Extraction of building rooffrom airborne laser scanning point cloud

Abstract

Nowadays three dimensional city modeling are necessary for supporting numerous management applications such as in smart city planning. However, the automated determination of precise, reliable and highly accurate city models is still a tedious and challenging task, requiring a pipeline comprising several processing intensive steps. Commercially available software's for building modeling require, generally, a high degree of human interaction. In case airborne laser scanning building are typically identified by their roof points, in this research a simple methodology has been proposed for building roof extraction from airborne laser scanner data. Methodology take input of laser scanner data points which are converted into a text file with the help Lastool. Methodology needs only X, Y and Z values of each point and perform the X-Y gridding by projecting the dataset at X-Y plane. After that vertical segmentation is performed at each grid for generation of area interest and removing the unnecessary points, then area of grid point is calculated with the help of convex hull. Flattering factor is calculated for detection of probable building roof points. At last connected component analysis has been performed with the help of Cloud Compare open source software.