Users can track economic activity and detect changes in the areas that are of most interest to them.
A new tool from French aerospace company Airbus and Silicon Valley startup Orbital Insight uses machine learning running on diverse geospatial datasets to let customers detect changes in infrastructure and land use in near real-time.
The tool, dubbed Earth Monitor, includes analytics that automatically detect and aggregate objects such as cars, roads, buildings and construction sites. It is being rolled into Airbus’s OneAtlas suite of geospatial imagery tools.
Earth Monitor use Orbital Insight’s computer vision and machine learning systems to detect changes in infrastructure by scanning through its petabytes of multi-source data that has been obtained via satellite and synthetic aperture radar (SAR) imagery, geolocation intelligence, and vessel traffic (AIS) data.
Dr. James Crawford, CEO of Orbital Insight said: “We’re proud and excited that Earth Monitor will leverage our dynamic algorithms so that users can track economic activity and detect changes in the areas that are of most interest to them, enabling custom analytic projects where they can define the what, where and when.”
In a promotional video the companies pointed to precision agriculture, deforestation monitoring and industrial uses, as well as the ability to build 3D models. For extractive industry users, the ability to assess and track detailed geospatial data like this could be crucial to informing investment decisions in high risk frontier markets.
Its launch comes as geospatial data is increasingly being used by a wide range of businesses and third sector organisations. At a recent AI DevCon in Munich, one United Nations representive detailed how AI trained on geospatial data is being used to automatically assess the size of Syrian refugee camps.
Orbital Insight’s existing tools include an energy analysis intelligence subscription that offers customers access to daily global and regional crude oil volume estimates, monitoring over 5 billion barrels of oil storage capacity across more than 25,000 tanks.
It has also used a convolutional neural network to identify industrial oil palm plantations for the NGO Global Forest Watchm training it on high-resolution satellite imagery from the imaging company, Planet.
As the company noted in a blog late last year: “We’ve made significant progress over the last few years, but we still haven’t completely accomplished our goal.”
“The current datasets pick up large-scale row plantations reasonably well, but we’re still seeing some misclassification between oil palm and other, similar-looking plantations like bananas. The algorithm also can’t identify plantations until they are mature enough to be visible in satellite imagery, which means we may not be able to attribute deforestation to oil palm until several years after it occurs.”