Study on extraction of micro-terrain features from 3D point clouds and its application
Addressing the challenge of ineffective recognition of micro-terrain features and small-scale geological hazard risks by general terrain characteristics,this paper proposed a new extraction and computation method for micro-terrain features applicable to LiDAR 3D point cloud data,based on discussing various commonly used terrain features.The micro-terrain features,including positive openness,negative openness,and slope,along with eigenvalue ratio,are selected to emphasize the features of the topographic concave and convex that have the highly nonlinear correlation with geological hazards.The high-dimensional information can be directly employed as inputs for machine learning algorithms focusing on terrain features,while the reduced-dimensional features can maximize the preservation of convex and convex features in 2D visualizations.Results from the case studies indicate that the fusion of openness and slope in 2D visualizations can clearly represent the concave and convex features of micro-terrain,and the eigenvalue ratio can map directional characteristics of mass movement of micro-terrain.These features provide effective terrain feature data support for recognizing geological hazard risks at different scales under micro-terrain conditions.
NETL
NSTL
万方数据
