Geoscience Reference
In-Depth Information
fulfill the purpose stated above; among these generalization operators there
is typification.
According to (McMaster and Shea 1989), typification replaces a large
number of objects by a smaller number of objects while trying to ensure
that the typical spatial structure of the objects is preserved. From this defi-
nition the two salient features that distinct typification from other generali-
zation operators stem out clear:
1.
it works on patterns
2.
it creates new objects
These two abilities, both to identify patterns and to draw new objects,
make typification algorithms complex and very difficult to develop.
Working on patterns implies that the typification algorithm should be able
to identify in the source data some information -the spatial structures- that
are implicitly coded in the characteristics of the geometries like direction,
dimension and shape in general. Because of the countless combination of
such characteristics, the extraction of patterns from geometries is almost an
impossible task if it is not constrained to just some characteristic (e.g.
alignment (Christophe and Ruas 2002)) or some type of patterns (e.g. grids
(Heinzle et al. 2005)).
The generation of new objects is too a very challenging task for an algo-
rithm, even more than pattern recognition if possible. To create a new
object the algorithm should be able to decide autonomously the shape, the
dimension, the direction and the position of it. Furthermore a deep under-
standing of the constraints on the new object should be embedded in the
algorithm that should be able to evaluate any topological and geometric
constraints imposed both by the semantic and the surrounding environ-
ment
1
.
In general we can say that the difficulty in developing a good typification
algorithm is related to the fact that the task that such an algorithm should
pursue, to identify the pattern that characterize a certain cartographic
phenomena and replace this phenomena with a new representation that
1
In this paper we will perform typification creating from scratch new geometries that will
replace the original ones. But typification could also be performed by selection and dis-
placement, i.e. deleting some elements and arranging the remaining ones in order to main-
tain the pattern. Nevertheless, in both cases, the considerations on the complexity of any
typification algorithm apply.
