Satellites Learn to Describe Their Own Imagery in Real Time
TL;DR
NASA/JPL and Loft Orbital's NAVI-Orbital enables lightweight on-board AI (DeepMind Gemma 3) to identify and describe features in satellite images via prompts, reducing ground-downlink and enabling real-time, autonomous analysis; initial tests show 88.2% accuracy across 7,960 images, with only two in-orbit captures so far; the approach could scale to a global, continuous monitoring network but raises ethics and robustness questions about on-board interpretation and surveillance.
Topics:sciencetechnology#artificial-intelligence#navi-orbital#onboard-processing#real-time-monitoring#satellites#technology
- In a First For Science, A Satellite Has Identified What It's Seeing From Space ScienceAlert
- Loft Orbital to test AI models on spacecraft for Earth observation SpaceNews
- Onboard Vision-Language Models Trend Hunter
- Loft Orbital Selected by NASA's Jet Propulsion Laboratory to Deploy Artificial Intelligence Software for Earth Science Applications Business Wire
- Loft Orbital – Weekly Recap TipRanks
Want the full story? Read the original reporting
Read on ScienceAlert