Page 16 - 电力与能源2023年第二期
P. 16

第 44 卷 第 2 期
                                                         电力与能源
                  110                                                                          2023 年 4 月

                                                                                       DOI:10.11973/dlyny202302003

                    基于多分段线性拟合的输电导线快速提取方法




                                    束庆霏 ,张纳川 ,蔡佳澄 ,何杨阳 ,方                            远  3
                                            1
                                                                  1
                                                                             1
                                                       2
                                  (1. 国网江苏省电力有限公司张家港市供电分公司,江苏 张家港 215600;
                                     2. 国网江苏省电力有限公司苏州供电分公司,江苏 苏州 215000;
                                  3. 国网江苏省电力有限公司苏州市吴江区供电分公司,江苏 苏州 215200)
                    摘   要:为从输电线路的激光点云中快速确定导线位置,通过降采样、高程滤波处理和多分段线性拟合方法,
                    确定了导线在 CGCS2000 坐标系平面内的分段线性方程。基于张家港市电网两条 220 kV 输电线路的激光
                    点云数据,算法结果如下:综合考虑时间成本和采样效果,体素下采样更适合导线提取算法;多分段线性拟合
                    效果较好,个别噪声点不影响拟合结果,平均每段线路可在 10 s 内完成拟合;对于非函数性质的线路分布,可
                    转换 x 和 y 坐标,或分段拟合。输电导线提取算法适用于平原地区的输电线路,可快速确定冗杂激光点云
                    (3 亿多点)中的导线位置,有助于后续无人机自主巡线的航迹规划。
                    关键词:激光点云;输电导线;降采样;高程滤波;多分段线性拟合
                    作者简介:束庆霏(1995—),男,硕士,主要从事无人机智能巡检工作。
                    中图分类号:TM75    文献标志码:A    文章编号:2095-1256(2023)02-0110-05
                 A Fast Extraction Method of Transmission Wire Based on Multi-piecewise Linear Fitting



                                                                                    1
                                                        2
                                                                      1
                                      1



                           SHU Qingfei ,ZHANG Nachuan ,CAI Jiacheng ,HE Yangyang ,FANG Yuan      3

                             (1.Zhangjiagang Power Supply Company,State Grid Jiangsu Electric Power Co.,Ltd.,


                                            Zhangjiagang 215600,Jiangsu Province,China;


                2.Suzhou Power Supply Company,State Grid Jiangsu Electric Power Co.,Ltd.,Suzhou 215000,Jiangsu Province,China;





                3.Wujiang Power Supply Company,State Grid Jiangsu Electric Power Co.,Ltd.,Suzhou 215200,Jiangsu Province,China)







                    Abstract:In order to quickly locate the conductor from the laser point cloud of the transmission line,the piece‑
                    wise linear equations of the conductor in the coordinate plane of CGCS2000 is determined by down-sampling,el‑

                    evation  filtering and multi-piecewise linear fitting. Based on the laser point cloud data of two 220 kV transmission

                    lines in Zhangjiagang,the algorithm results are as follows, Considering the time cost and sampling effect,voxel

                    doum-sampling  is  more  suitable  for  the  conductor  extraction  algorithm.  Multi-piecewise  linear  fitting  effect  is


                    good,individual noise points do not affect the fitting results,and averagely each section can be fitted within 10 s;

                    For non-functional line distribution,it can convert x and y coordinates,or piecewise fit. The transmission lines ex‑

                    traction algorithm is applicable to the transmission lines in the plain area,and can quickly locate the wires in the

                                                              .
                    jumbled laser point cloud (more than 300 million points) It is helpful for the track planning of the subsequent un‑
                    manned aerial vehicle autonomous patrol.




                    Key words:laser point cloud,transmission line,down-sampling,elevation filtering,multi-piecewise linear fitting

                    无人机是电力巡检作业智能替代的核心载                               多数学者采用各种聚类、搜索、拟合算法等提
                体。目前无人机自主巡检的工作模式是人工操控                            取还原点云数据中的输电导线。文献[4]提出了
                无人机将线路通道和杆塔完整地巡视一遍,采集                            一种基于密度的空间聚类方法,在平面内实现对
                激光点云数据并进行三维建模,为无人机规划航                            单条电力线的分割提取。文献[5]根据点云密度
                迹,从而实现无人机自主巡检             [1-3] 。航迹规划的前          差异将电力线与杆塔分离,通过半径搜索算法,实
                提是从激光点云数据中提取导线和杆塔的坐标。                            现复杂环境下单条电力线的提取。文献[6-9]通
                本文主要研究输电导线坐标的提取。                                 过高程阈值分割对电力线进行粗提取,然后利用
   11   12   13   14   15   16   17   18   19   20   21