3D Building Model Reconstruction Based On Airborne LiDAR And Remote Sensing Data

With the rapid development and application of digital city,three-dimensional(3D)building model reconstruction,as an important concept of digital city,has become an advanced research hotspot in geographic information science and remote sensing.It has been widely used in virtual environment,3D navigation,disaster responsing,and so on.With the advantages of all-weather,short cycle,and high-accuracy,airborne light detection and ranging(LiDAR)is able to obtain the 3D information of urban surface with high temporal and spatial resolution.Therefore,airborne LiDAR has become an important data source for 3D building model reconstruction in urban areas.Points near contour nodes and the direction change in outlines are rarely considered by most researches during regularizing process,leading to local distortion in the results.Besides,there are very few researches on reconstructing detailed 3D model.To make up for these two deficiencies,this paper focused on 3D building modeling in the urban area of Beilun District,Ningbo,China with airborne LiDAR and remote sensing data.The major studies are as follows:(1)Extracting outlines of the buildings in the study area.In this paper,the procedure of extracting outlines of buildings from airborne LiDAR and remote sensing data was studied.The main processes are as follows;1)A threshold segmentation technique was employed to remove both low vegetation and city components from the normalized digital surface model(nDSM),which was created from airborne LiDAR data.2)By combining Visible-band difference vegetation index(VDVI)and threshold segmentation technique,the residual vegetation can be detected and removed,getting accurate building infomation.3)Morphological closing and opening operations were used to fill holes in buildings and remove psedo outlines of buildings,resulting better building outlines for the study area.(2)Extracting outlines with different altitudes.Outlines of different heights are usually required to reconstruct detailed 3D building models.We studied the contour-based method to acquire outlines of different altitudes,which can be the data for regularization and 3D modeling later.(3)Regularizing building contour lines.The edges of the contour lines are usually so irregular,mainly in the form of zigzag,that should be regularized.In this paper,an algorithm of compressed polyline with minimum number of vertices was employed for the regularization of building contour lines.And the result achieved the maximum simplification and regularization as well as keeping the original shapes of buildings.(4)Reconstructing detailed 3D building model with CityEngine.This paper reconstructed detailed 3D building models in the study area by using the regularized buildings' contour lines and computer generated architecture(CGA)rules with CityEngine software.Compared with the real buildings,reconstructed models captured the details very well.The reconstruction method used in this paper might provide some ideas for 3D building modeling in urban areas at large scale.