We’re thrilled to announce that Blackshark.ai geospatial analytics and change detection capabilities, called Orca, are accessible via Microsoft Azure now. The Orca services are designed to integrate with several Azure services, especially Azure Synapse for end-to-end data pipeline manipulation and automated updates on geospatial insights.
Blackshark.ai is offering an end-to-end geospatial platform. Essential part of this AI platform are the geospatial analytics services detecting objects and extracting insights about planet infrastructure from satellite imagery. Such pre-trained and ready to use services can now easily be integrated in Azure Synapse analytics pipelines, providing fast and precise insights and vectors about building footprints, buildings heights or vegetation coverage. Details about the Orca integration into Azure Synapse will also be presented at Microsoft’s booth #1270 during the Space Symposium.
The containerized convolutional networks contained in the Orca services on Azure provide very fast, global-scale, and accurate insights based on satellite or aerial imagery data sets. Those input data sets are conveniently provided and licensable via Azure as well – or you can provide your own data if you prefer. Simply register to your Azure account, choose the right kind of geospatial imagery, select what kind of insights you want to get extracted and start the process. Bing! – you will be notified when the results are ready to consume.
The current machine learning algorithms are optimized for ortho satellite imagery with a resolution between 40 and 60cm and the containers are exporting either georeferenced pixel masks or regularized vectors about building footprints, building height or vegetation coverage. Blackshark.ai’s latest algorithms automatically provide very precise 3D polygonised models of buildings derived from high-resolution aerial imagery. These and other 2D and 3D extraction capabilities will be made available as Azure service as well in the future.
With the patented no-code data labelling tool even custom containers, extracting insights about object classes of your choice can be trained fast and easily. Examples for individual machine learning containers customers are typically asking for are specific types of vehicles, critical infrastructure, building types, or biome types.
Whenever fresh input data is available the Orca service can be started and will provide fast and precise insights. This fast, georeferenced object and change detection capabilities enable applications such as efficient 3D mapping services, logistic planning, risk analysis, telecom signal propagation planning or disaster relief planning.