Anaximander simplifies running geospatial deep learning models everywhere

Anaximander: Interactively Running Geospatial Deep Learning Models on Any Compute Backend

Computer Vision and Pattern Recognition

Summary

It is hard to use different deep learning models on satellite images because each one works in different ways and on different computers. The authors built Anaximander, a free tool that lets people run many models on any computer, from local machines to cloud servers, using one easy interface. It connects to a popular mapping program to handle cutting images into pieces and putting results back together on maps, showing status in real time. They showed it works by comparing three different models on a farming task without writing code. This tool helps experts quickly try out different methods on satellite images without confusing setup.

What this means in practice

  • For remote sensing teams: Test and compare different deep learning models on satellite images directly within GIS software without custom coding.
  • For precision agriculture services: Run multiple crop monitoring models efficiently on various compute setups to improve agricultural field analysis.

Authors

Satej S. Soman, Akram Zaytar, Girmaw A. Tadesse, Gilles Q. Hacheme, Muhammad S. Danish, Inbal Becker-Reshef, Rahul Dodhia, Juan Lavista Ferres

Abstract

Applying deep learning models to satellite imagery from within geographic information systems (GIS) remains high-friction for remote sensing practitioners. Models arrive in incompatible formats and target different compute environments, from local workstations to serverless cloud services. As a result, every evaluation demands custom deployment, tiling, and georeferencing code before a single prediction reaches the analyst's map. This friction discourages systematic comparison in a domain where model choice directly affects operational outcomes such as field delineation, crop monitoring, and disaster response. We present Anaximander, an open-source system that unifies model source and compute location choice behind one interactive interface. The system's backend is an inference server that loads models from multiple commonly-used sources and serves them on any accessible compute backend. The server provides session management and model caching, and streams results back per tile. The backend is paired with a QGIS plugin that drives tiling, result reassembly, georeferencing, and real-time per-tile status visualization. An additional user-interface path injects layer legends as prompts into vision-language models. We demonstrate the system in a code-free side-by-side comparison of three heterogeneous models on an agricultural field delineation task: gpt-image-1 via a cloud API, Segment Anything Model 3 (SAM3) on a remote GPU, and DelineateAnything on a local CPU. The inference backend and protocol are open-source and available at https://github.com/microsoft/nxmndr.