adroitly.

Introduction

This article presents a Model Context Protocol (MCP) server for high-revisit satellite constellation image tasking. The MCP server allows for ease of connection to most Large Language Models (LLMs), offering a flexible and agentic image tasking workflow with a natural language chat interface.

Oftentimes, the number of requested images are much more than the constellation of imaging sensors can service. Assigning a subset of imaging tasks (i.e., “task schedule”) to each satellite sensor is achieved by optimising for revisit frequency, ensuring cooperation between all satellites in the constellation, and taking into account a comprehensive set of hardware constraints that guarantees a realistic solution.