54
Armston et al. 2013; Morsdorf et al. 2006), above ground biomass (e.g., Babcock
et al. 2016; Popescu 2007; Nelson et al. 1988) and other structural measures (e.g.,
Palace et al. 2015; Higgins et al. 2014; Zimble et al. 2003), site or region specific
relationships are needed. These involve the correlation of LiDAR-derived metrics
associated with the vertical structure of the vegetation with ground-based field data.
To undertake LiDAR data processing, dedicated software processing tools are
required. LiDAR datasets are typically large and require a reasonable amount of
computing power and storage to handle these data. Table 7 lists a number of software packages available for analyzing LiDAR data where LAStools is probably the
most popular and widely used providing plugins for both ESRI ArcMap and the
QGIS software packages to provide an ‘easy to use’ environment.
Classification of Point Clouds
To produce a DTM product from LiDAR data, the classification of ground returns
and the quality of that classification is a critical processing step. There are many
publications demonstrating methods for this task (e.g., Mongus and Zalik 2012;
Evans and Hudak 2007; Zhang et al. 2003). However, you will most likely be limited to the algorithms implemented within the processing tools you have available.
The quality of the classification of the ground returns, and therefore the derived
DTM, is limited by the number and density of the LiDAR returns that have reached
the ground surface. If the LiDAR has not recorded the ground surface, then obviously the ground returns cannot be correctly identified. Where ground returns are
very sparse, it is likely that they will be identified as outliers (i.e., noise) rather than
true ground returns. Dense vegetation over-stories, particularly those close to the
ground (i.e., 1–2 m in height), often limit the number of ground returns. However,
a)
b)
First Return
Second Return
Third Return
First Return
Second Return
Fig. 7 (a) Multiple returns within a vegetation canopy and (b) multiple returns from the edge of a
building
P. Bunting
Armston et al. 2013; Morsdorf et al. 2006), above ground biomass (e.g., Babcock
et al. 2016; Popescu 2007; Nelson et al. 1988) and other structural measures (e.g.,
Palace et al. 2015; Higgins et al. 2014; Zimble et al. 2003), site or region specific
relationships are needed. These involve the correlation of LiDAR-derived metrics
associated with the vertical structure of the vegetation with ground-based field data.
To undertake LiDAR data processing, dedicated software processing tools are
required. LiDAR datasets are typically large and require a reasonable amount of
computing power and storage to handle these data. Table 7 lists a number of software packages available for analyzing LiDAR data where LAStools is probably the
most popular and widely used providing plugins for both ESRI ArcMap and the
QGIS software packages to provide an ‘easy to use’ environment.
Classification of Point Clouds
To produce a DTM product from LiDAR data, the classification of ground returns
and the quality of that classification is a critical processing step. There are many
publications demonstrating methods for this task (e.g., Mongus and Zalik 2012;
Evans and Hudak 2007; Zhang et al. 2003). However, you will most likely be limited to the algorithms implemented within the processing tools you have available.
The quality of the classification of the ground returns, and therefore the derived
DTM, is limited by the number and density of the LiDAR returns that have reached
the ground surface. If the LiDAR has not recorded the ground surface, then obviously the ground returns cannot be correctly identified. Where ground returns are
very sparse, it is likely that they will be identified as outliers (i.e., noise) rather than
true ground returns. Dense vegetation over-stories, particularly those close to the
ground (i.e., 1–2 m in height), often limit the number of ground returns. However,
a)
b)
First Return
Second Return
Third Return
First Return
Second Return
Fig. 7 (a) Multiple returns within a vegetation canopy and (b) multiple returns from the edge of a
building
P. Bunting