Automatic 3D Modeling Technique for Transmission Towers from 2D Drawings
Ziqiang Tang,Chao Han,Hongwu Li,Zhou Fan,Ke Sun,Yuntian Huang,Yuxin Chen,Chenxing Wang
DOI: https://doi.org/10.3390/math12233767
IF: 2.4
2024-12-01
Mathematics
Abstract:The 3D modeling of transmission towers currently depends on manual operations, resulting in high labor and time costs. To this end, an automatic 3D modeling technique based on 2D drawings is proposed. Using this method, the 2D drawings of transmission towers were analyzed first, then a 3D model of a tower was reconstructed using a counter-to-detail strategy. The analysis of the 2D drawings aimed to segment the geometric shapes and subsequently extract the vectors. All obtained vectors were classified into outer contour vectors and internal structure vectors. For each tower section, the 3D outer contour framework was constructed first using the wireframe model algorithm, followed by the assembly of internal details onto the 3D contour framework to fully reconstruct the 3D model. Experiments demonstrated that constructed 3D models exhibited high accuracy, with an average chamfer distance to the real scanned dense LiDAR point clouds of less than 0.05 m, which was less than 1% relative to the whole size of the created models. Furthermore, the automation of this technique implies its potential for various applications.
mathematics