摘要
为了自动检测建筑构件在生产及运输过程中产生的缺陷,提出了基于三维激光扫描和BIM模型的建筑构件检测方法;首先利用三维激光扫描仪获取构件对象的实际点云,并通过弦高偏差法实现点云去噪,同时基于BIM搭建构件的三维模型,通过stl文件将模型对象转换为期望点云;然后分别利用PCA算法和基于K-D树的ICP算法实现点云的初始配准和精配准;最后利用局部均方根值评估构件的误差大小,并通过基于霍夫变换的线性回归分析方法实现了误差量化;通过实例验证了所提算法的可行性与准确性。
To automatically detect the defects in the process of production and transportation, a building component detection method based on 3D laser scanning and building information modeling （BIM） model is proposed. Firstly component object＇ s actual point cloud is acquired by 3D laser scan ner. Noise of point cloud is removed through the chord deviation method, and construction components＇ 3D model is built based on BIM. Model objects is converted to desired point cloud through the STL file. Then the initial registration and precise registration of point cloud are realized by principle component analysis algorithm （PCA） and independent component analysis algorithm （ICP） based on K D tree, Finally, the error of the component is evaluated by the local root mean square value, and the error is quantified by the linear regression analysis method based on Hough transform. The lea sibility and accuracy of the proposed algorithm are verified by practical application.
出处
《计算机测量与控制》
2016年第2期14-17,共4页
Computer Measurement &Control
关键词
三维激光扫描
BIM
点云
配准
构件误差
3D laser scanning
BIM
point cloud
registration
component error
