October 21, 2012 at 22:34How we collect data for Nokia Maps
Maps and navigation have not recently disappeared from the focus of attention of the IT community. Since the web became social, navigation has proven to be one of its pillars, on which numerous services and applications rely. Today we will tell a little about how we collect data for our maps - maps.nokia.com .
Cartography, a science known to everyone from children's atlases, whose age is more than 5 thousand years, in conjunction with computers, turns out to be the most advanced frontier of combined realities, turning our world into databases suitable for processing. The finest part in creating maps is that indexing the physical world still requires people to drive around the cars and process the collected material at the computer.
Different cartographic companies have different approaches to creating their brainchild. For example, while competitors from the search business drive 5 million miles on their cars, Nokia receives a large amount of map data from various commercial carrier companies such as FedEx. Carriers drive about 3.3 billion miles per year. We receive from our partners about 12 billion data packets per month. In addition, when compiling maps, data from user navigation devices are used.
Also, these data allow you to track changes in existing roads and the emergence of new ones. In 2012, GPS data was used to refine 65,000 road segments. GPS data is also used to determine traffic congestion, as the speed of vehicles is also known.
It is worth noting here that due to concern for the rights of the user, information tracking stops every 30 seconds, so we do not fix specific routes that a certain person is moving.
In the future, we at Nokia expect to receive up to 30 different parameters from a GPS request, which will allow us to keep maps up to date without having to send employees to travel on roads. In the meantime, this system is, unfortunately, auxiliary.
In addition to FedEx and user devices, data-collection is carried out by driver researchers operating Nokia's True machines. Such work requires a special character warehouse, as it requires constant movement, mapping another major city. Months without loved ones, spending the night in one or another hotel are normal conditions for our researchers. The specifics of the cameras installed on the machine are such that they work best in dry conditions, so this work has a pronounced seasonal character. In the north, drivers work in the summer, and with the onset of winter they move south.
Inside a specific city, the route is planned according to an algorithm that determines the most effective way to drive and fix each road on a map. The algorithm, of course, is quite flexible and can make mistakes, but the system makes a certain route and tries to follow it. Drivers drive 8-9 hours a day, after which they return to the hotel. The next day they return to the point at which they stopped yesterday and continue to follow the given route.
A large amount of equipment is installed on the True Roof, one on top of the other. When fully deployed, the height of the module is about 2 meters.
1 - Six cameras that remove road signs. 2 - LIDAR. 3 - Panoramic camera.
The car is equipped with equipment with a total value of about $ 200,000. This includes 6 cameras that remove street signs, a panoramic camera for creating the Bing Street View gallery (recall, Microsoft uses our maps), two GPS antennas, three laptops and the main jewel - the LIDAR system, which “shoots” 64 lasers all 360 degrees around the car to create a three-dimensional terrain model on which the car is traveling.
LIDAR is used for various purposes, especially when combined with other visual data. He makes 1.3 million measurements per second, which are combined into a single street picture. This data is then superimposed on a picture taken by other cameras on the roof to create a digital cast of the area.
After we got a digital version of the environment, we can get almost any data from the terrain images. With their help, you can easily calculate, for example, the height and width of the bridge, as if "reading" the physical world around. And you can decode various signs and pointers to understand the location of the transport network around. We can extract and recognize 100 different types of signs and signs in 13 different countries of the world automatically.
The biggest problem of creating maps, because of which we have to actively use our own True vehicles, is not in three dimensions, but in the fourth - the world changes over time. Creating a map is a relatively simple task compared to its further support in the current state.
We collect up to 400 pieces of information on each segment of the road. This can be information about signs where the road leads, about the number of lanes or about repair work. This is a huge amount of information. When we talk about a percentage change, it can be a speed limit or a name on a certain track. This is an ongoing process of understanding the dynamic nature of changes in these road networks.
We have a concept called The Living Map. The idea is that as soon as a person starts using the map and additional location services, the map starts to find out what the person is looking for. And over time, you begin to connect the places you love, gradually turning this network into a new layer on top of all the road data carefully collected for you by the Nokia team.
Cartography, a science known to everyone from children's atlases, whose age is more than 5 thousand years, in conjunction with computers, turns out to be the most advanced frontier of combined realities, turning our world into databases suitable for processing. The finest part in creating maps is that indexing the physical world still requires people to drive around the cars and process the collected material at the computer.
Different cartographic companies have different approaches to creating their brainchild. For example, while competitors from the search business drive 5 million miles on their cars, Nokia receives a large amount of map data from various commercial carrier companies such as FedEx. Carriers drive about 3.3 billion miles per year. We receive from our partners about 12 billion data packets per month. In addition, when compiling maps, data from user navigation devices are used.
Also, these data allow you to track changes in existing roads and the emergence of new ones. In 2012, GPS data was used to refine 65,000 road segments. GPS data is also used to determine traffic congestion, as the speed of vehicles is also known.
It is worth noting here that due to concern for the rights of the user, information tracking stops every 30 seconds, so we do not fix specific routes that a certain person is moving.
In the future, we at Nokia expect to receive up to 30 different parameters from a GPS request, which will allow us to keep maps up to date without having to send employees to travel on roads. In the meantime, this system is, unfortunately, auxiliary.
In addition to FedEx and user devices, data-collection is carried out by driver researchers operating Nokia's True machines. Such work requires a special character warehouse, as it requires constant movement, mapping another major city. Months without loved ones, spending the night in one or another hotel are normal conditions for our researchers. The specifics of the cameras installed on the machine are such that they work best in dry conditions, so this work has a pronounced seasonal character. In the north, drivers work in the summer, and with the onset of winter they move south.
Inside a specific city, the route is planned according to an algorithm that determines the most effective way to drive and fix each road on a map. The algorithm, of course, is quite flexible and can make mistakes, but the system makes a certain route and tries to follow it. Drivers drive 8-9 hours a day, after which they return to the hotel. The next day they return to the point at which they stopped yesterday and continue to follow the given route.
A large amount of equipment is installed on the True Roof, one on top of the other. When fully deployed, the height of the module is about 2 meters.
1 - Six cameras that remove road signs. 2 - LIDAR. 3 - Panoramic camera.
The car is equipped with equipment with a total value of about $ 200,000. This includes 6 cameras that remove street signs, a panoramic camera for creating the Bing Street View gallery (recall, Microsoft uses our maps), two GPS antennas, three laptops and the main jewel - the LIDAR system, which “shoots” 64 lasers all 360 degrees around the car to create a three-dimensional terrain model on which the car is traveling.
LIDAR is used for various purposes, especially when combined with other visual data. He makes 1.3 million measurements per second, which are combined into a single street picture. This data is then superimposed on a picture taken by other cameras on the roof to create a digital cast of the area.
After we got a digital version of the environment, we can get almost any data from the terrain images. With their help, you can easily calculate, for example, the height and width of the bridge, as if "reading" the physical world around. And you can decode various signs and pointers to understand the location of the transport network around. We can extract and recognize 100 different types of signs and signs in 13 different countries of the world automatically.
The biggest problem of creating maps, because of which we have to actively use our own True vehicles, is not in three dimensions, but in the fourth - the world changes over time. Creating a map is a relatively simple task compared to its further support in the current state.
We collect up to 400 pieces of information on each segment of the road. This can be information about signs where the road leads, about the number of lanes or about repair work. This is a huge amount of information. When we talk about a percentage change, it can be a speed limit or a name on a certain track. This is an ongoing process of understanding the dynamic nature of changes in these road networks.
We have a concept called The Living Map. The idea is that as soon as a person starts using the map and additional location services, the map starts to find out what the person is looking for. And over time, you begin to connect the places you love, gradually turning this network into a new layer on top of all the road data carefully collected for you by the Nokia team.