Research On Slope Deformation Monitoring Based On 3D Laser Scanning Technology

In the increasingly developed social environment,more and more artificially excavated or naturally formed slopes have occurred,and accidents caused by slope disasters have also increase.In order to understand the deformation and stability of slopes in time,it is necessary to monitor the slopes.Aiming at the problems of insufficient data,long observation period of traditional measurement and observation stations,the 3D laser scanning technology is applied to slope deformation monitoring.In this paper,combined with the characteristics of 3D laser scanning data,through the registration and modeling of point clouds,the corresponding deformation analysis method is used to analyze the slope deformation from the point,line and surface characteristics.The main research contents of this paper are as follows:1?The traditional ICP algorithm requires strict initial position and pose of point cloud,and registration speed is slow.To solve this problem,PCA rough registration+ICP precise registration algorithm is used to register point clouds.The two algorithms are combined to complement each other.The initial position and attitude of point clouds are corrected by principal component analysis,and ICP precise registration is performed by using the obtained rough registration point clouds.The accuracy and efficiency of the algorithm are analyzed by the point cloud data of rabbit and orchard slope,and the reliability of the algorithm is verified.2?A set of slider simulation experiments are designed.Use the sliding block data of multiple scans to superposition the data of simulated landslide,and compares the sliding block displacement calculated by 3D laser scanner with the sliding block displacement measured by ruler.The accuracy of the 3D laser scanner and the improved ICP algorithm are verified by comparing the slider with or without multi-station splicing.3?In view of the large amount of slope point cloud data and incomplete closed surface,the DEM model of the original point cloud was constructed by using poisson surface reconstruction and NURBS surface reconstruction technology,it can be seen that poisson surface reconstruction technology is not suitable for unclosed surface reconstruction.For NURBS surface reconstruction,the target point of the marker points extracted by the 3D laser scanner are compared with those measured by the high-precision total station to verify the reliability of the reconstructed model.The experimental results show that the median error of the model in the x-axis is w_x=0.0024m,the median error in the y-axis is w_y=0.0020m,the median error in the z-axis is w_z=0.0022m,and the median error in the point position is w_p=0.0038m.It can be seen that the precision of model reconstruction is high and the detailed characteristics of slope can be well reproduced.4?For the reconstructed slope DEM model,the overall deviation of different stages is constructed by comparing the models,and the influence of error propagation on the overall analysis is obtained.Then,the multi-phase slope superposition and 2D profile line analysis were used to compare the displacement of corresponding points before and after slope sliding and the deformation of the whole slope.The results show that the slope as a whole is in a stable state,and only a small part of the area has been deformed.Then,by analyzing the difference of contour lines of the slope,the position of slope deformation in the point cloud data is obtained.According to the different slope data in the two periods,the uniform reference surface was established to calculate the change of soil volume of the slope,which further verifies the deformation situation and trend of the slope.