Google Maps. Apple Maps. OpenStreetMap. There is no shortage of free web maps these days.
These may be practical tools for a lot of use cases, but if your needs go beyond figuring out how to get from A to B the free web maps soon reveal their shortcomings. The most obvious limitation: their map data tends to be out of date, sometimes by as much as several years.
This is the need identified by Terramonitor. “There is a lot of geographical information available, but since it’s often outdated it doesn’t reflect the reality anymore,” says Joni Norppa, CEO of Terramonitor.
The company’s solution provides an up-to-date map of the globe by combining open-source satellite imagery with other data sources, such as aeroplane footage and field measurements. The bulk of the data comes from the European Space Agency, whose earth observation satellites continuously produce new geographical data.
The firm also makes use of a state-of-the-art algorithm that allows it get more out of the geographical data it collects. Finally, Terramonitor packages the map data in a format that is easy to integrate into clients’ existing workflows.
“Even though geographical data could be hugely valuable to a lot of people, many are discouraged by how difficult using it can be,” Norppa says. “Space data is hard, but we make it easy.”
Eye in the sky
For a Finnish company, it is perhaps not surprising that the forest industry is an important target market. Up-to-date geographical data helps forest owners determine the size and type of their forests, thus making forest inventory easier.
Continuously updated data can also be used to detect deforestation and illegal logging. “In South America, for example, farmers sometimes fell woods without permission from the forest owner in order to appropriate the land for farming,” Norppa says.
“Previously it could have taken months, even years, to discover illegal tree felling. But thanks to satellite imagery and machine learning it can now be detected almost immediately,” he says. “Even if you hired 100 people just to focus on illegal logging in Brazil, for example, they couldn’t do it as well as the AI can.”
Another industry that benefits from Terramonitor’s service is agriculture: since the company’s algorithm can analyse optical images for information that the human eye cannot see, it can be used to monitor the condition of the crop fields and provide actionable information to farmers.
Space data with meaning
Machine learning also gives Terramonitor a unique advantage in dealing with what Norppa calls the “cloud challenge” – the fact that satellite images are often obscured by clouds.
Most solutions to this compare two images in order to detect cloudy and non-cloudy areas. Instead, Terramonitor’s technology makes use of a large number of images, and uses cross-checking and machine learning to detect the non-cloudy ones.
In fact, machine learning and satellite imagery is a powerful combination that is unlocking entirely novel use cases, Norppa says. The company is looking at infrastructure and consumer behaviour specifically: analysing traffic patterns around shopping centres, for example, is likely to produce information that is highly valuable to businesses and marketers.
Space has quickly become a burgeoning industry, but Norppa says that the hype around the field misses some important elements. “A lot of the focus is on launching satellites and producing data, but we think it’s also crucial to produce information based on this data,” he says. “We make the data valuable by giving it meaning.”