SAN FRANCISCO – SpaceChain, a Singapore startup with blockchain nodes in orbit, is rolling out a service that integrates artificial intelligence with Earth imagery.
I-Sat, from SpaceChain’s U.S.-based SC Solutions, is designed to simplify the process of answering questions with Earth observation data.
“AI is convenient. You type in your question, and you get the answer,” Ziheng Xiang, SpaceChain director of technology, told SpaceNews. “But most of the AI is not supported by real-time analysis and there’s nothing combining the AI component with the Earth intelligence.”
Like OpenAI’s ChatGPT, I-Sat “generates an answer based on the natural language processing technology,” Xiang said. “But it also provides answers supported by real-time data analytics.”
SC Solutions is inviting Earth-imagery providers and applications developers to join its platform. SC Solutions will parcel out payment to vendors with its blockchain technology, said SpaceChain CEO Cliff Beek.
Sugarcane and Paper
To demonstrate the merit of I-Sat, SC Solutions conducted pilot projects focused on paper, pulp and sugarcane production in Brazil. By applying AI to satellite imagery, I-Sat was able to provide customers with detailed information on vegetation health and soil moisture as well as recommendations for improving plant health.
For the sugarcane plants, “we used optical imagery, synthetic aperture radar imagery and some open-source climate data,” Xiang said. “All of these packed into a machine-learning model provided high-precision prediction of the yield.”
Mining is another promising application, Xiang said.
Technology Partners
Before pivoting toward Earth observation and AI, SpaceChain established blockchain payloads in space. Seven SpaceChain nodes on satellites and the International Space Station process, transmit and store data.
To establish I-Sat, SC Solutions joined Nvidia’s Inception Program and Google for Startups, free resources for fledgling enterprises. SC Solutions also is forging ties with satellite imagery providers.
“It’s very difficult for the average user to access Earth observation,” Beek said. “Our platform will add a layer of analytics to bring together data from multiple providers.”
The company’s generative AI tool “allow us to interact with this chatbot so that the platform can provide the user with the accurate analysis and the explanation,” Xiang said.
If a user queried the platform for information on water pollution in New York in 2022, for example, the language model would “analyze my question and the user with accurate information and the explanation,” Xiang said.
