Institute of Statistical Mathematics

Miscellaneous Information

Miscellaneous Information

Abstract Reference: 31312
Identifier: P1.19
Presentation: Poster presentation
Key Theme: 1 Reduction and Analysis Algorithms for Large Databases and Vice-versa

Data Compression for optical movie data of the Tomo-e Gozen

Morii Mikio, Ikeda Shiro, Ohsawa Ryou, Sako Shigeyuki

The Tomo-e Gozen is a wide-field camera on 1-m Schmidt telescope on Kiso Observatory at Japan. By using the camera, we plan to monitor the wide area of the sky with high speed of 2 Hz, to search optical transients. Full observation will start on the next year. The camera consists of 84 CMOS chips and cover the fields of view with 20 deg^2. The data obtained by one night observation will become about 30 TB. Then, data compression is necessary. We have applied a matrix factorization method for the movie data and showed that the method can compress the data into one-tenth of original size, without losing scientifically important information. We demonstrated that the method can process the data with sufficiently high speed.