Our RnD team deployed a brand new “B&C” model for recognition of construction sites in urban areas. Soon it will be implemented into Geoalert processing workflow for automatic Urban Mapping and analysis.
So far we would like to share a demonstration app of the new model applied to one city area of Chelyabinsk, Russia. The map highlights the construction sites for each year in a raw and shows the difference between the previous and the current years.
Images taken from satellites cover larger areas that enables to analyse the context. Use the building recognition tool to get the vector mask of constructions and to monitor changes in a surrounding area.
It’s possible to use the data for calculation of construction areas statistics to perform some indicators such as overall development progress. You may figure out some ways to compare this statistics with the official one.
Our further plans include to apply “B&C” model (building & construction) to classify sites by progress stage: so it would be enough to analyse the single image to find out where there is “construction in progress”, “foundation pits” at the initial stage of construction or the “new buildings” ready to welcome the residents.
However the multitemporal layers based on historical satellite data provide the unique capability to calculate the overall indices of construction progress for the whole city or for the specific area of interest.
As Geoalert platform is connected to a global data source of satellite imagery, partnering with MAXAR, it’s possible to keep checking for a new data available by every selected construction site to process it and deliver the updated information to the subscribers.