摘要
激光雷达(light detection and ranging,LiDAR)技术能够快速高效地获取高精度三维点云数据。机载LiDAR不受天气影响且能在复杂地理条件下工作,已成为电力巡检的重要方法。基于机载LiDAR点云数据提取电力线是其中的关键步骤。首先利用电力线悬空的特性,得到非电力线点和电力线候选点;然后构建多类型、多邻域、多尺度的点云特征;再将离散二进制粒子群算法与支持向量机(support vec⁃tor machine,SVM)结合起来进行特征选择,根据选择出的特征子集提取电力线。对比实验表明,该方法的精度和效率都比较高。
The high-precision 3D point cloud data can be obtained quickly and efficiently by light detection and ranging(LiDAR)technology.Airborne LiDAR,which is not affected by weather and can work under complex geographical conditions,has become an important method for power line patrol,while power line extraction based on airborne LiDAR point cloud is one of the key steps.Firstly,the non-power line points and power line candidate points are distinguished through the suspension characteristics of power line.Secondly,multi-type,multi-neighborhood and multi-scale point cloud features are constructed.Then,the discrete binary particle swarm algorithm is combined with support vector machine(SVM)to select features.Finally,power lines are extracted efficiently according to the selected feature subset.Comparative experiments show that the accuracy and efficiency of the method are relatively high.
作者
黄一鸣
赖旭东
张尔严
HUANG Yiming;LAI Xudong;ZHANG Eryan(School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430079,China;Center of Digital Watershed,Wuhan University,Wuhan 430079,China;Shaanxi Tirain Science and Technology Co.,Ltd.,Xi’an 710054,China)
出处
《测绘地理信息》
CSCD
2023年第5期60-64,共5页
Journal of Geomatics
基金
国家自然科学基金(42130105)。
