In order to determine the emission height of the optical photons from pulsars we present an inverse mapping approach, which is directly constrained by empirical data. The model discussed is for the case of the Crab pulsar. Our method, which uses the optical Stokes parameters, determines the most likely geometry for emission including the magnetic-field inclination angle (α), the observer's line-of-sight angle (χ) and emission height. We discuss the computational implementation of the approach, and the physical assumptions made. We find that the most likely emission altitude is at 20% of the light-cylinder radius above the stellar surface in the open field region.