基于点云数据的变电站三维建模方法研究

Research on Substation 3D Modeling Method Based on Point Cloud Data

为提高变电站后期监测与运维管理能力,运用三维激光扫描技术,构建基于点云数据的变电站三维模型。首先,运用三维扫描设备采集变电站点云数据,通过概率统计方法、自适应密度体素滤波算法(adaptive density voxel filtering algorithm,ADVFA)及基于密度的空间聚类算法(density-based spatial clustering of applications with noise,DBSCAN)对变电站点云数据集进行顶部与地面切除、压缩和去噪预处理,减少数据运行的难度与时间。其次,构建变电站电气设备点云数据快速提取方法和匹配模型,由特征编码、特征解码和特征学习3个模块组成邻域特征聚合方法对预处理后的电气设备点云数据进行分割提取,基于图像识别的快速匹配方法,完成分类设备点云与模型库设备种类的相对匹配,提高电气模型精细匹配效率。最后,通过算例分析验证所提出的基于点云数据的变电站三维建模方法的有效性和合理性。

To enhance monitoring quality and further improve operation management level of substation in the later stage,three-dimensional laser scanning technology based on point cloud data is applied to build a three-dimensional substation model.Firstly,substation point cloud data are collected through three-dimensional scanning equipment,and then probability statistics method,adaptive density voxel filtering algorithm and density-based spatial clustering of applications with noise algorithm are used to preprocess top and ground cutting,compression and denoising of substation point cloud data set,which can make computing less challenging and reduce the time of data process.Secondly,the rapid extraction method and matching model of substation electrical equipment point cloud data are established.The neighborhood feature aggregation method,which is composed of three modules:feature coding,feature decoding and feature learning,can segment and extract the preprocessed point cloud data for electrical equipment.The rapid matching method based on image recognition can match the point cloud of classification equipment and the type of model library equipment,and thus improve the precision matching efficiency of electrical models.Finally,an example analysis results prove the validity and rationality of the proposed substation 3D modeling method based on point cloud data.