Geoscience Reference
In-Depth Information
Part II
Third Party Open Source Geospatial
Utilities
With the increasing popularity of open source software, the supply of packages with
utilities to process geospatial data is also growing. Functional overlap exists and
each package comes with its own
cation for the
selection of third-party tools in this topic is based on two requirements. First, their
design and usage are in line with GDAL and, second, they must be complementary
to some extent. The
avors and advantages. The justi
rst requirement excludes an important deal of packages that
provide their own integrated software environment. These environments are typi-
cally required to run the utilities that are written in an interpreted computer lan-
guage. Two examples of such environments are GRASS GIS and R. Without
question, both are powerful alternatives to the approach followed in here. Covering
them here would be beyond the scope of this topic and we would like to refer to
excellent existing topics (Neteler and Mitasova 2008; Bivand et al. 2008). In this
part, we exclusively deal with compiled utilities that run from the command line
shell (Bash in case of Linux) and are driven by command line options.
We believe the power of these tools is still underused. Their modular approach is
ideal to be combined in a processing chain. This chain is typically implemented in
Bash, Python, Tcl/Tk, or another scripting language. The second requirement
regarding complementarity is to
ll the gap between the low level geospatial data
processing and the need for higher level processing for practical applications. For
the context of this topic, this means applications in Earth systems data and models.
In particular, our emphasis is on remote sensing applications, with an emphasis on
land cover classi
cation.
We selected pktools (Chap. 12 ) and the Orfeo toolbox (Chap. 13 ) . They combine
the power and simplicity of the GDAL command line interface with more advanced
and state-of-the-art image processing techniques. Both tools are developed in C++
and are designed for high performance and large data processing.
In addition, this part deals with three-dimensional point clouds acquired with
laser scanning instruments (LiDAR). These data are covered here, because they are
not within the scope of OGR and GDAL and also require third-party utilities.
Search WWH ::




Custom Search