High-Performance Automated Image Tracking for Astronomical Observations

Many astronomical observations require extended image exposures in order to improve signal to noise ratio and to provide adequate data for quantitative analysis of image contents. Maintaining accurate pointing of a telescope over the desired observation time is challenging. We consider the problem of correcting for changes in the field of view by dividing the overall observation into a series of image frames, analyzing the individual frames to estimate the changes in the field of view between frames, and performing a corrected accumulation of image frames to form the final observation. We present an automated image tracking algorithm tailored for execution on resource-constrained platforms. We show the results of extensive algorithm performance characterization studies. Finally we describe an implementation of the algorithm for a radiation-tolerant FPGA.