Study on extraction method of slope surface deformation in open-pit mine by UAV remote sensing
Given the challenges prevalent in slope monitoring within open-pit mines,including blind spot monitoring,insufficient criteria for control point positioning,complexities in geological hazard interpretation,and the intricacies of reconstructing point clouds from Unmanned Aerial Vehicle(UAV)images,we introduce a novel method for extracting slope surface deformations leveraging UAV remote sensing.Initially,we collected point cloud and image data of the slope at various intervals using a UAV equipped with onboard Light Detection and Ranging(LiDAR)and a camera.Subsequently,the acquired slope point cloud and image data underwent preprocessing and alignment to establish a unified coordinate system.Leveraging the strengths of both slope point cloud and image data,we synthesized a sequential point cloud dataset.Next,taking into account the distinct surface characteristics found in open-pit mines,we enhance the conventional Iterative Closest Point(ICP)algorithm by integrating the Scale Invariant Feature Transform(SIFT)and cylindrical neighborhood search algorithm.This augmentation facilitates the extraction,mapping,and precise registration of feature point pairs between point cloud sequences,thereby enhancing the accuracy and efficiency of slope deformation extraction.Lastly,employing a two-phase Digital Elevation Models(DEM)overlay analysis and visualization approach,we achieved precise localization of critical slope deformation areas.This facilitated intuitive extraction of slope deformation positions and sizes.Additionally,by integrating image features from the Digital Orthophoto Map(DOM),we conducted a comprehensive analysis and interpretation of the deformation region.To illustrate and evaluate our methodology,we applied it to the Nanfen open-pit mine as a case study.The findings indicate that the enhanced ICP algorithm significantly improves the registration accuracy of dual point cloud datasets.Furthermore,the standard deviation of the slope deformation model is measured at 0.032 m.Comparative analysis against deformation values obtained from Global Positioning System-Real Time Kinematic(GPS-RTK)reveals a median error of 0.012 m.Hence,the method enables rapid scanning and extraction of deformations across large-scale,intricate slope surfaces.This capability offers critical technical backing for geological hazard analysis,monitoring deformations in blind spots,and strategically deploying ground monitoring apparatus.
万方数据
维普
