Data driven approach to binarization of astronomical images
Dimo Dimov, Aleksandar Dimov · 2010
The paper considers the problem of astronomical images binarization. The problem arises with the idea for effective application of Hough transform to the identification of stars in chain plate images. The astronomical methods based on chain plates are rather old, but the plates processing is an important task because of the large number of archives (and therefore priceless from an astronomical viewpoint) of chain plates in Bulgaria, Europe and across the world. After a brief analysis of the state-of-the art, the paper proposes a locally adaptive binary-tree method for binarization of wide photo plate images, as a modification and extension of a similar one that is the authors' former development. Theoretical and experimental approaches of the proposal are also discussed in the paper.