摘要
使用三维激光扫描仪在对建筑物进行外业扫描时,难以避免地会形成点云孔洞。点云孔洞的存在不仅会影响建筑物点云的浏览,还会对模型建立等后续操作产生影响。因此,在进行相关操作之前,需要对点云孔洞进行修补。使用BP神经网络方法和切片方法对点云孔洞进行修补。根据不同方法的特点,使用BP神经网络方法修补建筑物不规则区域孔洞,使用切片方法修补规则区域孔洞。根据对实际案例的应用分析,两种方法相互结合能够较好地实现对建筑物点云孔洞的修补。
It is difficult to avoid the formation of point cloud holes when 3d laser scanner is used for field scanning of buildings.The existence of point cloud holes not only affects the browsing of building point cloud,but also affects the following operations such as model building.Therefore,the point cloud hole needs to be repaired before the relevant operation.BP neural network method and slice method were used to repair the point cloud holes.According to the characteristics of different methods,the BP neural network method was used to repair the irregular area holes of the building,and the slice method was used to repair the regular area holes.According to the application analysis of practical cases,the two methods combined together can better realize the repair of the building point cloud holes.
作者
孟庆年
王德宏
张洪德
王智
胡玉祥
Meng Qingnian;Wang Dehong;Zhang Hongde;Wang Zhi;Hu Yuxiang(Qingdao Surveying & Mapping Institute,Qingdao 266032,China;Qingdao Engineering Research Center of Geographical Information of Underground Space,Qingdao 266032,China;Shandong Joint Innovation Center of Coast Survey and Monitor Engineering,Qingdao 266032,China;CRRC Science and Technology Park (Qingdao) Co.,Ltd.,Qingdao 266000,China)
出处
《城市勘测》
2021年第4期118-123,共6页
Urban Geotechnical Investigation & Surveying
关键词
建筑物点云孔洞
修补
BP神经网络
切片
the building point cloud holes
repair
BP neural network
slice
