Planet Labs announced it is making geospatial data available through Amazon SageMaker, a fully managed machine learning (ML) service from Amazon Web Services (AWS).
Now, Planet data can be directly embedded into Amazon SageMaker, allowing data scientists and ML engineers to acquire and analyze global, daily satellite data. With this data, customers can train, test, and deploy ML models all within Amazon SageMaker.
Planet operates the largest constellation of earth observation satellites in the world, with the capacity to provide daily medium- and high-resolution imagery of Earth’s landmass every day.
Planet is using AWS to better serve its customers who can now benefit from the simplicity and speed of Amazon SageMaker’s new geospatial ML capabilities to build, train, and deploy ML models using Planet’s geospatial data at up to 10x the speed.
These enhanced capabilities create new opportunities for Planet customers to accelerate data access within geospatial tools and cloud platforms.
Due to the challenging work required to use geospatial data for ML, access to ML has historically been out of reach for many geospatial data customers. With Amazon SageMaker, customers can pull in their proprietary data sources, such as Planet satellite data, from Amazon Location Service and AWS Data Exchange.
It’s a first-of-its-kind partnership and the only on-demand, high-cadence satellite imagery ML model training, inference, and visualization platform available in the market.
“Planet understands the challenges associated with ingesting and supplying mass volumes of data,” said Planet President Kevin Weil. “Using Amazon SageMaker, we can now offer ML capabilities with deep integrations of multiple datasets, and we look forward to developing and collaborating on complex models with our customers.”
“Building, training, and deploying ML models using geospatial data is a manual process for most organizations. The primary barrier is the lack of purpose-built capabilities for performing ML on these types of massive datasets,” said Kumar Chellapilla, GM AI Services at AWS.
“Amazon SageMaker can help simplify and accelerate geospatial ML from months to minutes by enabling Planet customers to enrich their datasets, train geospatial models, visualize the results, and more. Together with Planet, we are excited to unlock the value of geospatial data for their customers and help them make accurate predictions,” he added.
Purposely built for geospatial data, Amazon SageMaker geospatial ML capabilities offer easy-to-use tools to orchestrate data chunking and pre-processing operations for satellite imagery, creating web-based visualizations, and creating seamless scaling up for large datasets. The pre-built algorithms can reduce development to just a few days and one-click deployment to the cloud, freeing up customers’ time.
