Blacksky Reaches Final Phase of AI Satellite Image Analysis Program
American company Blacksky has successfully advanced to the final phase of the long-term program called Smart, initiated by the agency of advanced research in the field of intelligence. The primary objective of this project is to develop artificial intelligence (AI) tools for analyzing satellite images, with a focus on national security.
Throughout the various stages of the Smart program, Blacksky has been actively engaged in the development of new machine learning architectures. As a result, they have created the Spectra platform, which is already capable of efficiently processing data from tens of millions of square kilometers.
Using state-of-the-art artificial intelligence techniques, Spectra has analyzed archival satellite images spanning a period of 5 years, successfully identifying over 2 million changes at 120,000 different locations worldwide. This achievement exceeds the initial performance expectations of the system, which demonstrates its remarkable capabilities.
“For the Smart program, Blacksky has developed a high-performance machine learning architecture (Mlops), originally designed for analyzing one million square kilometers,” explained Patrick O’NNIN, chief technologist of Blacksky. “Now, this technology is applicable to other programs involving information search and analysis, with a monthly processing capability of over 30 million square kilometers.”
The impressive technology has already captured the attention of American intelligence agencies and is poised to be utilized in numerous other government projects in the near future.
A large-scale search is a crucial component of Blacksky’s mission within the Smart program, which holds significant value for intelligence organizations. This technology enables the obtainment of a comprehensive understanding of changes occurring across vast areas of Earth’s surface over time. Such an approach