A method is presented that will enable the bivariate luminosity/surface brightness distribution of galaxies to be determined from a relatively small HIselected sample. This will be taken from the HIParkes All Sky Survey (HIPASS). The advantages of using an HIsample in order to avoid the selection effects that are present at optical wavelengths are discussed. We are developing an algorithm to automatically extract a uniform sample of galaxies from the HIPASSdata cubes and to determine the parameters of these galaxies. We have so far conducted tests involving both simulated sources injected into cubes with real noise and data from the Multibeam Deep survey. Results from these are encouraging.