Image acquisition and processing using a PC cluster toward real-time solar image improvement

This paper describes a PC-cluster-based observation system for improving the angular resolution of solar images. The system consists of two equipments separately working: one equipment realizes automatic acquisition of one hundred images at every fifteen seconds, and selects ten best images to be used in image processing. In the other one, image restoration and resolution improvement are carried out with a blind deconvolution method and a super-resolution method, respectively, using a PC cluster system. The system can take images at every fifteen seconds with high angular resolution. A test observation at the Hida Observatory confirmed a satisfactory performance of this system. The two equipments in the current hybrid system will be united in a future system that will realize image processing in pseudo-real time. A condition toward pseudo-real time processing is also investigated in this paper.