大桥工字梁点云模型特征面智能识别及应用
作者:

Intelligent recognition and application of characteristic surfaceof point cloud model of bridge I beam
Author:
  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献 [20]
  • |
  • 引证文献
  • | |
  • 文章评论
    摘要:

    独立钢梁预制构件的质量检测是现代桥梁建设的一个重要环节,传统手工作业的方式工作效率低且费时费力。为快速准确地对独立钢梁构件进行质量检测,提出一种基于点云模型的桥梁构件特征面语义智能识别算法,对构件质量参数所涉及的特征面进行智能提取。首先,利用三维激光扫描技术建立钢梁构件的精细化点云模型;其次,采用一种基于超体素的区域聚类复杂场景分割算法,对模型的不同特征区域进行划分,将同一特征明显的大区域组合体进行生长融合,将细小的、独立无关的特征体进行去除;再次,根据模型的空间几何特征(密度、法向量、空间连通性)建立语义提取规则,根据语义规则智能提取钢梁构件检测所涉及的关键特征语义面;最后,基于提取的特征语义面,提出钢梁构件梁高参数和拱度参数的检测算法思路。算法用于某省特大桥工字形组合钢梁构件检测,实验结果可知,梁高最大绝对误差165mm,最小绝对误差037mm,拱度最大绝对误差58mm,最小绝对误差07mm,均满足规范要求,该算法可以准确快速地识别构件关键语义特征面,为钢梁构件质量安全检测提供新的思路。

    Abstract:

    The quality inspection of prefabricated components of independent steel beams is an important link in modern bridge construction,but the traditional manual operation is inefficient and time consuming.A semantic intelligent recognition algorithm of bridge component feature surface based on point cloud model is proposed to intelligently extract the feature surface involved in component quality parameters in order to quickly and accurately inspect the quality of independent steel girder components.Firstly,a refined point cloud model of steel beam components is established by using three dimensional laser scanning technology;secondly,a region clustering complex scene segmentation algorithm based on super voxel is used to divide the different feature regions of the model,grow and fuse the large region combinations of the same feature that are obvious,and remove small,independent and irrelevant feature bodies;thirdly,the semantic extraction rules are established according to the spatial geometric features of the model(density,normal vector and spatial connectivity),and the key feature semantic aspects involved in the intelligent detection of steel beam components are identified;finally,based on the extracted feature semantic surface,the detection algorithm of beam height parameters and camber parameters of steel beam components is proposed.The algorithm is used for the detection of I beam composite steel beam components of a provincial super large bridge.The experimental results show that the maximum absolute error of beam height is 1.65mm,the minimum absolute error is 0.37mm,the maximum absolute error of camber is 5.8mm,and the minimum absolute error is 0.7mm,all of which meet the specification requirements.The algorithm can accurately and quickly identify the key semantic feature surfaces of components,providing new ideas for the quality and safety detection of steel beam components.

    参考文献
    引证文献
引用本文

刘德儿,司玄玄,杨大海,殷涛,刘志刚.大桥工字梁点云模型特征面智能识别及应用[J].激光与红外,2023,53(4):528~536
LIU De-er, SI Xuan-xuan, YANG Da-hai, YIN Tao, LIU Zhi-gang. Intelligent recognition and application of characteristic surfaceof point cloud model of bridge I beam[J]. LASER & INFRARED,2023,53(4):528~536

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 在线发布日期: 2023-04-19