Abstract
The determination of astrometric positions from a mosaic of direct CCD frames with partial overlap taken with a long focus reflector telescope and the block adjustment technique for data reduction are proposed. The influence of errors introduced by: the measurement method itself, instruments, image centering, and reference positions is evaluated by means of numerical simulations. Finally the achievable accuracy in a 25 arcmin square field covered with sixteen frames is determined. It is found that errors introduced by the method itself and the reference catalog lead to positions with accuracy the same order than the reference catalog. Since errors are not significant if the ACT Reference Catalog is the source of reference positions, this method could be used in a measurement of second epoch positions for the determination of proper motions using Carte du Ciel plates as first epoch. It is also found that errors of the measured positions are dominated by aberrations of optical system and centering error of stellar images.
