Adaptive Document Image Binarization with Application in Processing Astronomical Logbooks

Abstract

Recently, the digitalization of the astronomical scientific heritage has been considered an important task that can facilitate much researches in astronomy. The creation of digital libraries and databases of astronomical photographic plates brings up the problem of digitalization astronomical logbooks, since the data contained in them is crucial for the usage of the plates. An optical character recognition (OCR) system for the handwritten numerical data is needed in order to speed up the process of database creation and extension. In this paper document image binarization is considered since it is a critical stage for the subsequent steps in an OCR software system. A specific method is proposed which outmatches the state-of-the-art techniques in the case of the images of interest.