Toponym Extractor

Turns MapReader text detections in pixel coordinates into a latitude/longitude table, using the ground control points in an Allmaps Georeference Annotation. Everything runs in this browser tab — your files are never uploaded anywhere.

1 Inputs

MapReader detections

GeoJSON in pixel coordinates. Drop a file here, choose one, or paste below.

No detections loaded.

Measured on a different image?

Pixel coordinates only mean something against the image they were measured on. If MapReader ran on a different scan or size than the one the annotation georeferences, everything lands in the wrong place. Drop that image here — only its header is read, so a 400 MB TIFF is instant — or type its size.

× px

Assuming the same image as the annotation.

Allmaps Georeference Annotation

The .json saved from Allmaps Editor, or its annotation URL.

No annotation loaded.

Every coordinate below inherits the accuracy of this georeferencing. Loosely placed or sparse ground control points produce confidently wrong lat/long with nothing on screen to indicate it — check a few known locations against a basemap before trusting the output.

2 Options

The mask is the polygon drawn in Allmaps Editor around the cartographic area of the sheet. Text in the margins — titles, legends, imprints, credits — falls outside it.

Detections scoring below this are left out of the results.

Substring match anywhere in the transcribed text, ignoring case and accents — segou finds both SEGOU and Ségouba, and bou finds Diambour.

Allmaps ground control points use image pixels with Y growing downward from the top-left corner. MapReader output that has passed through a GIS often has negative or upward Y instead, which has to be corrected or the result comes out mirrored.

Leave this alone to use exactly what you chose in Allmaps Editor.

Load both inputs to continue.