Using AI to control energy for indoor agriculture
30 September 2024
Published online 21 March 2011
Scientists use satellite imagery to document regime's crimes against civilians.
The Gadaffi regime in Libya announced a ceasefire in its military action against protestors today, after the United Nations Security Council last night agreed on a resolution to authorize a no-fly zone over the country and air strikes against the regime's military assets.
The international community's action comes as the regime's forces march on Benghazi, Libya's second-largest city, which is currently held by protestors. The town is the seat of the Interim Transitional National Council, which has been recognized by France, the United Kingdom, Portugal and the Arab League as Libya's legitimate government. Many fear that if Benghazi were to fall, the regime would inflict massive retribution and human-rights abuses on the city's population.
Scientists and human-rights activists are using satellite images and geographical information systems to document abuses committed by the Gadaffi regime since it began to brutally crack down on protests last month. Their data could be vital in bringing the regime to justice — in February, the United Nations referred senior regime officials to the International Criminal Court, to face charges of possible crimes against humanity. Susan Wolfinbarger, a senior programme associate at the Geospatial Technologies and Human Rights project of the American Association for the Advancement of Science in Washington DC tells Nature how her team is using technology to hold the Gadaffi regime to account, and monitor other oppressive governments.
How do you go about demonstrating human-rights abuses using satellite images?
One of the biggest problems in documenting abuses in Libya has been that observers are being prevented from entering the towns, but with satellite imagery we can document the destruction and see whether civilian sites have been targeted. We collect existing on-the-ground information about the locations of buildings such as schools and hospitals, so that we can find them on images. We don't need lots of satellite pictures, just recent ones of specific areas, which we compare with archival images to assess the occurrence of attacks such as air strikes.
It takes a lot of patience, and specialized software. We look, for example, for evidence of burning, the presence of shell craters, whether roofs are intact and so on. During past work in Sri Lanka, we were able to pick out craters from different types of mortars.
We can also look for mass graves. We found one during a project on Afghanistan, but it is very difficult, and you need on-the-ground intelligence to tell you where to look. You can then check for things like ground disturbance, and changes in the colour or shadowing of terrain. Our focus is on human-rights abuses during armed conflicts, forced displacements and the situation of refugees — for example, by inspecting whether tents show damage. We try to provide scientific evidence that documents abuses.
How available are the satellite images that you need?
The situation has got way better even since I started at the project in 2006; things have become a lot easier. Satellite-image databases are much larger and cheaper. Google frequently updates its Google Earth and Maps services with high-resolution commercial imagery of zones that are experiencing humanitarian emergencies, and many other organizations also purchase and publish such pictures. Humanitarian and human-rights groups have also become far better coordinated. We all know each other now, and share information and images to avoid duplication.
We get most of our images from commercial suppliers, with whom we often have partnerships to get reduced prices — we work with GeoEye, DigitalGlobe, ImageSat International and Spot Image. The costs can add up fast: ordering a new image of even a small area can cost US$10,000 or more, although purchasing an image that a company already has in its database is cheaper — around US$400 for a zone measuring 5 kilometres by 5 kilometres. That is often all the area we need, because it covers most towns.
The resolution of images has improved. Back in the Landsat era, when resolution was in the tens of metres, you couldn't even contemplate the sort of work we do. There are still things that are too small too see, but it's much better than it was. Some commercial satellites are getting down to 40- or even 20-centimetre resolution, but the US government puts a limit of 50-centimetre resolution on commercial images. Having fine detail can be important — for example, when analysing images of shell craters.
What other challenges do you face?
In the Democratic Republic of Congo one of our main challenges has been that there is almost constant cloud cover. We tried using radar imagery but the results were disappointing. Radar imagery can be useful in other situations, however; it's great for quickly assessing large areas. Metal shows up brightly, which is useful for tracking human activity and habitations.
Another challenge is getting on-the-ground information to corroborate satellite data. It's best if you have reports from human-rights groups in the area: they can tell you who did what, when and where. In Libya, such information is currently lacking.
Funding is also an issue. We currently have a 3-year $750,000 grant from the Oak Foundation, which provides the sort of minimum annual budget we need. We are very good at stretching dollars.
Is your work making a difference?
Our analyses are currently being used in cases involving human rights abuses in three countries, before international courts: Darfur before the International Criminal Court; Georgia before the European Court of Human Rights, and Zimbabwe before the African Court on Human and Peoples' Rights. One of our major focuses in the coming years will be on ensuring that our analyses meet the necessary standards, norms and best practices to be used as evidence in court.
This article is reproduced with permission from Nature (18 March 2011) doi:10.1038/news.2011.172 Published online 18 March 2011
doi:10.1038/nmiddleeast.2011.35
Stay connected: