Home
Back to portfolio

GeoRefMaps

Functional MVP for semi-automatic georeferencing of geological maps using coordinate detection, OCR-assisted control points, manual correction, and GeoTIFF export.

Overview

GeoRefMaps is an MVP tool for the semi-automatic georeferencing of geological maps distributed as PDF documents and raster images.

It was designed to detect map-border coordinates, identify the main cartographic frame, and generate geographic control points (GCPs) with visual assistance and manual correction support.

Current Capabilities

The project already integrates:

  • A functional Python graphical interface
  • Vector PDF reading
  • Raster image processing
  • OCR fallback with Tesseract
  • Automatic corner detection
  • Recognition of projected UTM coordinates
  • Experimental support for geographic coordinates

The current workflow is focused on geological cartography from Peru.

Workflow

GeoRefMaps currently supports maps in PDF, PNG, JPG, and TIFF formats.

The workflow can:

  • Extract vector text directly from PDFs
  • Apply OCR on borders and selected areas of interest
  • Estimate the main cartographic frame and grid lines
  • Assign coordinates to the four map corners
  • Allow manual validation and correction from the interface

Outputs

The application can export:

  • The full map sheet as a georeferenced GeoTIFF

Export and reprojection are already supported for:

  • EPSG:4326
  • EPSG:32717
  • EPSG:32718
  • EPSG:32719

Repository

Release

Applied Value

At its current stage, GeoRefMaps demonstrates a solid technical foundation for reducing manual georeferencing time and supporting geological, mining, and cartographic workflows.

Special thanks to Fabian Yacila for his conceptual contributions.