The Research On Key Technology Of Large Slope Deformation Extraction Based On 3D Laser Point Cloud

A large number of slopes emerge with the rapid development of China’s economy.Under the action of external load,slope body is usually showing heterogeneous and anisotropic characteristics,such as excavation,rainfall,surcharge,river scouring,reservoir water level fluctuation and earthquakes,it is prone to showing local or overall instability phenomenon.The slopes will cause serious harm to the people’s lives and property once fail.In order to prevent the damage caused by slope fail,we need to use certain means to monitor slope engineering timely and effectively,extract slope deformation information and further analyze the possibility of slope instability,put forward targeted disaster prevention measures,which is of great significance to ensure that the national economic development and people’s lives and property safety.Compared with the traditional single point monitoring method,the 3D laser scanning technology has been widely used in slope safety monitoring with its high precision,high efficiency and non-contact advantages.Based on the 3D laser scanning,the most common method of slope monitoring is calculating the difference of two-stage model ultimately.We can only roughly extract a large area of regional deformation by the classical regional model from the massive scanning point cloud,however,which can’t be used to accurately monitor the local change of the small area.Therefore,in this paper,the key technologies of point cloud registration and taking into account the best search neighborhood deformation information extraction are studied in depth,combined with three-dimensional laser point cloud data.The main research contents are as follows:(1)Improved Point Cloud Registration Algorithm.An improved point cloud registration algorithm is proposed based on the classical iterative approach.Complete the point cloud of the overall registration work through the point cloud rough registration and fine registration.Firstly,the featured points set are used to complete the initial registration of the point cloud.Then,based on the rough registration,the stable points set is introduced into the improved ICP algorithm to complete the accurate registration of the point cloud,the cloud point of the entire registration work is completed ultimately.(2)Determine the best search neighborhood.In this paper,the K nearest neighbor search algorithm is introduced into the slope deformation information extraction,and a method of quantitatively determining the best search neighborhood is proposed according to the nearest neighbor search model.The nearest neighbor search algorithm is used to find the corresponding points.By the connection neighbor points with point cloud coincidence error,analyzing and gathering statistics the displacement of the corresponding points to determine the best search neighborhood.(3)Extract the slope deformation information.In the process of slope deformation information extraction,the deformation information in the local area of the slope is extracted with the method proposed in this paper by means of the combination of field simulation and practical engineering case,and the deformation analysis of the slope surface is completed.The feasibility and the applicability of the method are verified,which can make up for the shortcomings of the regional model reconstruction.The three-dimensional laser scanning technology is applied to the extraction of large-scale slope deformation information,which can monitor the deformation trend of the slope in time and efficiently.Based on the large-scale slope point cloud data,the slope data of different phases are unified into the same coordinate system through the improved point cloud registration algorithm,and then the nearest neighbor search algorithm is used to extract the slope surface deformation information.The feasibility and engineering applicability of the method are verified by means of field simulation and practical engineering case.