Scientists have digitized 94 thousand old images of the starry sky

Consortium of European Scientists completed APPLAUSE project. Within its framework, 94 thousand photographic plates of 1893-1998, which were stored in the archives of European institutions, were digitized. All images were analyzed and cataloged using machine learning and other technologies.

This project is important because it gives everyone interested open access to unique data. Each shot captured the celestial bodies at a particular moment in time, and the exact position will never be repeated. Perhaps these photos will help scientists make discoveries that could not be confirmed by observations in the 21st century.

In addition to photographs of space objects, scans of observation logs, envelopes in which records are stored, and other documents are presented.

When digitizing such materials, calibration accuracy is important. To provide it, we developed the open source program PyPlate. It uses machine learning techniques to detect defects. With its help, you can detect scratches and dust on the plates and remove them so that the final image is better and more accurate.

You can view the images at project page. You can configure the search for a specific archive, select a specific year, month, and even date. If you were born before 1998, you can find out what scientists photographed on your birthday.