Research on Simplified Method for 3D Reconstruction of Grid Structures

Three-dimensional (3D) reconstruction technology plays a crucial role in various fields, especially in civil engineering, where it is essential for the monitoring and management of infrastructure such as buildings, bridges, and roads. This paper presents a simplified 3D reconstruction method specifically designed for large-span and high-precision grid structures. The method reduces computational complexity and enhances reconstruction accuracy by describing grid structure nodes with only four basic parameters: three spatial coordinates and a radius parameter. Utilizing image data captured by Unmanned Aerial Systems (UAS) and combined with deep learning algorithms, this study successfully achieved high-precision 3D reconstruction of a partial grid structure in a sports stadium in Hebei. Additionally, this paper reviews single-image 3D reconstruction methods based on deep learning, UAS-assisted 3D reconstruction techniques, and the application and challenges of 3D reconstruction technology in civil engineering. Finally, the paper discusses future research directions, including enhancing algorithm robustness, processing large-scale data, fusing multi-source data, and developing real-time reconstruction systems.