The Spectra AI-based system will now analyze satellite images.
The American company BlackSky has entered the final phase of the multi-year SMART program, which is implemented by the Agency for Advanced Research in the Field of Intelligence. The goal of the project is to create AI tools for analyzing satellite images in the interests of national security.
During several stages of the SMART program, BlackSky has been actively working on new machine learning architectures. As a result, the Spectra platform was created, which is already capable of processing data from tens of millions of square kilometers.
Using artificial intelligence methods, Spectra analyzed archived images over 5 years and identified more than 2 million changes in 120,000 different locations around the globe. This is three times higher than the originally planned system performance.
"For the SMART program, BlackSky developed a high — performance machine learning architecture (MLOps), originally designed to analyze one million square kilometers," explained Patrick O'Neill, Chief Technology Officer at BlackSky. "This technology now brings significant benefits to other information retrieval and analysis programs, processing more than 30 million square kilometers per month."
The technology has already attracted the attention of US intelligence agencies and will soon be used in many other government projects.
Large-scale search is a key element of BlackSky's SMART mission, which is of great importance for intelligence organizations. The technology will allow you to get a general picture of the total changes in large areas of the Earth's surface over time. Such an approach, encompassing both environmental change and security-related activities, is critical for monitoring and understanding global processes.
The American company BlackSky has entered the final phase of the multi-year SMART program, which is implemented by the Agency for Advanced Research in the Field of Intelligence. The goal of the project is to create AI tools for analyzing satellite images in the interests of national security.
During several stages of the SMART program, BlackSky has been actively working on new machine learning architectures. As a result, the Spectra platform was created, which is already capable of processing data from tens of millions of square kilometers.
Using artificial intelligence methods, Spectra analyzed archived images over 5 years and identified more than 2 million changes in 120,000 different locations around the globe. This is three times higher than the originally planned system performance.
"For the SMART program, BlackSky developed a high — performance machine learning architecture (MLOps), originally designed to analyze one million square kilometers," explained Patrick O'Neill, Chief Technology Officer at BlackSky. "This technology now brings significant benefits to other information retrieval and analysis programs, processing more than 30 million square kilometers per month."
The technology has already attracted the attention of US intelligence agencies and will soon be used in many other government projects.
Large-scale search is a key element of BlackSky's SMART mission, which is of great importance for intelligence organizations. The technology will allow you to get a general picture of the total changes in large areas of the Earth's surface over time. Such an approach, encompassing both environmental change and security-related activities, is critical for monitoring and understanding global processes.
