Cross-Correlation Methods in Problem of Search and Identification of Transient Objects in Astronomical Images

Pattern Recognition and Image Analysis. 2023. Vol. 33. No. 2. P. 157–168.

E. A. Shekotihin, Pankov N.S., Pozanenko A.S., Belkin S.O.

The paper considers the effectiveness of various methods of cross-correlation of astronomical
catalogs in solving the problem of search and identification of the sources of rapidly varying brightness, i.e.,
of the transients. Comparison of the methods is carried out in terms of precision and recall of cross-correlation
and is considered by an example of circular fields of the optical catalogs Pan-STARRS PS1 and
SDSS DR12 with the radii of 60 and 61 arcmin, respectively. In addition, the possibility of using the photometric
information in cross-correlation is investigated, and the description of the developed algorithm
for searching for transient events is provided. The list of candidate transients detected in the mentioned
field is also included in the paper.

Keywords: machine learning, cross-correlation, Transients, Binary classification, brightness matching, cross-matching