4.2 Acquiring and Preparing Data for PSDB 
4.2.1 Remotely Sensed Data: Landsat TM 
images were processed using image processing 
system as follows: 
a. Two tapes containing the images of the 
study area for years 1990 and 1993 
were acquired and have been used to 
achieve the aim of this study. 
The extent (patches) of the study area 
was determined then down loaded to the 
hard disk. 
Image enhancements such as histogram 
equalization were applied on the images 
in order to enhance the distinction 
between its features. 
Preparing color composites were 
performed using standard bands of Red, 
Green, and Blue (RGB), respectively. 
determining best band combinations to 
the to insure feature separability. 
Performing classifications for selected 
band combinations using unsupervised 
modules and post-supervised 
classification techniques for (1990) and 
(1993) TM images. 
Obtained data were prepared for coding 
them in the knowledge base of the 
system in a qualitative sense. 
4.2.2 Laboratory Data: Field trips to the study 
area and collecting samples for laboratory 
testing were conducted. Two soil samples, 
from two different areas with two different 
characteristics were collected and brought to the 
environmental and soil laboratories at the civil 
engineering department at KSU. 
Two major laboratory tests were performed. 
The first was the classification test for different 
soil types. The other test was a comprehensive 
test that last for about ten days. It covered all 
aspects of swell/shrinkage properties of the soil 
samples. It should be stated here that sample 
locations were defined on aerial photographs as 
well on the satellite images. Finally, these 
data were prepared carefully to be suitable for 
coding proper information in the knowledge 
base. The field data, also used as a real-world 
reference data for testing purposes as well as 
53 
for verifying identities of unsupervised classes. 
Table 2 is a sample of laboratory soil tests. 
5. DISCUSSION AND RESULTS 
According to the findings of the laboratory test 
and image interpretation, soil was classified 
into clay, silt, and sand. Comparing these 
types in (A) and (B) sites showed that (A) 
contained high percentage clay, whereas it was 
less in (B). On the other hand, the soil 
swelling test showed that both areas were 
subjected to swelling. However, the swelling 
was higher in (A) than it was in (B) due to the 
existence of the mineral "montmorillonite" in 
site A. Montmorillonite is a mineral that cause 
soil swelling under certain weather conditions 
and with water availability. It should be stated 
here that the absence of montmorillonite does 
not necessarily mean the absence of swelling. 
A correlation between swelling soils that were 
approved by lab tests and interpreted on TM 
images was verified. ^ Accordingly, from 
interpreting images and re-entering its input to 
the 'SDB system soil swelling behaviors can be 
reported. 
In case that soil is permanently dry, or 
permanently saturated, no soil swelling are 
expected in areas that contain factors causing 
swelling; that is stable weather condition 
prevents soil change. 
I’SDB contains many recommendations each of 
which is a guidance for civil engineers on how 
to treat sites of projects according to the 
degree of the damaging behavior of bed soils. 
For instance, having proper slopes to prevent 
Still water from accumulating at or near 
structures, preparing very efficient draining 
systems to prevent water from reaching 
foundation soils of structures, and laying the 
structures on separate foundations on resistible 
Soil with large factor of safety are samples of 
recommendations contained in PSDB in case of 
identifying expansive soils. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996