In order to perform the astrometry on the images we first aligned all the frames in the same band taken on the same night for the same object using the IRAF task IMSHIFT and the positions of a few field stars. Then, using the positions of a larger number of field stars ( > 100), as given by the IRAF task STARFIND, we registered the images in the different bands using the GEOMAP and GEOTRAN tasks. Finally, once all the frames were combined and the various bands registered, we used the IRAF tasks STARFIND and CCXYMATCH to cross-correlate the positions of the field stars in pixels with their FK5 equatorial coordinates in the USNO-A2.0 catalogue. The cross-correlation was done using the positions of the stars in our final R-band image. The H (line+continuum) image was used in those cases where a large number of USNO-A2.0 stars were saturated in the R-band image (e.g. in those cases of exceptional seeing). Once the plate solution was computed and stored in the image header (using the IRAF tasks CCMAP and CCSETWCS) we copied the WCS information to the rest of the bands using the task WCSCOPY. Typical (rms) errors in the astrometric calibration of our images are 0.1 arcsec both in RA and DEC.