3-Dimensional Mapping of Corneal Topography and Thickness

Abstract

Optical sections of the cornea are obtained by illumination with a collimated beam expanded in a fan shape by a small rotary cylindrical lens. The light diffused from the cornea is observed by two cameras and processed in order to yield the surfaces' profiles. The optical system used to project a thin rotating line on the cornea consists of a white light source provided with optical fiber bundle output which is first conditioned by a set of lenses so that it would produce a spot on the cornea. A small cylinder lens is used to expand the beam in one direction so that a thin line illuminates the cornea, rather than a spot. The cylinder lens is provided with motor driven rotation about an axis normal to its own in order to rotate the line on the cornea such that the projected line scans the whole cornea; the illuminator is completed with a slit aperture. The cornea is not perfectly transparent, scattering some of the light that traverses it; this fact is used for its observation by two cameras. These are placed at pre-defined angles with the illumination axis, so that optical sections of the cornea can be seen; the use of two cameras avoids the need for camera rotation in synchronism with the cylinder. The two cameras' images can be combined in order to simulate a single virtual rotating camera. Image processing is used to extract information about the corneal surfaces profiles and thickness from the optical sections. Several peculiar aspects of processing are discussed, namely the corneal edge detection algorithm, the correction for angle of view and deformation due to observation of the inner surface through the cornea.