Management Of Massive Point Cloud Data Of Metro Tunnel And Design Of The System

The traditional method of metro tunnel monitoring is measuring the points which were laid by people on the tunnel through using total stations.These monitoring-points are always divided into several groups and different group is on the different cross-section of the tunnel.There is drawback of the method that the number of monitoring-point is limited,and meantime,the precision of measurement is easily affected by environmental factors,such as the humidity and pressure of tunnel environment.The disease detection of metro tunnel is always a manual method which is not only very slow but also prone to having some omissions due to the limited viewing distance of people's eyes.The 3D laser scanning technology is a non-contact measurement technology which is rapid developed in recent years.The work efficiency would be improved a lot if we use the 3D laser scanning technology to do the monitoring and disease detection of metro tunnel.Meantime,we can get high-density points data which can be used to build the model of metro tunnel by using this technology.The massive point cloud data of the tunnel has brought a huge challenge to the computer hardware and software because of the limited computer memory and the CPU's computing ability.We can not handle all points in the same time by putting all data into computers memory.So we should build a reasonable index structure to manage the massive point cloud data.At present,domestic and overseas scholars have taken up large-scale research on the management and visualization of massive point clouds,and put forward different point cloud management models,such as the quad tree model,the octree model,the R tree model and so on.Though there are so many models we can use,each of them has its own applicability and none of them can be suitable for all different spatial distribution of point cloud.In this paper,a new method is used to get metro tunnel point cloud that we use the tunnel scanning car to get the data.The basic hardware of the scanning car is introduced.As a result of the dynamic scanning,this paper also introduces the method of point cloud registration.In order to manage the massive tunnel point cloud data,considering the character of the spatial distribution of the tunnel point cloud,this paper proposes a hybrid index model which is suitable for the point cloud of the tunnel.Meanwhile,an efficient storage method of point cloud data is introduced as well.In order to get a smooth visualization of massive tunnel point cloud,this paper proposes an out-of-core visualization method based on the OSG.By using the DatabasePager technology provided by OSG,this paper realize an efficient point cloud rendering effect combing with the index structure built before.For the application of he tunnel point cloud management and visualization theory,this paper designed and implemented a tunnel point cloud processing platform,which provides point cloud management,visualization and interactive and the function that getting the points of tunnel's cross sections are added into the platform.At last,this paper selects a region of a line of Wuhan Metro,getting point cloud data of this region by using the tunnel scanning car and disposing the point cloud data in on the platform.The disposing effect shows that the point cloud management and visualization theory is suitable for the massive tunnel point cloud data.