Geoscience Reference
In-Depth Information
15.4 LiDAR Data Derived Products and Applications
15.4.1 Digital Elevation Models
Although there is no common terminology, we use digital elevation model (DEM)
as a generic term for a three dimensional representation, typically in a raster format.
Many processing steps can be performed on the point cloud itself, but in the end the
data are often rasterized in a gridded format. This facilitates further analysis in a GIS
and overlaying with other geospatial data.
The creation of a DEM involves filtering and interpolation . Filtering removes
unwanted pulse returns. For instance, as a subset of a DEM, a digital terrain model
(DTM) is based on ground returns only. Returns obtained from above ground targets
must therefore be filtered. Conversely, for a digital surface model (DSM), pulse
returns from the ground must be filtered.
Interpolation, on the other hand, is needed to fill raster grid cells that are not cov-
ered by any pulse returns. This has also been discussed in Sect. 10.1 on gridding point
vector data. We must also decide how to handle multiple points that cover the same
grid cell. This is similar to creating raster composites, explained in Sects. 8.1 and 12.1 .
The implementation of interpolation and filtering techniques for creating a DEM
from point clouds often remains in proprietary software. LAStools provides the
executable las2dem , lasgrid and lascanopy for restricted use 22 inaWindows
environment.
In PDAL, drivers.p2g.writer can be used in combination with pdal
pipeline to convert point clouds to a gridded raster. Each cell in the output grid
can give one of the: minimum value, maximum value, average value, median value,
inverse distance weighted interpolation (for sparse points), or density. With some
background in Python programming, more advanced processing techniques can be
added.
In SPDLib, the filtering steps are performed on the pulse in SPD format (see
Sect. 15.3.5 ). To convert the SPDpulse data to a raster dataset, the utility spdinterp
can be used. An example is provided in the case study in Chap. 18 .
In pktools (see Chap. 12 ) , the utility pklas2img creates a gridded raster dataset
from a point cloud in LAS/LAZ format. You can filter the points to be retained
in the output grid. For instance, to select the first pulses only, use the option
-fir|-filter first .
As a simple example, we can create a digital surface model of the point cloud
s t-helens.las .
pklas2img -i st-helens.las -o st-helens_dsm_10.tif -p epsg:26910
-dx 10 -dy 10 -ot Float32
22 http://www.cs.unc.edu/~isenburg/lastools/download/LICENSE.txt
 
 
Search WWH ::




Custom Search