Following on from the launch of Ubotica’s CogniSAT-6 in March, the European Space Agency (ESA) is gearing up to take the next step in revolutionizing Earth Observation (EO) with the launch this June of Φsat-2, an AI-powered CubeSat. Powered by Ubotica’s SPACE:AI, Φsat-2 will break free from traditional limitations and unlock a new era of real-time insights from space by demonstrating the ability to deploy different AI applications from many vendors, this deployment being achieved while in orbit.
Lead by Open Cosmos, the mission partners CGI, Simera, DEIMOS, KPLabs, GEO-K and Ceiia have collaborated to deliver a satellite which showcases SPACE:AI’s ability to execute AI applications on demand.
The Problem: Wasted Time, Data, and Costs
Today’s EO methods involve capturing vast amounts of data in space (“mow the lawn”), downlinking it to Earth, and then processing it to extract useful imagery – a time-consuming, data-intensive, and costly process. These inefficiencies become even more problematic with higher-resolution cameras.
Φsat-2: AI-Driven Innovation in Space
Φsat-2 breaks this paradigm by deploying AI directly onboard the satellite. With a multi-spectral camera and SPACE:AI, it analyzes imagery in real time, extracting key insights and transmitting only the most valuable data. This slashes wasted time, data, and costs.
App Store infrastructure comes to Space
Think of Φsat-2 as an orbiting smartphone. ESA’s Φsat-2 satellite, incorporating Ubotica’s SPACE:AI technology, enables custom AI apps to be developed, installed, and operated on the satellite, even while in orbit. This flexibility means Φsat-2 can adapt to changing needs, maximizing its value for scientists, businesses, and governments.
Initial Φsat-2 AI Apps:
The mission launches with exciting apps pre-installed:
- Cloud Removal: Eliminates cloud-covered imagery, reducing clutter.
- Streetmaps from Satellites: Generates emergency response maps on-the-fly.
- Ship Detection: Identifies and classifies vessels in near-real time.
- Compression App: Demonstrates on-board AI reducing data storage and transmission needs without significant information loss. It uses a Deep Autoencoder for efficient image compression, allowing direct ground reconstruction for tasks like object detection.
Orbital AI Challenge Winners
ESA’s 2023 Orbital AI Challenge yielded two more transformative apps:
- The Saint Exupéry’s Marine Environment Monitoring: Detects oil spills, algae blooms, and sediment in real-time.
- PhiFire AI’s Wildfire Detection: Provides critical real-time information for firefighters.
Ubotica’s SPACE:AI platform is the key to Φsat-2’s intelligence:
- SPACE:AI Processing Layer: CogniSAT Edge-AI for onboard analysis.
- App SDK Layer: Enables developers to build custom apps.
- Applications Layer: Diverse, customizable applications, which can run concurrently and apply multiple data processing passes on the same imagery.
The Future of Earth Observation Starts Now
Ubotica is pioneering Live Earth Intelligence, having launched CogniSAT-6 in March this year. CogniSAT-6 uses AI to deliver information from captured images straight to end-user devices such as your phone within minutes. It makes autonomous decisions based on what it observes, and allows users to interact in real-time with the spacecraft. It is the first satellite that will provide real Live Earth Intelligence. Φsat-2 is the latest example of a smart, connected and collaborative Earth Observation system enabled by Ubotica technology. Applications will be cross compatible between these spacecraft – powered by Ubotica SPACE:AI.
Get expert insights on Smart Satellites, EO, Live Earth Intelligence, and SPACE:AI delivered straight to your inbox. Sign up here!
Recent Comments