administrative units where the evaluation is provided and
considered the relations and contextual and temporal
background for the data processing.
Object-oriented image analysis is based on an object-oriented
approach to image analysis. In contrast to the classical image
processing method, the basic processing units are image
objects or segments. Further motivation for the object-
oriented approach is the fact that the expected result of most
image analysis tasks is a set of real world objects. The
topological relations of single or adjacent pixels are given by
the raster implicitly. The association of adjacent image
objects must be explicitly worked out, in order to address
neighboured objects.
The increasing resolution of the sources results in the
increasing number of objects of course and moreover the
complexity of object structuring hierarchy is rapidly growing
too [9]. We consider: the house as well as the block of
houses, or the type of the urban site. It is very similar in case
of trees: scattered trees can create one object and dense
standing trees another one and what about the forest? 'The
huge amount of data asks for automation of the classification
and interpretation processes of the spatial data.
There are many reasons for the knowledge-based spatial data
network building and sharing. The pixel-oriented
classification can be accepted as the pre-processing phase
that is followed by object-oriented contextual classification.
This opinion is supported by the long experience with texture
analysis and it represents only the local type of context
incorporation into the processing and it represents the lowest
level of contextual modelling.
2.2 Object hierarchy
The average resolution of image objects can be adapted to the
scale of interest and resulting information can be represented
in the scale based on the average size of image objects. This
fact is coherent with the hierarchical networking and
representation of image objects. In the hierarchical structure
each object knows its neighbours, sub objects and super
objects.
We distinguish three types of elemental object features:
attribute (physical properties of objects that follow from real
world or image or different information layers related to the
object), fopological features (describe the geometric
relationships between the objects or the whole scene) and
context features (represent the objects’ semantic
relationships. The gardens are inside the urban area; the
island is surrounded by the water, and so on. It means
between class (object) relationships.
The features are the means to assign an object to the certain
class. They can be used simultaneously or separately as well
as in exactly defined hierarchy and their essence comes from
spectral, textural, local, global, temporal, contextual or any
other of the available property of image entity.
They help, together with the class hierarchy that makes
possible to use the inheritance, to create groups of the object
and set up the structure, to give additional meaning to the
classes of objects. The class hierarchy defines the
requirements that an object must meet to be assigned to the
certain class — the first step of segmentation.
124
The class hierarchy, which can be continually changed during
the processing, it is the base of knowledge for the image
object classification. In many applications the desired geo-
information and the objects of interest are than extracted step
by step, by iterative process of classifying and processing. It
is very similar to human image understanding processes. This
kind of circular processing results in a sequence of partial
states, with an increasing differentiation of the classification
result and the increasing abstraction of the original image
information.
a
Figure 1. Aerial data - identification of swimming pools.
On the each step of the abstraction new information and new
knowledge is generated and can be used beneficial for the
next analysis step. High beneficial is the fact, that after
successful analysis, a lot of interesting, additional
information can be derived.
The class hierarchy approach is also very good tool to test the
ability of students to understand the task and process the final
study for the subject GIS.
I- | =
| | PT — rr Ld j
i | i F E |
js | | 7 =
| Xi ! E pe
| t | af)
A CH
— |
Figure 2. Swimming pools compared with parcels —
identification of owners.
The different segmentation techniques can be used to
construct a hierarchical network of image objects and each
level in this hierarchical network is produced by a single
segmentation run.
The h
informe
simulta
structut
neighb
the clas
(attribu
operatic
depend
size of
semant
Inherit
Class d
to their
more tl
hierarcl
comple
Group:
It is the
meanin,
functioi
automa
the gro
group.
Structu
Differs
structur
Differei
basis fi
make p
to sing]
The ir
comple:
inherita
differen
group’s
resultin;
2.3 Ap
The lo
develop
approac
monitor
paramet
Agriculi
Land u
Infrasti
And Int
control
control.
One exa