| 注册
首页|期刊导航|交通运输研究|基于车载激光点云的普通干线公路违法建筑自动化排查方法

基于车载激光点云的普通干线公路违法建筑自动化排查方法

聂世刚 诸雨 刘锦阳 徐阳 胡颖雷

交通运输研究2025,Vol.11Issue(1):67-78,12.
交通运输研究2025,Vol.11Issue(1):67-78,12.DOI:10.16503/j.cnki.2095-9931.2025.01.007

基于车载激光点云的普通干线公路违法建筑自动化排查方法

Automated Inspection Method of Ordinary Arterial Highway Illegal Construction Based on Vehicle-Mounted Laser Point Cloud

聂世刚 1诸雨 2刘锦阳 2徐阳 1胡颖雷3

作者信息

  • 1. 河南交通发展研究院有限公司,河南 郑州 450000
  • 2. 东南大学 交通学院,江苏 南京 211100
  • 3. 河南省交通事业发展中心,河南 郑州 450000
  • 折叠

摘要

Abstract

To achieve efficient and accurate inspection and management of ordinary arterial highway ille-gal construction,ensuring traffic safety and improving transportation efficiency,an automated inspection method of ordinary arterial highway illegal construction based on vehicle-mounted laser point clouds was proposed.Firstly,a hybrid road edge extraction algorithm was constructed based on the fusion of two-dimensional and three-dimensional features.Secondly,the point cloud clustering technology based on GB-ADBSCAN was adopted to identify illegal construction point cloud as a whole,and the Point-Net++model was improved by adjusting the number of features and introducing parallel symmetry function to segment the single illegal construction,achieving the accurate inspection of illegal con-struction in sheltered environment.Finally,the proposed method was verified using laser point cloud data from national highway G207 in Henan Province.The results show that the proposed hybrid road edge extraction algorithm has an extraction accuracy of 98.7%,and the extraction effect was signifi-cantly better than that of three commonly used road edge line extraction methods based on scanlines,feature images and spatial features.As verified by field surveys and real-scene image comparison,the accuracy of unauthorized construction identifying based on vehicle-mounted laser point clouds reaches 87%,indicating that the proposed method can effectively improve the recognition accuracy and speed of the point cloud of illegal construction,and achieve stable and efficient automated inspection of high-way illegal construction under the complex environment of roadside.

关键词

激光点云/违法建筑/特征融合/PointNet++/自动化排查

Key words

laser point cloud/illegal construction/feature fusion/PointNet++/automated inspec-tion

分类

交通运输

引用本文复制引用

聂世刚,诸雨,刘锦阳,徐阳,胡颖雷..基于车载激光点云的普通干线公路违法建筑自动化排查方法[J].交通运输研究,2025,11(1):67-78,12.

基金项目

河南省交通运输厅科技项目(2021T1) (2021T1)

交通运输研究

1002-4786

访问量0
|
下载量0
段落导航相关论文