基于激光点云的公路箱梁表面裂缝识别方法研究
Research on Surface Crack Identification Method of Highway Box Girder Based on Laser Point Cloud
刘春阳 1曹少华 1杨维双 1张舒烨 2马会环 2赵辰洋2
作者信息
- 1. 中铁四局集团第五工程有限公司,江西九江 332000
- 2. 中山大学土木工程学院南方海洋科学与工程广东省实验室(珠海),广东珠海 519080
- 折叠
摘要
激光扫描作为一种新兴的无接触式测量技术,具有测量速度快、精度高的特点,常用于物体表面的损伤检测中.为从激光点云中提取裂缝所在位置,提出一种基于点云扫描线的混凝土预制箱梁表面裂缝识别方法,该算法根据点云在不同条带上的高度差异获取突变点,并使用小波分解增加突变点的拾取精度,最后通过连接突变点得到裂缝的平面分布.点云实验表明,依赖于小波分解对突变点的精确检测,该算法能够实现对预制箱梁表面独立单一走向裂缝的精确识别,其中使用Coif2小波基进行4层小波分解能够实现最佳的识别效果.
Abstract
Introducing laser scanning as a novel non-contact measurement technology known for its rapid speed and high preci-sion,commonly applied in surface damage detection.This paper presents an algorithm for detecting surface cracks on concrete prefabricated box girders based on laser point cloud scanning.By analyzing height differences along scanning lines,the algo-rithm identifies crack locations and employs wavelet decomposition to enhance detection accuracy of mutation points.Connect-ing these points across different scanning lines reveals the crack distribution.Experimental results demonstrate the algorithm's capability to accurately identify individual linear cracks on prefabricated box girder surfaces,with optimal performance achieved using 4-layer wavelet decomposition with Coif2 wavelet base for mutation point detection.
关键词
工厂预制/激光雷达/点云处理/裂缝检测/小波变换Key words
factory prefabrication/LiDAR/point cloud processing/crack detection/wavelet transform引用本文复制引用
出版年
2024
NETL
NSTL
万方数据