TL;DR
- Geological maps are difficult VLM inputs beyond sheer scale: a map sheet’s legend, cross-section, correlation diagram, graticule, and projection notes are mutually referential, so a simple question like “what’s the dominant lithology in this quadrant” requires multi-step reasoning.
- Tiling or downsampling addresses the scale problem but worsens the compositional one, since it can separate a polygon from the legend entry it depends on to be interpreted correctly.
- Existing approaches are surveyed: AI4CMA and DIGMAPPER (and Eigenform’s own tools) treat this mainly as a vectorisation problem, while Microsoft’s PEACE takes a multimodal-LLM approach, decomposing map QA into information extraction, domain-knowledge integration, and answer generation, reporting its GeoMap-Agent scoring 0.811 against GPT-4o’s 0.369 on its own 124-map, 3,864-question GeoMap-Bench.
- Eigenform’s objection to PEACE is about packaging rather than accuracy: its repo ships as a fixed end-to-end QA tool and benchmark, not a modular toolkit.
- The post introduces “Stratigraphic Amenity,” which repackages PEACE into a small Python SDK plus a local MCP server exposing map layout and legend detection, geo-referencing, knowledge retrieval, and overlay operations, deliberately stopping short of OCR, VLM inference, or answer construction so users can bring their own agent and harness.
Key Takeaways
- The contribution here is architectural, not a new accuracy result. Decomposing an end-to-end QA system into composable primitives is what lets other builders reuse the underlying map-reasoning work without inheriting a fixed agent design.
- The benchmark numbers cited (0.811 vs 0.369) are PEACE’s own published results, not a claim about Stratigraphic Amenity’s performance. The tool’s value proposition is modularity and reuse, not a demonstrated improvement on that benchmark.
- Stopping deliberately short of OCR, inference, and answer construction reflects a general preference for composable building blocks over monolithic end-to-end systems. “Bring your own agent and harness” is a scope boundary, not an omission.
Geological maps are awkward inputs for VLMs (circa Fable), and the sheer scale and resolution of these maps is only the beginning of the problem.
A typical map sheet is a large raster containing the map, legend, cross section, correlation of map units diagram, graticule, projection notes, and other marginalia. These components are individually readable, but they are mutually referential: a polygon depends on its legend entry, the legend depends on stratigraphic ordering, and geographic interpretations depend on the map frame and CRS.
Tiling or downsampling can fix the scale problem but exacerbates compositional ones. So a geological question like “what is the dominant lithology in this quadrant” can require multiple steps and operations for an AI assistant to cook up an answer, at varying levels of accuracy.
There is already substantial work on geological-map digitalization: AI4CMA and DIGMAPPER treat primarily as a problem of vectorizing these maps. (And us too, we also have our own tools for vectorizing these maps). While PEACE takes a MLLM approach: their GeoMap-Agent decomposes map QA into information extraction, domain-knowledge integration, and answer generation. The also released GeoMap-Bench: 124 maps and 3,864 geologist reviewed QA pairs, upon which their agent scored 0.811 vs GPT-4o’s 0.369.
However, their repo is currently packaged as and end-to-end QA tool and benchmark. We were looking for something more modular.
Hence, Stratigraphic Amenity: PEACE repackaged into a small Python-SDK + local MCP server exposing operations for map layout and legend detections, geo-referencing, knowledge retrieval, and overlays. It deliberately stops short of doing OCR, specifying VLM inference, or answer construction. Bring your own agent and harness.