Semi-automatic Feature Extraction And Reconstruction From Airborne LiDAR Point Clouds

Airborne LiDAR (Light Detection and Ranging, LiDAR) is a technology whichis a set of three technologies of laser, Global Positioning System (GPS) and InertialNavigation System (INS) Witch can access three-dimensional space information ofobjects on a real-time or quasi real-time. This technology has been applied in a widerange of various fields, for example military, surveying, oil and act. At present,although the LiDAR technology has been more mature in terms of hardwaredevelopment. But the data-post-processing and information extraction theory andmethods are seriously lagging behind relatively. This paper did some meaningfuldiscussion of hot and difficult problems about feature extraction and3Dreconstruction of airborne point cloud.Some of them has been analyzed and realized.Specifically, the major research contents and results of this paper are.First, based analysis of airborne LiDAR laser point distribution law andreferenced to exist method, this paper put forward point clouds classification methodbased on decision which is used to extracting regular feature points. It firstlyseparated the point cloud to different segments according to the coplanar of points andderived the segments outline edges used Delaunay triangles. And then calculate thesegments’ geometric attributes, including the direction, height, area and etc. Separatethe different target feature points according to set constraint conditions by experimentvalue. Experiments show that this method can derive target feature points effectively.It has high degree of automation, strong target, accuracy, and reduces the quantity ofwork. This can help workers improve work and save costing effectively.Second, airborne LiDAR point cloud can be used for3D model reconstructionbecause of its accurate three-dimensional information. This paper purpose themodel-reconstruct method of point cloud data based on the characteristic of airbornepoint cloud and exist modeling reconstruction method. It firstly separated buildingpoints from the whole point cloud. And then separate points to different modules according their connectivity used connected components. The next step isreconstructing every building. The method of building reconstruction can assume theregular of modeling about borne LiDAR point cloud data.Last, this paper supplied several modeling reconstruction tools based differentpoint segments and modules. Deriving segment’s outline includes rectangle, circle,convex, concave and TIN edges. Editing tools of topology data includes movingvertex, deleting vertex and etc the basic editing tools. Deriving simple model containssingle roof building model and gable roof building model. The experiments show thatthis method can effectively complete building reconstruction work used airborneLiDAR point cloud data. It has higher speed and efficiency than drawing model byfull manual. It can improve production efficiency, shorten the production circle andsave production cost if it can be used production.As the constant improving of hardware technology and the further optimizationof relevant theories and methods, airborne LiDAR technology will be applied morewidely.