Research on 3DZI technology and its application in apparent inspection of concrete continuous beam bridge

To address the inefficiency, safety risks, and limited accuracy associated with traditional visual inspections of concrete continuous beam bridges, this study proposes a novel inspection method called 3DZI (3D Reality and Deep Zoom Image). This approach integrates 3D reconstruction techniques with panoramic image unfolding to enable comprehensive and accurate bridge appearance inspections. The process begins with UAVs equipped with RTK positioning capturing high-resolution images of various bridge components, including the deck, beams, piers, and underside. These images are used to reconstruct detailed 3D point cloud models through Structure from Motion (SfM) algorithms based on image overlaps. Subsequently, parameterized gridded design models are created for each structural component. By intersecting these gridded models with the reconstructed 3D model using normal equations, surface deviation data is extracted. This deviation data serves as the basis for generating distortion-free, high-resolution panoramic images that accurately reflect the condition of the bridge’s surfaces. Furthermore, a mapping relationship between the 3D models and the 2D panoramic images is established, enabling seamless navigation and integration of inspection data. This integration supports systematic defect detection—such as cracks and spalling—and facilitates longterm bridge health monitoring. Field tests conducted on the Jiujiang Bridge demonstrated that the 3DZI method significantly improves the completeness and accuracy of inspection data. It enables efficient defect analysis and contributes to more informed, data-driven decision-making in bridge maintenance and management.