Artificial Intelligence used to convert maps to satellite images
November 1, 2020
Technology studies manage to develop a technique that can convert old maps into satellite images. This incredible technology is transforming the way map files can be viewed.
By checking the maps that were made in the past, it is possible to analyze what the environments were like years ago or even hundreds of years ago. This is a simpler way, with few details, as we are used to seeing in books.
One question that can be asked is how would it be if they were asked today? The quality and benefits would certainly be much better.
With this questioning, Henrique de Andrade, Polytechnic School of the University of Pernambuco, Recife, made a great analysis on several maps of different years of the city where he was born, Recife.
Henrique explains what gave him the strength to create this technology and what were his incentives:
“I gathered all these digital copies of maps and ended up discovering things about my hometown that is not so well known. I feel that in Recife people were denied access to their own past, which makes it difficult for them to understand who they are, they can do about their own future.”
A machine learning algorithm
This idea was presented to Professor Bruno Fernandes. The student’s idea was to create a machine learning algorithm that was able to make old maps turn into images similar to what Google does.
For the author of this research, this technology would help show people as many people who live in the places as in other parts of the world, what the changes have been over time.
With that, it would be visible what were the changes that happened over time, such as the social and economic impacts of urbanization. It is possible to analyze areas that used to be a large plantation of trees and are now occupied by buildings and houses.
This can also facilitate the creation of social projects, as it is easier to analyze the areas in more detail and thus make improvements in the regions.
Artificial intelligence in the project
To be able to plan and develop this technology, they used artificial intelligence (AI) and a technique called Pix2pix, which depends on 2 neural networks to work.
The first neural network is responsible for creating the images based on the input of information, while the second is responsible for analyzing the whole image and ensuring if it is false or not.
From these systems, the networks are developed to deceive each other, and after that, create the best and most realistic images based on the information that was entered.
Last month, Henrique and Bruno published the study in the IEEE Geoscience and Remote Sensing Letters. In this article, as a demonstration of the technology, a map of Recife from the year 1808 was shown to generate modern satellite images of the area. Henrique says:
“When you look at the images, you have a better idea of how the city has changed in 200 years. The geography of the city has changed dramatically – landfills have reduced water bodies and green areas have all been removed by human activity.”
Henrique mentions that the way that artificial intelligence works is very effective in the information input part since a large volume of information is not necessary.
However, some corrections need to be made. There needs to be a historical context to be able to check the inserted images and the quality of which the map is generated is inferior to what the developers are looking to create. Henrique says:
“Moving forward, we are working to improve the resolution of the images and testing various inputs.”
This technology still has a lot to be developed
It is still a new study, but still, the use of this technology and the improvements it can provide are very promising.
For the researcher, this tool that makes it possible to create maps is widely applicable and can be used in different locations in various locations.
Another factor that can also be very useful, according to Henrique, is that the images created by the satellites can be used by urban planners, anthropologists, and historians, for example, for future analyzes and developments.
Going deeper into this analysis of this technology, it may still have more advantages that we can take advantage of. These images can show us the environmental changes that affect the quality of life of people and with that we can analyze what was the cause of a certain problem that has been harming the population of a certain region for years, affecting the public health, studied by the sanitarians.