Abstract: Aiming at the problem that not only the slope point cloud is void and uneven distribution,but also the effect of establishing the slope model directly and comparing analysis of deformation is poor,this paper puts forward a method which used an improved inverse distance weight interpolation to fast compress and reconstruct point cloud and procedures to achieve automated processing.Then it established the NURBS surface model fast based on the reconstructed point cloud cond compared analysis of deformation based on the NURBS surface model.The research achieved the purpose of using laser point cloud data to rapidly construct slope model and deformation analysis.By Riegl VZ-1000 scanners collected two point cloud data of a slope in mining subsidence area.The experimental analysis showed that the proposed method could implement quickly and automatically to compress and reconstruct the slope terrain point cloud.And the proposed method could implement quickly and accurately to build the slope model and from the surface,line,point to conduct a comprehensive analysis and comparison.The calculated results were in agreement with the measured values,and conformed to the actual terrain changes.The research of this paper supported the application of 3D laser scanner in slope deformation monitoring.

Key words: slope deformation monitoring, three-dimensional laser scanning, point cloud reconstruction, NURBS surface modeling, deformation analysis