2026-05-06 23:37:01 +03:00
2026-04-14 15:57:42 -07:00
2026-05-03 16:19:43 +03:00
2026-05-03 16:19:43 +03:00
2026-05-03 16:19:43 +03:00
2026-05-03 03:32:46 -07:00
2026-05-03 16:47:02 +03:00
v4
2026-04-14 14:48:54 -07:00
2026-05-03 15:09:17 +00:00
2026-05-03 16:19:43 +03:00

This is a project that consists of both extracting and digitalizing/modernzing previously existing vector garmin img maps.

  • BgMountains project
  • Should be applcable to other such img maps
  • I've made sure to make the dynamic type recognition and configuration a thing, just for those types of scenarios

Thus, since, it's been 17 release cycles, each adding more digital granularity available from the previous maps into a digitally convertible format, and it since it has been 17 such cylces, we've come full circle and exhausted bgMountainsMap and must circle back to the original BGTopo Map Project of the Plovdiv Universtiy.

Agenda

  1. Reverse the garmin img format [x] Extract - coordinate + unique type pairs [x] Parse them [x] Output them into different formats:
  • OSM/OBF(with converter)
  • CSV
  • KML
  • KMZ
  • GPX
  1. Cover bulgaria's watersources fully
  • A [x] 75% of Bulgaria is covered (the north plateu of Bulgaria seems to not have corresponding BGMountains locators for wetersources)
  • B [ ] Parse the existing BGTopo Plovdiv Uni Project and it's military map with OCR to determine more such points.

To achieve B, the current plan is to:

  • Export and centre the whole military map
  • Train a simple ocr model on the existing pairs
  • Start cross verifying across the extracted digital coords and the detected ones by the model
  • Name them

No Water Points

Uses:

  • OsmAnd~ - OBF / GPX
  • Google Earth Pro - KML (tested) / KMZ
Description
My journey on extracting data from garmin img non-documented format.
Readme 172 MiB
routes-v3 Latest
2026-06-04 10:44:04 +00:00
Languages
Python 74%
Jupyter Notebook 25.9%