Geography Reference
In-Depth Information
Value
High : 548.22
N
Low : 1.00
Inland water
0
20
40 km
Figure 10.6
Mean of elevation for a 3
¥
3 pixel moving window.
averaging. Figure 10.7 shows the standard deviation, also for a 3 ¥ 3 pixel moving win-
dow. In this case, it is apparent that the edges of more notable topographic features
(particularly those with high elevation values) are highlighted.
h e mean i ltered DEM (recall that this is a grid-based representation of topographic
form) in Figure 10.6 is smoother in appearance than the uni ltered DEM in Figure 10.5.
h is is particularly clear on the edges of areas with higher elevation values. Other i lters
are dei ned by Sonka
et al.
(1999) and Lloyd (2006). Smoothing i lters (such as the mean
i lter) are used in many dif erent contexts. For example, smoothing i lters are ot en used
to reduce the ef ect of 'noise' in remotely sensed images (Mather, 2004), although the
median i lter, for example, may be more suitable in such cases than the mean i lter (as
the mean is af ected by outliers, whereas the median is not). h e focus of the following
section is the analysis of DEMs specii cally (as opposed to other forms of raster grids).
Derivatives of altitude
10.5
h e remainder of this chapter focuses on the analysis of DEMs. Various products are
ot en derived from DEMs—derivatives such as gradient and aspect (the two making
up slope) are ot en computed. Gradient refers to the maximum rate of change in altitude
while aspect refers to the direction of the maximum rate of change (e.g. gradient may
be north facing) (Burrough and McDonnell, 1998). h e terms 'gradient' and 'slope' are
sometimes used interchangeably, but here the convention of dei ning slope to comprise
Search WWH ::
Custom Search