摘要
针对输电铁塔塔头、塔身等不同部件之间差异较大,难以统一建模的问题,该文提出了一种基于空间结构特征分割和混合驱动的输电铁塔分区建模方法。首先,对铁塔重定向并采用基于空间结构特征的自适应分割方法,将铁塔分割为塔头和塔身,解决当前基于数据统计的分割方法易受噪声干扰和参数设置影响导致分割错误的问题。然后,利用数据驱动的方法重建塔身。当铁塔内部点云缺失时,采用基于横隔面约束和角度规则的方法重建横隔面和斜材。最后,构建新的塔头模型库,采用模型驱动的方法进行重建,并优化塔头三维方向的参数,解决塔头重建模型与原始点云匹配度低的问题。选择多种类型的铁塔点云进行实验,并与现有方法进行对比。结果表明,本文提出的方法分割精准度可达98.67%,塔身四棱框架重建精度为0.06m,横隔面重建精度为0.03m;塔头建模平均精度为0.18m,提高了0.04m,平均重建时间为6.68s,提高了11.66s。结果表明本文方法针对不同类型的铁塔点云数据均表现出较好的鲁棒性和自适应性。
In view of the large differences between different components such as the head and body of the transmission tower,a partition modeling method for transmission towers based on spatial structure feature segmentation and hybrid drive was proposed in this paper.Firstly,by reorienting the tower and adopting the adaptive segmentation method based on the spatial structure features,the tower was divided into the tower head and the tower body,which solves the problem that the current segmentation method based on data statistics is susceptible to noise interference and parameter setting.Then,the data-driven method was used to reconstruct the tower body,and the diaphragm and inclined material were constructed by using the diaphragm constraint and the method based on angle rule when the point cloud data inside the tower was missing.Finally,a new tower head model library was constructed,which was constructed by a model-driven method,and the parameters in the three-dimensional direction of the tower head were optimized to solve the problem that the tower head reconstruction model did not match the original point cloud with a low degree.Various types of transmission tower point cloud data were selected for experiments,and the proposed method was compared with the existing methods.The results showed that the segmentation accuracy of the proposed method can reach 98.67%,the construction accuracy of the four-sided frame of the tower body is O.06 m,the construction accuracy of the transverse surface is 0.03 m,the average modeling accuracy of the tower head is 0.18 m,which is increased by 0.04 m,and the average construction time is 6.68 s,which is increased by ll.66 s.The results showed that the proposed method shows good robustness and adaptability in the face of different tower LiDAR data.
作者
张正鹏
刘建华
谢欣余
胡天硕
刘树辉
ZHANG Zhengpeng;LIU Jianhua;XIE Xinyu;HU Tianshuo;LIU Shuhui(Xiangtan University,Xiangtan,Hunan 411105,China;Great Wall Information Co.,Ltd.,Changsha 410199,China;Beijing Hua Ke Zhi Xing Technology Ltd.,Beijing 102600,China)
出处
《测绘科学》
北大核心
2025年第8期134-146,共13页
Science of Surveying and Mapping
基金
湖南省科学技术厅面上项目(2022JJ30561)
湖南省研究生科研创新项目(QL20220161)
国家重点研发计划项目(2020YFA0713503)
关键词
LIDAR点云
空间结构特征
点云分割
混合驱动
铁塔分区建模
LiDAR point cloud
spatial structural features
point cloud segmentation
hybrid-driven approach
transmission tower partitioned modeling
