Building Detection from LIDAR Point Cloud Data

Abstract

LiDAR (Light Detection and Ranging) is an emerging technology now and has proved to be one of the best techniques for 3D city modeling. Building detection is an important aspect of 3D city modeling as it can help in urban planning, utility services, disaster management, traffic management, environmental monitoring and many other applications. In this paper we propose a method in which building structures can be accurately discriminated from vegetation. Pre-processing of the data is done using a remote sensing tool called Las Tools. The data is structured using kd tree and then further segmented using Euclidean distance clustering. Then we process this data using another open-source remote sensing tool called CloudCompare. This paper will help in imparting a clear idea of how efficiently and accurately building detection can be performed with the help of these open-source tools without deploying much complex algorithms.