Research On Slope Deformation Monitoring And Dangerous Rock Mass Identification Based On Laser Point Clou

With the development of China’s economy and the establishment of large-scale infrastructure,the stability of the natural environment has been greatly affected,resulting in the frequent occurrence of geological hazards on slopes.The deformation monitoring of dangerous rock masses,which are hidden and sudden,can effectively grasp their deformation dynamics,and then take measures to prevent and reduce the damage caused by dangerous rock collapse in advance.As rockfalls are usually located on high and steep rock faces,it is difficult to monitor them efficiently by traditional measurement methods.Compared to conventional monitoring methods,3D laser scanning technology has the advantages of non-contact,high accuracy and speed.Therefore,this paper uses the advantages of 3D laser scanning technology to carry out deformation monitoring of the slope near the Beehive Hotel in Pukou District,Nanjing,to analyse the deformation information of the slope and extract the dangerous rock mass.The following are the main research contents and conclusions of this paper:(1)To study the key principles of 3D laser scanning technology,analyse the factors leading to its measurement errors on this basis,derive the acquisition method to control the data quality in view of the error influencing factors,and successfully collect the point cloud data of the slope in four phases.The method of using 3D laser scanning technology and techniques related to 3D point cloud processing are investigated in order to reduce errors and improve data accuracy.These include the acquisition of point cloud data,point cloud thinning,noise reduction and point cloud stitching processing techniques.(2)Study point cloud alignment-related algorithms,based on the ICP point cloud alignment algorithm,to address the problem of high initial position requirements of the ICP algorithm.The SAC-IA point cloud coarse alignment algorithm is used to adjust the initial attitude of the point cloud data,and the ICP alignment algorithm is combined with the ICP alignment algorithm to stitch and align the acquired multi-station and multi-period data.The final four phases of point cloud data were aligned to an accuracy of 2.7 mm,2.6 mm,2.8mm and 2.7 mm respectively,which laid the foundation for subsequent model reconstruction and deformation analysis.(3)The NURBS surface modelling algorithm was studied to construct a digital surface model,and the accuracy of the constructed surface model was evaluated using Geomagic Studio.On the basis of the digital surface model,the multi-phase point cloud was differentially processed to analyse the 3D deformation of the slope.The results showed that the slope as a whole was uplifted by 8.1mm in the Z-axis,with large settlement and uplift in some areas reaching a maximum uplift of 1.7607 m and a maximum settlement of 1.7411 m,with an average uplift of 21.2mm and settlement of 9.4mm.(4)Research on techniques related to the extraction of surface features of dangerous rock bodies,using inconsistency indicators to extract edge point clouds and nearby point clouds,combined with K-means clustering algorithms for rock body delineation.Superimposed on the slope deformation information to achieve the extraction of dangerous rock bodies in the slope point cloud data.The slope is finally divided into 113 rock objects and 7 dangerous rock bodies are identified and extracted.