Study Of Rock Mass Surface Information Identification In High Steep Slopes Based On Nap-of-the-object Photography

With the rapid development of modern cities and transportation networks,the number and scale of infrastructure such as high and steep slopes are increasing.However,high and steep slopes are also the main sites of geological disasters such as landslides and debris flows,which endanger people's property and safety.Therefore,it is very important to carry out detailed investigation and long-term monitoring of high and steep slopes.In order to overcome the disadvantages of low efficiency and high risk of traditional survey methods,this paper proposes a high and steep slope survey method based on UAV close to photography technology.A semi-automatic discontinuities identification and extraction algorithm is proposed,which relies on Knn clustering algorithm and PCA algorithm.This algorithm can be used to identify the orientation information of discontinuities.With the support of the National Key Research and Development Program of China and Changsha Science and Technology Plan Project,the main research contents and results of this paper are as follows:(1)The feasible experimental scheme and its parameters to survey steep slopes are presented,which based on UAV nap-of-the-object photography.Taking the frame beam slope near Changsha Super Computer center as the experiment object,multiperiodic 3D real-scene models are established.By comparing the error and detail quality of the models,the 3D models established by nap-of-the-object photography possess excellent intuitiveness and clear details.The root mean square error of the three directions of the model is all better than 1cm,and the overall three-dimensional error is also less than 1cm,which is enough to meet most engineering investigation requirements.The method is simple,efficient,safe and flexible,which is a new geological survey technology.(2)A semi-automatic discontinuity identification and extraction algorithm is studied with Knn clustering algorithm and PCA algorithm as the main structure.Firstly,the candidate planes are constructed by Knn algorithm,and the candidate planes are tested for coplanarity by PCA algorithm.Then,calculating the plane parameters of the best fit plane and its normal vector by the least squares method,assigning this normal vector as the normal vector of the initial candidate point.The connectivity between points and the maximum angle between normal vectors are taken as constraints to draw the boundary of discontinuities.Finally,the normal vector of the plane is used to determine the orientation,completing the semi-automatic identification and extraction of the discontinuities.(3)Based on the 3D point cloud of Changsha Super Computer Frame beam slope,the size and distribution of each discontinuities of the slope can be obtained by calculating the processed point cloud using the algorithm above.The strike error of the identified discontinuities is better than 2.8 degrees,and the dip error is better than1.5 degrees,which meet the requirements of most engineering investigation,and also verify the accuracy and effectiveness of the algorithm.(4)Applying the Nap-of-the-object photography and algorithm on Changsha Dingzi town slope renovation project of abandoned mining area,we reconstructed the3 D real scene model and 3D point cloud model whose accuracy is better than 2cm.what's more,we also obtained the discontinuities distribution and location of the five identified regions,and then recorded their orientations.The spatial distribution of discontinuities in the regions is demonstrated by plotting stereogram of poles,discontinuities division diagram and pole contour diagram.Preferred plane are extracted in 5 identified regions for further qualitative analysis of its effect on slope stability.Finally,four groups of dangerous discontinuities are identified,which provided a reliable basis for the engineering support design.