GRASS GIS for the distinction of vegetation from buildings using LiDAR altimetric data.



Downloads per month over past year

Antolin Sánchez, Roberto and Brovelli, M. A. (2008) GRASS GIS for the distinction of vegetation from buildings using LiDAR altimetric data. In II Jornadas de SIG Libre: comunicaciones de las Jornadas. Universitat de Girona, SIGTE, Girona. ISBN 9788469114339

[thumbnail of Antolin04.pdf]

Official URL:


The LiDAR (Light Detection and Ranging) technology, based on the scanning of the territory by airborne laser telemeters, allows the construction of Digital Surface Models (DSM) by a simple data interpolation, and Digital Terrain Models (DTM) by the identification and removal of attached and detached object (such as buildings, bridges,
power lines or trees). For this purpose, the Laboratory of Geomatica of the Politecnico di Milano - Campus of Como – developed a filter algorithm based on bilinear or bicubic spline interpolation with Tychonov regularization in a least squares approach. However, in many cases it is
still necessary a more accurate and complex DEM in which a distinction between buildings and vegetation is needed, like in some hydrological hazard prevention models, where only vegetation has to be removed, or in automatic 3D city modeling where vegetation is problematic because it
interferes with the vectorization of the building models.
Therefore the filter was revised and further improvements were carried out in order to allow an automatic classification between buildings and vegetation. The procedure uses LiDAR altimetric data as unique information and takes advantage of the differences in height between first and last LiDAR pulses. Firstly, the old filter (consisting in an edge detection, a region growing and a correction algorithms) is applied to these differences. Then, a vectorization is made by considering clustered points classified as 'double pulse'. Vegetation is obtain by discriminating the shape and size of the polygons obtained in the vectorization.
The filter has been developed into the free and open source GRASS 6.2 GIS software (under general public license GPL) as part of the LiDAR tools, in such a way to have an integrated environment suitable to enter, visualize and process the data.

Item Type:Book Section
Additional Information:

Congrés Jornadas de SIG Libre (2es : 2008 : Girona)
1 disc òptic (CD-ROM)

Uncontrolled Keywords:GIS, free software, GRASS, LiDAR, vegetation.
Subjects:Sciences > Physics
ID Code:21934
Deposited On:19 Jun 2013 16:40
Last Modified:29 Sep 2015 11:18

Origin of downloads

Repository Staff Only: item control page