How Facebook's AI is scanning Satellite images to connect 10% of the World without Internet

“This is an impressive project from our team developing solar-powered planes for beaming down Internet connectivity and our AI research team, [sic]”

Facebook co-founder Mark Zuckerberg post on their use of AI to scan millions of satellite images to determine who has Internet access

  

Facebook is trying to connect the 10% of the world’s population that has no Internet access.

Facebook is looking into ways to more efficiently deploy their drones to deliver Internet to more people as noted in  “Facebook to beam Internet using drones”, published March 15, 2016 By Niranjan R, Internet Business Times.

For this reason, Facebook's Connectivity Lab in collaboration with Facebook's data science division, infrastructure unit, and machine learning and artificial intelligence groups, has begun to analyze satellite imagery for 20 countries as noted in “Facebook is using AI to make detailed maps of where people live”, published February 22, 2016 By Nick Statt, The Verge. 

This data, some 350TB of data or 51.6 billion images of satellite data, covers some 21.6 million square kilometers of places on Earth with no Internet Access. The 20 nation’s part of this exercise was as follows:

1.      Algeria

2.      Burkina Faso

3.      Cameroon

4.      Egypt

5.      Ethiopia

6.      Ghana

7.      India

8.      Ivory Coast

9.      Kenya

10.  Madagascar

11.  Mexico

12.  Mozambique

13.  Nigeria

14.  South Africa

15.  Sri Lanka

16.  Tanzania

17.  Turkey

18.  Uganda

19.  Ukraine

20.  Uzbekistan

So why exactly is Facebook doing this?

Facebook AI scanning Satellite images - How to scan satellite images of remote areas to help the Unconnected

It's all a part of their Internet.org initiative that uses drones, satellites, and lasers to deliver Internet to rural areas and developing countries. According to Facebook, this is more than just about getting more people to use Facebook, as the data can be used to help assess the risk for Natural Disaster: “We believe this data has many more impactful applications, such as socio-economic research and risk assessment for natural disasters”.

This method, which is assumed to be the least error-prone way of finding out who has Internet and who doesn’t, will be combined with data from Columbia University's Center for International Earth Science Information Network to create a more accurate population data set later in 2016.

This data is collated by an AI (Artificial Intelligence) program into a map showing areas where Internet access is lacking and helps Facebook determine whether Wi-Fi or Cellular Technologies can help to bridge the digital divide for this disconnected 10%. Once physical structures on the maps are added, Facebook can add census data, redistribute population data sets evenly across each location, to determine how many people lived in each building.

Google has similar ambitions as well using low orbit satellites to connect the last 10 percent of the Earth to their Google Services as explained in my blog article entitled “Google and O3B Network Limited partner on Satellite Broadband Project – How to set up a Satellite Broadband Network”. 

In so doing, it would help to power their Global Drone delivery ambitions as explained in my blog article entitled “How @Google X Labs Project Wing uses Global Satellite Network for Drone Delivery”. 

Hopefully, this effort will eventually result in free Internet to everyone on the planet, as Facebook stands to benefit immensely if they can cut out the telecom Providers and deliver Internet and Facebook directly to people who have no access.