AAO Newsletter October 1996 - Page 6
Figure 4: Plots of median RMS centroiding errors in xy
for various survey-grade emulsions.
comparison is shown in Figure 2b. The histograms of shift vectors remaining are shown in Figure 3b, and the single co-ordinate, single measure standard errors of these distributions are x = 0.15µm and y = 0.15µm; once more, these are consistent with the absolute errors measured previously.
Centroiding precision for survey emulsions
Three plate pairs were chosen from the UKSTU collection to illustrate the typical centroiding errors with SuperCOSMOS measures for relative stellar positions from survey grade material. Figure 4 show xy residuals for 103a-O, IIIaJ and 4415 emulsions. This demonstrates the advantages of the fine-grained IIIa and 4415 emulsions over the coarse 103a-O, and shows that SuperCOSMOS comes close to achieving the theoretical grain-noise limiting accuracy of ~ 0.3µm for well-exposed images.
By careful optimisation of the xy measuring table of SuperCOSMOS, we have achieved repeatabilities and absolute positional accuracies having single co-ordinate, single measurement standard errors of < 0.1µm and sigma ~ 0.15µm respectively. SuperCOSMOS is therefore capable of providing accurate co-ordinates for fibre spectrographs like the 2dF facility. For the purposes of relative astrometry of stars the precision of SuperCOSMOS enables, for example, the determination of relative proper motions at an accuracy of a few mas/yr using the first and second epoch sky survey Schmidt plate material in both the northern and southern hemispheres. There is now a great potential for studies in, for example, stellar kinematics and Galactic structure.
Many people have contributed to the success of the SuperCOSMOS project. Among them are Dennis Kelly, Steven Beard, Clive Davenhall, Bernard McNally, Magnus Paterson, John Cooke, Tom Paul, Joel Sylvester, Richard Bennett, John Harris and Eve Thomson.
N. C. Hambly, L. Miller, H. T. MacGillivray, J. T.
Herd & W. A. Cormack
Royal Observatory, Edinburgh