We have used maximum entropy eclipse mapping to recover images of the visual surface brightness distribution of the primary component of the RS CVn eclipsing binary SV Cam , using high-precision photometry data obtained during three primary eclipses with STIS aboard the Hubble Space Telescope .
These were augmented by contemporaneous ground-based photometry secured around the rest of the orbit .
The goal of these observations was to determine the filling factor and size distribution of starspots too small to be resolved by Doppler imaging .
The information content of the final image and the fit to the data were optimised with respect to various system parameters using the \chi ^ { 2 } landscape method , using an eclipse mapping code that solves for large-scale spot coverage .
It is only with the unprecedented photometric precision of the HST data ( 0.00015 mag ) that it is possible to see strong discontinuities at the four contact points in the residuals of the fit to the lightcurve .
These features can only be removed from the residual lightcurve by the reduction of the photospheric temperature , to synthesise high unresolvable spot coverage , and the inclusion of a polar spot .
We show that this spottedness of the stellar surface can have a significant impact on the determination of the stellar binary parameters and the fit to the lightcurve by reducing the secondary radius from 0.794 \pm 0.009 R _ { \odot } to 0.727 \pm 0.009 R _ { \odot } .
This new technique can also be applied to other binary systems with high precision spectrophotometric observations .