an unmanned aerial vehicle (UAV) route-planning and three-dimensional (3D) reconstruction modeling method to achieve intelligent monitoring of the quality of the high pier. The all-around image data acquisition can be achieved by a UAV combining a fixed-point surround with a layered flight image data-collection strategy. The key pixel points are mapped to the 3D space by the coordinate transformation technique after extracting the target features from the image. Subsequently, an irregular triangular mesh reconstruction method is employed to generate a refined 3D model, enabling visual display and multi-angle analysis of the high pier. As a result of the field tests, the heading and side overlap rate of the images reached 75.8% and 84.2%, respectively, after the UAV completed the set route-planning task. The R_k value of the generated 3D model of the bridge pier was less than 1.0 after coordinate transformation and 3D reconstruction, which confirms the high modeling accuracy under the set UAV route. The error rate of the 3D reconstruction model was controlled below 0.2%, which proved the high applicability and accuracy of the proposed method.