For the first time, an Earth observation satellite has autonomously identified objects of interest using a vision-language model (VLM) in orbit. This milestone, achieved in April using Google DeepMind's Gemma 3 VLM onboard Loft Orbital's Yam-9 spacecraft, demonstrates AI's potential to revolutionize space-based sensing. The technology allows for on-orbit data triage, reducing the need for extensive ground analysis, and opens possibilities for continuous, interactive monitoring from space. AI
IMPACT Enables autonomous on-orbit data analysis, potentially reducing data transmission needs and enabling new space-based AI applications.
RANK_REASON Demonstration of a vision-language model running on an Earth observation satellite for autonomous object identification.
- Mastodon
- mastodon.social
- TechCrunch
- Gemma 3
- Google DeepMind
- Jet Propulsion Laboratory
- Jetson Orrin AGX GPU
- Juan Delfa Victoria
- Kepler Communications
- Loft Orbital
- NASA
- NAVI-Orbital
- Nvidia
- Paul Lasserre
- Planet Labs
- Taran Cyriac John
- Yam-9
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
