Brothers who live in remote, remote areas will soon have a simple, easy-to-find home address placed by artificial intelligence . The system will be developed by Facebook and MIT , which uses the machine learning algorithm to identify and assign addresses to locations within the satellite imagery. Researchers believe that this approach will be cheaper, more efficient than other mapping techniques.
Statistics show that there are still about 4 billion people in the world without a clear physical address. That's why researchers want to find an effective way to solve that. And after the research, they trained a deep learning algorithm that allowed them to scan satellite images and identify pixels containing paths. At the same time, they use a different algorithm to get these pixels and to group them together into a network of paths that can be analyzed and broken down into quadrants.
When everything is "spread out" clearly, the algorithm will use the numbers and characters assigned to the paths and from there the addresses will be formed. Researchers say that with this technique intersections or nearby locations will be easy to identify because it is easy to rely on numbers and identifiers to find the connection between the roads. different.
In fact, this is not the first time researchers have sought to address the problem of addressing remote areas. In 2015, Google introduced the Open Location Code project, which aims to identify remote areas on their maps. In addition, what3words has also been studying how to divide a globe into a 3x3 meter grid, then randomly generate 3 phrases to name the cells.
And the biggest problem that many of the solutions that come up before the meeting is how to make the addresses they have easily created can be widely accepted. This is a real challenge because it requires approval in the community of the people and the government of the country. However, if successful application of a naming system, it will open the door to ease the delivery of medical care equipment, delivery, etc. On the other hand, it also supports the planning and build the infrastructure.
Statistics show that there are still about 4 billion people in the world without a clear physical address. That's why researchers want to find an effective way to solve that. And after the research, they trained a deep learning algorithm that allowed them to scan satellite images and identify pixels containing paths. At the same time, they use a different algorithm to get these pixels and to group them together into a network of paths that can be analyzed and broken down into quadrants.
When everything is "spread out" clearly, the algorithm will use the numbers and characters assigned to the paths and from there the addresses will be formed. Researchers say that with this technique intersections or nearby locations will be easy to identify because it is easy to rely on numbers and identifiers to find the connection between the roads. different.
In fact, this is not the first time researchers have sought to address the problem of addressing remote areas. In 2015, Google introduced the Open Location Code project, which aims to identify remote areas on their maps. In addition, what3words has also been studying how to divide a globe into a 3x3 meter grid, then randomly generate 3 phrases to name the cells.
And the biggest problem that many of the solutions that come up before the meeting is how to make the addresses they have easily created can be widely accepted. This is a real challenge because it requires approval in the community of the people and the government of the country. However, if successful application of a naming system, it will open the door to ease the delivery of medical care equipment, delivery, etc. On the other hand, it also supports the planning and build the infrastructure.
Comments
Post a Comment