Identification of inclined buildings from aerial LIDAR Data for disaster management

Abstract

Airborne laser scanning has demonstrated a strong ability to obtain the location of collapsed buildings and the dimension and characteristic of their damage. This paper focuses on the identification of inclined buildings from aerial LIDAR Data which contains amount of noise and small details. To achieve the goal, several steps were carried out in the approach: Firstly, the building's roof was classified and majority of roof types were taken into account in our research; Secondly, a chosen algorithm was used to extract the roof facets of building and also the foundation, and it was proved to be robust to eliminate the noise and give a good result; Thirdly, the geometric axis line of a building was extracted; fourthly, a method was proposed to determining whether buildings are inclined. Finally, an experimental study was made to demonstrate the feasibility of the proposed solution on disaster management.