Satellite imagery of Port Hedland in Australia. Photo: BlackSky

The Intelligence Advanced Research Projects Activity (IARPA) has awarded BlackSky a second award to support a research program to automate analysis of satellite imagery. On Thursday, BlackSky announced the $4.4 million award to support the second phase of the multi-year Space-based Machine Automated Recognition Technique (SMART) program.

This follows after BlackSky received a contract in January 2021 for the first phase of the program. IARPA is within the Office of the Director of National Intelligence. The SMART program is working to automate the analysis of satellite imagery to improve change monitoring. According to IARPA, the program is using heavy construction as the initial use case. 

BlackSky Chief Innovation Officer Patrick O’Neil said BlackSky created a high-performance machine learning operations (MLOps) framework as part of its work on the first contract. BlackSky’s framework was selected as the core infrastructure for the entire SMART program, O’Neil said. 

“The MLOps framework will extend BlackSky’s Spectra AI tasking and analytics capabilities and accelerate the development of SMART technologies that make quick sense of the immense volume and complexity of space-based image data,” O’Neil added.

BlackSky said it has developed both supervised and semi-supervised machine learning algorithms that automatically recognize heavy construction projects such as military bases, dams, and airports.

This is the latest on a string of government contract wins for BlackSky, including the 10-year Electro-Optical Commercial Layer (EOCL) contract from the National Reconnaissance Office (NRO) worth up to $1 billion. The company also recently won an artificial intelligence contract from the Joint Artificial Intelligence Center worth up to $241 million over five years.