3D Modeling Reconstruction of Bridges Based on 2D Digital Images

Traditional bridge modeling methods often use 2D drawings for modeling, which is difficult to show the authenticity of the model. With the development of computer technology and the proposal of the concept of digital twin, the visual model is required to be closer to the entity. In this paper, an image 3D reconstruction method is proposed to establish the 3D model of the real bridge. Firstly, the image is collected around the target bridge, then the sparse point cloud model of the bridge is reconstructed by SFM(Structure From Motion) algorithm, the point cloud is densified by MVS(Multi-View Stereo) algorithm, and the model surface is constructed by Poisson Surface Reconstruction algorithm. Finally, the color and texture of the model are given by Texture Mapping algorithm to complete the reconstruction of the 3D model. The test shows that the maximum error of bridge reconstruction is only 3.3%, which can meet the needs of bridge integrity research.