Why rods and cones?

A key problem in vision is to normalize one's lightness scale so that surface reflectances are always assigned the same gray value regardless of the illumination level. The solution requires an assessment of the relation between the strength of the illuminant and the strength of the image signal-information that is not available in the image alone. However, the level of scattered light in the optical system does provide an independent measure of the illuminant strength, and can be used to solve the lightness scale normalization problem. To do this requires a comparison between two imaging systems, each of which respond differently to the internal optical scatter. The rod and cone systems have properties that are ideally suited for such a role.