of the American Society for Information Science

Go to
 Bulletin Index

Volume 26, No. 1

bookstore2Go to the ASIS Bookstore

October  / November 1999



Digital Imagery in Global Disaster Information

by Einar Bjorgo

Editor's Note : Margarita S. Studemeister, director of the Jeanette Rankin Library Program at the U.S. Institute of Peace in Washington, DC, served as guest editor for this special section of the Bulletin. She can be reached by mail at the U.S. Institute of Peace, 1550 M Street, NW, Suite 700, Washington, DC 20005-1708; by phone at 202/429-3850; or by e-mail at mss@usip.org

Very High Resolution Satellites: A New Source of Information in Humanitarian Relief Operations

by Einar Bjorgo

Geographic information is one of the key parameters that need to be assessed in humanitarian relief operations. Disasters often occur in less developed regions, where existing map material tends to be poor, if it exists at all. When relief agencies take on a humanitarian mission, they need updated information on such things as road networks, water sources, possible aircraft landing field and, of course, the number of people in need of assistance and where they are located. This information could in many cases be provided by military sources. Unfortunately, this task seems not to be routine in complex emergencies. Fortunately, a new source for providing relief agencies with up-to-date information is about to be available: commercial very high resolution satellites.

Satellite imagery has in general not been used by the relief community on a routine basis. There might be various reasons, but the relatively high level of specialist interpretation traditionally needed to process and analyze satellite images, in combination with relatively high costs, are important reasons why the relief community chose not to utilize this technology extensively in the past. Furthermore, the amount of detail that can be distinguished in images acquired for civilian purposes has been relatively low, which has also been an important limitation for the relief community.

The Development of Satellite Imagery for Non-Military Use

Satellite images were originally a tool for military specialists and civilian researchers who to a large extent were also involved in designing the technical characteristics of the various satellite sensors. The resulting images captured by these satellites were then analyzed by trained staff and interpreted by experts. Thus, the resulting image products were not aimed at people outside the military and scientific community. However, a gradual change occurred as the scientific studies were verified and various applications proved beneficial, such as within vegetation monitoring and geologic surveys.

Since the end of the cold war, declassified, but degraded, versions of older Russian intelligence satellite imagery have been commercially available. However, these images are analog data with limited coverage and have not been utilized extensively. They have a spatial resolution of 2 meters, that is, objects as small as 2 meters across can be detected in the images. To respond to the demand for better data, the U.S. government allowed commercial companies to build, launch and operate satellites to disseminate images with spatial resolutions down to 1 meter. Figure 1 illustrates the improvement in detail that can be detected when comparing imagery at spatial resolutions of 10, 3 and 1 meters. Until now, the most widely used sensors for civilian applications have been those on board the U.S. Landsat satellite at 30 meters resolution and European SPOT satellite at 10 meters resolution. Lately, the Indian IRS 6 meter resolution satellite sensor has also proved useful.

The combination of improved spatial resolution and less technically demanding processes for analyzing images might prove very useful for relief agencies. Image providers and value adding service providers also need to understand the needs of the relief agencies. For staff members to understand the satellite images and information derived therefrom, it is useful if images are in near-natural colors and have a very high resolution, resembling aerial photographs, which most people can relate to. That way, relief staff in the field can also do their own interpretation utilizing hard copies or standard off-the-shelf image processing software or simple photo editors freely available. Thus relief staff can complement and verify information that is difficult to interpret away from the field, such as at a regional office or headquarters.

As mentioned above, commercial very high resolution satellites now under development can detect objects as small as 1 meter across. At this resolution individual tents and houses can be distinguished in a satellite image, larger groups of people can be located and the numbers of people can be estimated. It is important to note that no individuals can be detected or identified with this type of satellite imagery. The ReliefSat (www.nrsc.no/reliefsat) and ENVIREF (www.enviref.org) studies, funded by U.S. Institute of Peace and European Commission, illustrate how historic Russian very high resolution satellite imagery can be used to map road networks, water sources, rivers, vegetation and individual refugee camps. Figure 2 shows a detail of the now demolished Site 2 refugee camp in Thailand. This figure shows the dense refugee housing and relatively organized campsite as acquired in a 3.3-meter resolution near-natural color satellite image. Individual trees and buildings can be detected in this image. However, much of the refugee housing is so dense that individual tents are difficult to detect.  (Note to readers: to see the color version of this photograph, please refer to the Bulletin Web site at asis.org.)

Applications and Limitations

The applications of such images are not restricted to emergency efforts. One of the most promising ways of utilizing these upcoming satellites is for refugee camp planning. The average refugee camp lasts for approximately seven years, and well-planned camps are important for sustainable use and more controlled environmental impacts from the many persons settled there. In order to provide decision-makers with realistic views, as well as take possibilities for landslides and flooding into account, terrain models of the planned refugee hosting area is necessary. Such models can be derived from satellite imagery into so-called Digital Terrain Models (DTMs). Using standard image-processing software, high-resolution satellite images can be draped over the DTM. Thus, complex humanitarian relief operations can utilize various products and methods related to satellite image technologies to get updated and objective information on the environment. Figure 3 shows how a DTM with a high-resolution satellite image draped over can illustrate a camp's environment. The figure shows refugee camps in Kukes, northern Albania, during the recent Kosovo situation and how these camps were located along the road from Kukes airport to the City of Kukes.

However, it is important to identify clearly the various limitations with this technique for the relief agencies. Satellite image applications have had a tendency to be oversold in the past, and the Earth Observation (EO) community must be careful not to promise too much to relief agencies with very limited budgets. The near-future very high resolution satellite sensors are all designed to acquire information in the visual and near-infrared (IR) part of the electromagnetic spectrum. Consequentially, during cloudy conditions, ground features cannot be detected in the imagery. Cloud cover is particularly a problem in Central Africa and many equatorial regions, but also at higher latitudes, it can significantly limit the amount of useful satellite images available. During the recent conflict in Kosovo, it took several weeks until a high-resolution cloudless image of the refugee hosting area was available. This situation is expected to improve as more and more satellites orbit the earth, and thus the possibility of acquiring a cloud-free scene when there is a gap in the cloud cover increases.

Another major limitation for the relief agencies is the cost of imagery. Even though the images may be cost efficient, this assumption still needs to be proven and quantified to some extent before at least smaller relief organizations are willing to invest $3000 in an image. One way of reducing this problem may be for relief organizations to cooperate in image purchasing. Interested organizations, such as InterAction, should be able to make arrangements with major data providers on behalf of their members. Thus, each member can receive an image or image-derived information at significantly reduced rates. Arrangements such as that recently made between the SPOT Image satellite company and the United Nations (UN) on significantly reduced rates for UN organizations are very promising.

Management and Policy Concerns

This new type of very high resolution satellite imagery also raises some policy concerns related to relief staff security, reliability of access to imagery, and inter- and intra-institutional information dissemination. Relief organizations should develop policies regarding what type of images and information should be disseminated to field staff. A hypothetical situation where field staff members have received a recent detailed satellite image showing the location of refugees, but also national military personnel, might jeopardize the security of the staff should the military forces demand to see the image. Thus, in some cases it may be better to disseminate only satellite-image-derived information directly relevant to the relief operation by means of such things as standard maps with annotations or simple text messages.

Another issue is that of so-called "shutter control." According to the license granted the various U.S. companies launching commercial satellites, the U.S. government has the right to limit access to the imagery in situations where U.S. interests are at stake. This provision raises an interesting question: If these satellites had been launched before the Kosovo crisis, would the U.S. government have enforced "shutter control" to protect the NATO mission? Even though there are other non-U.S. companies scheduled to operate very high resolution satellites, one can image a situation where the United States at least suggests that these companies not distribute imagery, as was the case with the European SPOT satellite system during the Gulf War. No SPOT images of the conflict area were then available to the public.

The issue of information dissemination within and between relief organizations should also be addressed. Since Internet technology facilitates simultaneous distribution of satellite-derived information to all levels within an organization or group of organizations, standard procedures will need to be decided upon prior to image order. Many relief organizations have a top-down information structure, which can be useful for security issues. However, such a structure may also delay important satellite-derived information on such things as the number of refugees and where they are located from reaching field staff. This issue needs to be resolved within each agency.

In order to streamline image acquisition, analysis and information dissemination it is important to develop relationships among image providers, value adding companies and relief agencies. Previous efforts clearly show how important these relationships are. The time from image acquisition until the end product has been disseminated to the relief staff has been much too long to have any practical use in humanitarian emergencies. It is important to realize that the technology to achieve rapid dissemination exists, but the institutional aspects need to be streamlined.

Proving the Case for Satellite Data in Humanitarian Relief Operations

One of the most important issues for the near future is to prove, as much as possible, the cost efficiency of the satellite data in humanitarian relief operations. The EO community developing new products and methods for utilization of very high resolution imagery will likely need one good case to prove the full extent of this technology. When this event might occur is difficult to estimate. First the new satellites, such as Ikonos-2, OrbView-3 and QuickBird, need to be launched; then a scenario has to develop where these satellites are used successfully. There is still some way to go before the scientific EO community will have developed pre-operational tools, but once the methods are streamlined, the results seem very promising. Hopefully, within two to three years, relief organizations will have near real time access to detailed satellite-derived information, making relief operations more efficient.


Figure 1: Improvements in spatial resolution. Left: 10 meters; middle: 3 meters;  right: 1 meter. The amount of detail that can be detected increases by a factor of 100 when resolution improves from 10 to 1 meters. [Resampled aerial photography of Capitol Hill. Image copyright Orbimage 1995.]

Figure 2: Detail from Site 2 refugee camp, Thailand. Image copyright Sovinformsputnik 1998.

Applications and Limitations

Figure 3: Preliminary results from ReliefSat and ENVIREF studies: Radar satellite image derived DTM combined with recent 6-meter high-resolution optical image. Kukes refugee camps in northern Albania. Image acquisition 27 May 1999. Image copyright: ESA/Eurimage 1999. Image processing and analysis: NERSC/ERDAS.

Einar Bjorgo  is with the Nansen Environmental and Remote Sensing Center in Bergen, Norway. He can be reached by e-mail at einar.bjorgo@nrsc.no


How to Order

@ 1999, American Society for Information Science