Most supernovae are thought to be unobserved because they occur in regions which are either highly reddened, or are behind regions of extreme reddening such that, in the visible, they suffer many magnitudes of extinction. To obtain the true rate of supernova events in different classes of galaxy it is important to be able to observe all supernovae, reddened, or not. A method of doing this would be to survey a sample of external galaxies at 10-microns. Highly reddened supernovae in relatively nearby galaxies would be bright at this range and subject to a minimal extinction. Detection efficiencies for an appropriately selected sample would thus be ~100%. This would allow the true rate of supernovae in external galaxies of different types to be calculated very exactly. A survey project of this kind would be a potentially unique contribution to the field of stellar evolution and supernova rates in different classes of galaxy.