基于LiDAR点云数据的电力线自动提取算法研究  

Power Line Automatic Extraction Algorithm Based on LiDAR Data

作　　者：刘晓磊[1] 郑顺义[2] 许磊[2] 

机构地区：[1]中国飞行试验研究院,陕西西安710089 [2]武汉大学遥感信息工程学院,湖北武汉430079

出　　处：《测绘地理信息》2017年第4期13-16,共4页Journal of Geomatics

基　　金：国家自然科学基金资助项目(41671452)

摘　　要：通过分析LiDAR点云中电力线的特点,提出一种基于动态阈值滤波和种子点跟踪相结合的电力线自动提取算法。首先,根据电力线高程特征动态获取过滤阈值,将地面点滤除;然后,将点云投影到水平面,并建立位置索引;接着,对影像进行线特征检测,并利用建立好的索引寻找对应的电力线点;之后,对初始电力线点云进行加密;最后,利用种子点跟踪的方法对初检测结果进行精化,在过滤掉噪声点的同时实现电力线点的矢量输出。经实验验证,提出的电力线提取算法可以将实验区内单条电力线的识别率提高至90%以上。According to the analysis of the characteristic of power lines in the LiDAR data, this paper presents an algo- rithm which combines dynamic threshold filtering and the seed point tracking together to extract the power line auto- maticly. Firstly it uses the threshold to wipe out the ground points according to the elevation characteristics of power line. Secondly, it casts the point cloud to the plane and crea- ting indexes of points＇ position. Thirdly it locates lines in the image and find the power line point＇s position according to the effective indexes. Next it makes power line points be- come denser, Finally, it takes advantage of the seed points to improve the accuracy of original results and wipe off the noise points at the same time, Experiment results show the rec- ognition of single power line can reach over ninety percent.

关 键 词：LIDAR 点云 动态滤波 线检测 精提取 

分 类 号：P225[天文地球—大地测量学与测量工程] TM7[天文地球—测绘科学与技术]