Identification of hidden geological hazards based on real-scene 3D models:A case study of Shangping town,Guangdong Province
To overcome the limitations associated with identifying concealed hazards at higher altitudes and discern the distinctive characteristics of sudden geological hazards within intricate terrains,it is imperative to achieve precise identification of potential geological hazard sites.Additionally,this study aims to develop an efficient and accurate method for hazard identification and address the limitations of blind areas or unknown sites by finding effective ways to avoid them.Furthermore,it seeks to gain a better understanding of the characteristics of typical potential geological hazards and provide robust data support for disaster early warning,risk prevention,and control.A case study was carried out in Shangping Town,Lianping County,Guangdong Province.This study utilized UAV oblique photogrammetry technology to construct a series of high-resolution(2～3 cm)three-dimensional(3D)models using ContextCapture Center software within eight groups of GPU clusters.Aerotriangulation procedures were also completed simultaneously.Subsequently,point clouds,digital surface models(DSMs),and other derived data were generated step by step from these 3D models.For landslides located mainly in areas with little vegetation coverage,where landslide bodies and deformation cracks are clearly visible.Pieced DSMs derived from 3D models were mosaicked together using ArcGIS software,followed by generating hillshaded maps based on the mosaicked DSMs.Finally,EarthSurvey software was used in combination with the 3D models and hillshaded maps to delineate landslide ranges.As for collapses that are difficult to confirm solely relying on DSM and real-scene models regarding their scope and characteristics,this study adopted Feima UAV manager's smart point cloud plate,which separated vegetation points from ground points,thereby eliminating surface vegetation interference.DEMs were then interpolated based on the ground point cloud,followed by generating corresponding hillshaded maps based on these DEMs.In Comparison to the DSM that does not account for vegetation interference,the utilization of DEMs and hillshaded maps enables a more accurate depiction of collapse shape and occurrence range.Following all necessary steps,potential geological hazards were identified using EarthSurvey software.Furthermore,field surveys were conducted to analyze the characteristics of typical geological hazards within 50 meters from villages.The identified potential geological hazards in this study align with on-site verification,achieving a 100%accuracy rate.A total of 23 potential geological hazards(14 landslides and 9 collapses)were identified,resulting in an approximate density of 4 points per square kilometer.This study integrated methods such as"air-ground linkage","two-three-dimensional integration",and"point-area-volume collaboration"to digitally represent three-dimensional geological hazards while significantly improving identification accuracy for those located in blind spots or higher elevations within mountains.The proposed methods and obtained results hold great significance for disaster early warning systems,risk prevention,and risk control.
real-scene three dimensional modelgeological hazardsUAV Oblique photogrammetryrecognition of potential geo-hazardsGuangdong Province
NETL
NSTL
万方数据
