Poverty could be predicted by reading the satellite images using artificial intelligence. By indicating the areas where the most help is needed, these images could help eradicate global poverty.
One can make an idea of a country’s wealth by examining how much it shines at night. A comparison between China and South Korea’s intense brightness and North Korea’s dark mass could be one of the best examples found by the scientists.
However, the nighttime light cannot make differences between neighborhoods or villages within a large region. This kind of information could only be obtained by sending legions of survey-takers in populated rural areas. These actions usually cost millions of dollars and are very time-consuming, so, the researchers at Stanford have discovered a way to make computers and satellites do the work for them.
Their computer model proved to be far more effective than the methods that rely on data gathered from surveys and has a predictive power that can provide information.
The tests were conducted in five African countries: Nigeria, Tanzania, Uganda, Malawi, and Rwanda. Capturing images as part of the U.S. Air Force Defense Meteorological Satellite Program they then assumed that the brighter areas were more economically developed than those less enlightened.
A team of computer scientists and satellite experts created a self-updating world map to locate poverty. They used a type of artificial intelligence, a computer algorithm that recognizes signs of poverty through a process called machine learning. Marshall Burke, assistant professor in Stanford’s Department of Earth System Science said that the system shows the computer an image which it must identify.
The computer was then given data from the five African countries and then asked to use the data to find signs of poverty in a separate set of high-resolution daytime satellite images that contain information from poor regions that also appear dark in night photos. The computer could then link other things to poverty like roads, urban areas, farmlands, and waterways.
While the idea of machine learning goes back to the early days of computing in the 1950s, it has recently become a mainstream field of computer science.
How do you feel about computers being able to predict poverty through satellite images? Please let us know in the comment section below.