Historical map images georeferencing

There was a need to georeference a large number of raster images of historical maps. The total number of images was about ten thousand.

Existing georeferencing solutions, such as QGIS, were not suitable because they were not effective in massive stream processing of a large number of images. It was decided to develop a specialized desktop-application taking into account the existing requirements.

The developed application is based on the functionality of the GDAL library. Image transformation parameters obtained as a result can be used for mass parallel transformation of a large number of images.

This was an internal project of GeoICON.

The job was done within four months during the year 2015.

All the necessary volume of images was successfully processed and transferred to the customer. At the moment, the application is being refactored in order to transform it into a commercial product.

Here is what I was doing in the project:

  • Analysis and requirement study;
  • Leading the programming team and core development;
  • Development of the database for the PostgreSQL DBMS with PostGIS extension;
  • Development of the backend part of the solution;
    • Development of the REST API with Node.js and Express;
    • Development of Telegram Bot for remote control of the georeferencing process;
  • Development of the cross-platform desktop application with the Qt framework;
    • Intensive usage of the  GDAL library;
    • Integration of web components inside the desktop application with the QtWebkit. Intensive usage of OpenLayers;
  • Deployment of the solution on the server running CentOS 6 with Docker;

Below are a few illustrations to give you a visual impression of the web application.