地球觀測中的人工智慧/機器學習及賦能技術

"地球觀測服務供應商如果想在不斷發展的地球觀測市場中保持競爭力，就必須採用能夠進行即時資料分析的智慧、分散式自適應系統。"

本報告為在地球觀測 (EO) 中使用人工智慧和機器學習 (ML) 提供了策略指導。它還解釋瞭如何將人工智慧/機器學習與基礎模型等新興賦能技術相結合，為最終用戶提供客製化的解決方案。

報告中解答的問題

• 地球觀測衛星營運商和服務提供者如何利用代理人工智慧、聯邦學習和基礎模型來提供高效能的客製化解決方案？
• 各利害關係人該如何因應人工智慧日益普及並從中受益？
• 在太空和地面採用邊緣運算的策略有哪些？
• 利害關係人可以利用哪些合作關係來增強和提升其人工智慧/機器學習能力？
• 在建構支援 AI 的 EO 解決方案時應考慮哪些因素？

"Earth observation service providers must embrace intelligent, decentralised and adaptive systems that enable real-time data analytics if they wish to stay competitive in the evolving Earth observation market."

This report provides strategic guidance about using AI and machine learning (ML) for Earth observation (EO). It also describes how AI/ML can be used together with emerging enabling technologies such as foundation models to offer tailored solutions for end users. It outlines implementation strategies for various stakeholder groups, and lists the benefits of, and requirements for, fulfilling customers' needs.

Vendors can also use the recommendations to further strengthen their value propositions (particularly for downstream applications) and build solutions that address market needs.

Questions answered in this report:

• How can EO satellite operators and service providers use agentic AI, federated learning and foundation models to offer high-performance, tailored solutions?
• What should various stakeholders do to address, and benefit from, the increasing adoption of AI?
• What are the adoption strategies for edge computing in space and on the ground?
• Which partnerships will enable stakeholders to enhance and improve their AI/ML capabilities?
• What are the key considerations when building AI-enabled EO solutions?