Geography Reference
In-Depth Information
Crisp Set Analysis of CCM
The final speed of given vehicle was evaluated according to formulas ( 7 ) and ( 8 ).
Particular obstacles were compared with the given vehicle properties in a pixel of
1 by 1 m. Total of 27 raster layers were created for each particular coefficient C ij
where the value of given pixel was the “cost” of the pixel—reclassified value of
coefficient to the range of 0-1.
The final cost map was created using map algebra from combination of all
particular layers. This cost map enables to find the cheapest way from start point
to destination.
Precision, uncertainty, vagueness and similar properties were not taken into
account in the calculations of cost map because of crisp set analysis. The final
result is easily understandable but the great disadvantage of such analysis is that
user has no information about the properties of features entering calculations. While
determination of feature borders can be different (for example, the footprint of
building can be defined in resolution of centimetres, the border between two types
of soil is defined in resolution of approximately 100 m), therefore it is necessary to
include other parameters in the calculation. One possible approach is application of
a fuzzy logic.
Fuzzy Set Analysis of CCM
Certain processes that differ only in the input conditions are repeated with all
coefficients. These are selection processes as well as conversional ones, etc. For
solution of vagueness, in-built processes Fuzzy Membership and Fuzzy Overlay are
used. While the use of Fuzzy Membership is different, Fuzzy Overlay is the same
for all coefficients. Conditions for Fuzzy Membership are dependent on geometric
accuracy that shows the types of geographical objects in the database. There is only
the example of the coefficient C 1 —Terrain relief calculation in the following text.
Coefficient C 1 is very complex, it includes the influence of slope gradient as well
as the influence of micro-reliefs. The slope gradient is one of the input layers and it
can be—just like in this case—prepared from the actual factor calculation. The
calculation is divided into several branches.
In one branch, fuzzification is calculated directly from the values of gradient
(expressed in degrees). The limiting value for the chosen vehicle (36 ) was taken
for the calculation. The conditions for fuzzification are as follows
8
<
:
0,
x ¼ 0
36 x
36
μ S ðÞ¼
ð 9 Þ
, 0
<
x
<
36
1,
x 36 :
Other branches of calculation concern micro-relief. The input layers are point, line,
Search WWH ::




Custom Search