How can you decide if a country is rich according to satellite images? In this article, you will find the answer!
Can a machine predict income of the particular region of the Earth by using satellites? The Journal Science published a new study, where a combination of a certain algorithm and satellites images may provide a humanity a prediction machine for poor regions. It`s roughly predictable that nighttime pictures of satellites may show you a wealth area, but how to count income in areas where the night light is limited? That was a problem to face, but with the help of Neal Jean, who is the lead author of a new method of processing data. This method uses nighttime images and daytime available data in combination.
READ ALSO: What is the largest church in the world?
The overall algorithm of satellite images can be divided into two parts, and now the team of scientist used data of Malawi, Rwanda, Nigeria, Tanzania and Uganda. The satellites provide a survey of day and night data of these countries. In this phase, the machine searches for the houses and tries to predict whether there would be a light during the nighttime in the particular house. The second part is adding economic context to a collected date. This helps the system to recognize villages and houses, districts. If there is a nearby village with a similar pattern in the database, the machine can process the data according to the similarities.
This new model of calculating the overall income of households is certainly better than simple night pictures of rich regions. Nevertheless, it has limitations. The system can`t find similarities off economic patterns in dense urban areas. Therefore, it may help providing the data from the satellites of agricultural areas mostly. Right now, the algorithm is being tested in five African countries. However, the constructors of the machine hope that with the available internet data and pictures of the satellites, the machine will be able to develop and enhance the methodology of the survey on its own.