|
|||||||||||||||||
David
B. Davis
Department of Geography University of Utah Salt Lake City, UT 84112 USA carpincho@att.net Implementing GIS in Developing Countries: A Case Study in Uruguay Abstract:
This research was done to determine the feasibility of creating a fully
functional GIS in a developing country of the world using inexpensive hardware
and software with readily available data.
Although this study involved a city in Latin America the methods used
should work well in almost any other developing country of the world. IDRISI and ArcView software packages were
used because of their availability in various languages, price discounts for
research, experience throughout the world, and ease of use. The majority of the data was obtained from
Uruguayan military mapping and imagery agencies. All of the data was written to a CD-ROM
for ease of use and distribution. This
data was presented to various civic and government organizations in the city of
Sarandi del Yi in 1999. A public
planning committee is currently using the data. The city government and the water department are planning to
begin using the data this year. INTRODUCTION The improving economic and political conditions throughout much of the world, most notably in Asia and Latin America are resulting in a more dynamic global population. People are able to move from place to place more easily than ever before in search of education, employment, and improved living conditions and opportunities. This movement or migration can be from country to country but more often is from city to city or from one part of a city to another. All of
this movement of people in and out as well as within cities has placed great
demands on all levels of government.
Trying to keep track of the population and their vital records is only
part of the difficulty. The myriad
changes to the infrastructure and the environment associated with the movement
of people or totally new settlements have been almost impossible to monitor and
update.
Traditionally this monitoring and record keeping has been done using
paper maps, paper records and other filing systems, aerial photography, and
field mapping. These static record
keeping and data systems become outdated and inaccurate almost
immediately. These difficulties and
challenges are basically the same whether dealing with a small city in Uruguay
or the largest of cities such as Mexico City.
Even if a city has the resources to obtain new data such as aerial
photography on a regular basis they often times do not have the manpower or
technical skills needed to integrate this data into their records. During
the past decade new and improving technologies and methodologies have made most
or even all of this monitoring or record keeping much more feasible. Government and now private company remote
sensing data, high capacity yet lower cost computers, easier and more powerful
software, GPS, and the Internet are all helping to make this possible. Much of this data is available on a
consistent, real time basis. Because it
is often in digital form it can be immediately integrated into a computer
system. With increasing demands and
tightening budgets governments are almost obligated to investigate and
implement as much of these new technologies and systems as possible. This project was an attempt to show that
the new and improved data and technologies available today have made it
possible for city governments in developing countries to keep up with
population and infrastructure dynamics.
The city of Sarandi del Yi, Uruguay was used as a case study or model of
how this can be done. RESEARCH DESIGN
This
research was done to determine a process, not the process for GIS
implementation in developing countries.
Many of the steps and procedures are similar to those involved in
setting up a GIS in the USA although several modifications were needed. Differences in culture, technical abilities,
governments, language, and training all needed to be incorporated into the overall
process. The processes used in this research
project can be divided into three distinct phases: determining the current
capabilities and data, determining the needs for implementation, and the actual
implementation. When determining the
capabilities and data currently available to the city government such things as
quantity and type of computer equipment, maps, and technical skills of the
employees were evaluated. Five items
were looked into regarding what would be needed for GIS implementation. The first was cooperation among the various
government agencies and utilities in the city, something that is usually easier
said than done. The second item was
what computer equipment would be needed now and what could wait until later in
the process. The third item was closely
associated with the second, being what software would do what was needed and
still be affordable. Data was the
fourth item to be investigated. The
city had paper maps that were drawn to scale but were not georeferenced. A georeferenced map of the city was
available from the Uruguayan Military Geographic Service and aerial photography
from the Uruguayan Air Force remote sensing division. The last item to be evaluated was determining what training was
needed, who would receive the training, and who would provide the training. None of the city government employees have a
university level education. CREATION OF THE RASTER DATA As was mentioned earlier, data was
available from the city as well as from Uruguayan military sources. Maps and aerial photography was obtained on
a trip to Uruguay in 1994, as well as through in-laws that sent newer maps and
aerial photography in 1997. The
following maps were obtained for use in this project: 1:50000-scale topographic map, 1:10000-scale city street map,
1:20000-scale soils map, sewer map, and water lines map. The street map also displayed the ID number
for each of the city blocks. Black and
white aerial photography was obtained at two different scales and time
periods. The oldest photography is at a
scale of 1:20000 and was flown in 1967.
The newer photography was from 1994 and is at a scale of 1:10000. All of the maps and aerial photography
was scanned at 400 dpi and 256 shades of gray.
This dpi was used to achieve a good viewing resolution while at the same
time manageable file sizes. Each of the
9” by 9” photographs were approximately 17 MB in size. The scanned topographic map had a much
larger file size of nearly 62 MB. Since
this map was much larger than the city boundaries a subset of the scanned map
was “cut out” digitally. This resulted
in a map of the city area with a much smaller file size of only 19 MB. After evaluating many different GIS and/or image processing packages IDRISI was chosen as the best system for this project for the price. The capabilities of IDRISI for GIS and image processing were as good or better than any of the other systems. None of the other organizations had the experience in developing countries like the makers of IDRISI (Clark University, Worcester, MA) and none could even come close in price. In October of 1995 IDRISI version I for
windows was purchased at the student discount price of $247.50, a 75 percent
discount. The IDRISI version II upgrade
was purchased and installed in March of 1998.
The Spanish language module of IDRISI was installed in April of
1998. The software can now be used in
either English or Spanish with a simple menu selection. Language modules for IDRISI are available at
no cost for several different languages.
Currently modules are available for Spanish, Chinese, French,
Portuguese, and Russian. With the data and software in the
computer the GIS layers could now be started on. Histogram equalization was performed on the aerial photography to
lighten the images. This made
everything easier to identify and more pleasing to the viewer. The next step was one of the most important
goals of the project, to georeference all of the raster and vector layers. Because the maps the city agencies were
using were not georeferenced a means for obtaining control points needed to be
found. Various sources for control were
evaluated. These included Landsat and
SPOT imagery, GPS, and other georeferenced maps. Since one of the main goals of this research was to try to use
inexpensive and easily obtainable data the only viable option was to use the
1:50000-scale topographic map obtained earlier. This map has the UTM grid clearly marked and numbered. The IDRISI software makes it very simple
to georeference or digitally rectify maps or imagery. This process was first done on the scanned topographic map. The x, y coordinates on the map were
replaced with the UTM coordinates. This
was done by placing the cursor on the UTM intersection lines on the screen and
entering these values into a table. The
UTM values for these same intersections were noted from the map and also placed
into the table. With these values the
software was able to georeference the scanned map. To digitally rectify the aerial photography a similar process was
used. By displaying the georeferenced
map and an aerial photograph side by side on the screen control points were
easily selected. By placing the cursor
on a road intersection or other feature that could be seen on both the map and
photograph the x, y and UTM values could be determined. As many as 46 control points were selected
for each photograph. These images could
then be viewed in IDRISI. All of the
maps and imagery were also exported as TIFF images with an accompanying TIFF
World file. This enables the images to
be viewed in practically any GIS or image processing software. The image below shows a portion of the
base map with an overlay of two of the digitally rectified aerial
photographs. As can be seen in Figure 1
the roads, hydrology, and other features visible on the photography line up
quite well with the underlying map. All
of the layers, both vector and raster were referenced to the map, not to the
aerial photography.
Figure 1. Rectified Aerial Photography with Base Map. CREATION OF THE VECTOR DATA The ability to digitize on screen was preferred to doing so on a digitizing table. The emphasis of this project was to do everything possible with the data and equipment on hand. On screen digitizing would eliminate the need for a digitizing table. Not only for this portion of the project but also for the city should they make updates or create new data layers in the future. Although the IDRISI software does have the ability to do on screen digitizing it is quite limited. Research was done to determine if relatively inexpensive software with better on screen digitizing capabilities could be found. ArcView GIS version 3.1 was evaluated first because of previous experience with this software. Also a 30-day evaluation copy of the program was available for testing. After successfully digitizing and attributing a railroad line and several roads it was decided to contact ESRI for further information about purchasing this product. ESRI offers a substantial discount on the purchase of their products to university students and faculty when used for educational research. The regular price for ArcView GIS 3.1 is $1,195 where as the educational discount lowers the price to a much more accessible $250. The following layers of information were digitized and attributed: - water lines - sewer lines - streets - rivers - flood zones - city block numbers- utilities - city parks - government buildings - various (banks, clubs, businesses) - schools - churches - train track - route 6 - route 14.
Kollin, Cheryl, “Growth
Surge: Nationwide Survey Reveals GIS Soaring in Local Governments”, Geo Info
Systems, Vol. 8, No. 2, (1998), pp. 24-30. |
|
||||||||||||||||
|
|||||||||||||||||
|
|||||||||||||||||
