Intelligent Recognition Of Slope Discontinuities Based On 3D Laser Scanning

With the increase of the slope height of open-pit mines in China,the instability phenomena such as landslides are increasing year by year.The structural plane is an important factor that affects the stability of the slope.It is the premise and foundation of the slope stability evaluation of the open pit mine to obtain the information of the structural plane of the rock mass accurately and quickly,which has an important engineering practical value.3D laser scanning technology is an advanced automatic high-precision stereo scanning technology.Compared with traditional manual measurement and photogrammetry,it breaks through the constraints of single point measurement,and can quickly and accurately obtain the 3D point cloud data of the slope rock mass.However,the technology produces a large amount of point cloud data in the measurement process.How to reconstruct the surface model of structural surface and group the dominant occurrence from the complex and irregular point cloud data is the key problem of the engineering application of the technology.Therefore,based on the above-mentioned key problems and the rock point cloud data of slope,this paper carried out the systematic research on the intelligent recognition of rock structural plane,the main research contents are as follows:(1)Based on the fact that the surface interpolation method is easy to be affected by noise points and the data segmentation of point cloud is tedious,an improved fuzzy c-means algorithm is proposed to reconstruct the surface of structural surface,and the surface reconstruction process is successfully realized.This method takes all points in the point set determined by principal component analysis as the clustering center of each cluster of fuzzy c-means algorithm.By solving the membership degree of spatial sample data to set the clustering center and iterating continuously,the minimum value of function JM is calculated,so as to achieve the purpose of structural surface model reconstruction of rock slope,which is one of theinnovations of this paper.This method avoids surface interpolation and simplifies the operation process,so it has a certain practical value.(2)Based on the reconstruction of rock slope model,the unit normal vector of structural plane is transformed into the occurrence information of structural plane by the method of equal angle lower half projection,which provides the basis for the advantage grouping of structural plane.(3)The neighborhood propagation clustering algorithm overcomes the problem that the traditional K-means clustering algorithm is sensitive to the initial clustering center and easy to fall into the local optimum.The previous research has been successfully applied to the advantageous grouping of the occurrence of rock structural planes,but it still does not solve the problem of automatic and fast selection of the input parameter p value that affects the clustering results and determines the final clustering number.For this reason,an improved algorithm of adaptive p-value scanning technology is proposed to group the structural surface occurrence information extracted from point cloud data.The sharp angle between two structural surfaces is used to replace the Euclidean distance as the similarity measurement criterion.The mathematical model of structural surface grouping is established,and the P-value is dynamically reduced in the iteration process.The validity of the silhouette clustering is introduced As the standard of selecting and evaluating clustering,index is the second innovation of this paper,which is to obtain the optimal grouping of structural planes.The application of the artificial random structural surface and the actual engineering survey structural surface shows that this method has high reliability in the occurrence grouping,and has higher robustness and calculation efficiency compared with Shanley and mahtab methods,fuzzy c-means clustering,spectral clustering and particle swarm optimization clustering methods.