2. IMAGES AND EXPERT SYSTEMS
Aerial photographs are a sort of remote sensing
product. Remote sensing can be described as an
art, science, and tool of acquiring information
about targets or certain phenomenon by a
sensor that is not in contact with these targets
or phenomenon (Cracknell, et al., 1993;
Curran, 1985; Lillesand, et al, 1987).
Accordingly, remote sensing is an acquisition
tool, that can be used to investigate soils and
other terrain parameters.
Interpretation process of space images is
conducted according to systematic operations
using low level and high level interpretation
elements (Al-Garni, 1994; Carbone, et al.,
1996). Key elements of interpretation contain
an intelligent data-base of robust experience
built throughout many years of experience in
the minds of professional human interpreters.
Terrain analysis techniques for image
interpretation is a powerful tool (Al-garni, 1992
and Way, 1973; Mintzer, 1989).
A.I. is described as the study of how to make
computers do things which at the moment
people do better (Rich, et al., 1991). Expert
systems are the applied portion of the A.I.
science through which A.I. programs can code
the experience of an expert in a field and
transfer his experience to needed person
(Bowerman, et al., 198; Hayes, et al., 1983).
3. SYSTEM DEVELOPMENT
Even though the total development of the PSDB
system is an increment process with overlapped
phases, there are five ordered phases of
developing the system explained as follows:
3.1 Identifying the Problem
Expansive soils can be identified using tools
such as stereo-aerial photographs, satellite
images, radar images, and laboratory testing of
soils. However, very experienced group of
people are required to conduct proper tests
from the above list of sources. Accordingly,
50
inexpensive tool that are able to guide a civil
engineer should be available to help him
conducting all tests by himself. This can be
achieved through an intelligent knowledge base.
3.2 Data Acquisition
Real geographic location that contains soil
problem must be selected for PSDB. For this
study, an area in Saudi Arabia called Tabouk
that has soil problems was selected. Then all
possible information that can be obtained about
expansive and regular soils of Tabouk were
acquired. This includes:
Maps from Ministry of Municipality and Rural
Affairs (MOMRA), Control points from
MOMRA, Aerial photographs from Military
Survey Department (MSD), and Satellite
images from King Abdulaziz City for
Science and Technology (KACST).
3.3 Data Preparation
Very careful analysis and investigation of the
soils was conducted. This was performed using
manual and automatic techniques of
interpretation. First, visual interpretation of
stereo coverages of aerial photographs and
maps was applied. Some visual factors or key
elements of interpretation such as drainage,
tone, and associations were extracted.
Second, satellite images for the area at different
dates (1990 and 1993) were processed using
proper image processing tools. Classification
of soils, structures, and vegetation were
conducted carefully. All these sets of data were
prepared in suitable forms in order to code
them in an object-oriented intelligent data base.
3.4 Data Presentation
Data presentation is an expression of building
the structure of the intelligent data base. It is
similar to regular data base structures, fields,
and records. In object oriented expert systems
there are classes, attributes, and instances.
These three elements were the basic elements of
structuring the knowledge base of PSDB.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
The se
system
method
an expe
interpre
in this
The pri
a numt
was th
topogra
and lab:
used in
in conc
ELSE).
rules in
IF The 1
AND The
AND The
Gullies :
Vegetatio
: = "The
3.5 Sys
The sy
evaluati
editing
knowlec
system
complet
evaluati
compare
expert).
system \
to the er
4. |
The maj:
input dat
output ai
these co
The sets
interpret
Was assi;
The task
