STEREOPHOTOGRAMMETRIC MAPPING OF THE ANTERIOR SURFACE OF THE HUMAN CORNEA 
Dr Jon Osborn, Department of Surveying and Spatial Information Science, University of Tasmania, Australia 
Dr Gordon Wise, Senior Ophthalmologist, Tasmanian Eye Clinics, Wentworth St, South Hobart, Tasmania, Australia 
Commission V, Working Group 5 
KEY WORDS: close range, non-metric, accuracy, medicine, corneal topography 
ABSTRACT 
The major refractive element of the human eye is the anterior surface of the cornea.. An accurate knowledge of the topography 
of this surface is important in a variety of ophthalmic applications. Existing commercially available methods of mapping 
corneal topography rely on measuring the images of mires reflected from the corneal surface. These methods are inadequate, 
particularly for abnormal corneas and intra-surgical application. 
This paper describes a stereophotogrammetric method for measuring corneal topography. Photogrammetric verification of the 
method shows that accuracy in the order of 20um (r.m.s) has been achieved. The instrument has been installed at an eye 
diagnostic clinic and tested on patients with corneal abnormalities. 
corresponding photokeratoscope images. 
Some results are illustrated and contrasted with 
The design of a digital implementation of the instrument is described, including details such as camera type, array size and 
resolution, target recognition and camera calibration. The clinical parameters that affect the commercial viability of a digital 
system are discussed. 
1. CORNEAL TOPOGRAPHY 
The topography of the anterior surface of the human cornea 
is complex. Descriptions usually divide it into zones — the 
central, paracentral, peripheral (transitional) and limbal — 
although these clinical divisions are probably not 
anatomically real (eg Dingledein and Klyce 1989). The 
central Zone usually has an approximately uniform spherical 
shape but with an irregular periphery and with considerable 
variability between individuals (Bogan et al 1990). The 
average cornea measures about 11.5mm vertically and 
12.1mm horizontally with a typical mean anterior radius of 
curvature of 7.8mm (Clark 1973b, Ruben 1975, Smith 1977). 
Large variations of corneal topography can be expected 
between healthy individuals (eg Clark 1974, Guillon et al 
1986, Dingeldein and Klyce 1989, Bogan et al 1990, Bores 
1991). Topography is affected by endogenous factors such 
as ethnic race, age, or congenital anomalies (Duke-Elder 
1970, Smith 1977, Guillon et al 1986, Bores 1991). These 
may not interfere with vision but are important if any detailed 
analysis of individual corneas is to be attempted. There are 
also many external determinants that affect the shape of a 
cornea and the performance of the eye. These include factors 
such as the presence of corneal pathology, the effects of 
surgical intervention, of drug induced changes and of 
mechanical forces (Duke-Elder 1970, Smith 1977). Probably 
the most important of these is keratoconus which is a disease 
that induces severe irregular astigmatism. Surgical 
procedures to improve defective vision have recently 
culminated in techniques such as radial keratotomy and 
photo refractive keratectomy  (P.R.K.). The clinical 
importance of these new refractive surgical procedures is a 
major impetus to recent research into methods of measuring 
corneal topography. 
Predominantly, because accurate measurement of individual 
corneas has not been possible, statements about corneal 
443 
topography have had to be generalisations based on relatively 
large populations (eg Clark 1974, El Hage 1976a, Kiely et al 
1982) or models derived from a theoretical consideration of 
the eye as an optical system (eg Patel et al 1993). The 
relationship between corneal topography and vision is 
critical, complicated and poorly understood. Quantification 
of what constitutes normal differences between individuals 
and normal changes in a single individual is an important 
step towards understanding corneal abnormalities. 
2. KERATOMETRY AND PHOTOKERATOSCOPY 
The two most commercially successful approaches to 
measuring corneal topography rely on measuring the images 
of mires reflected from the corneal surface. The first of these 
reflected mire techniques is keratometry, the second is 
photokeratoscopy (or videokeratoscopy when digital cameras 
are used). The value of any new method of measuring 
corneal topography must be established in terms of its 
benefits over these existing techniques. The principals of 
keratometry and photokeratoscopy have been addressed in a 
previous publication (Wise et al 1986) and the problems 
related to these techniques are only briefly reviewed here. A 
critical review of these and other methods of measuring 
corneal curvature is provided by Osborn (1995). 
Keratometry, by far the most widely adopted technique, has a 
number of inadequacies. 
e Curvature or net power is presented as a mean value for 
the central portion of the cornea (typically 3.8 to 4.2mm), 
but based on measurements at the edge of that area. Over 
90% of the corneal surface — the central and peripheral 
regions — is not measured. 
e It is assumed that the curvature of the cornea between two 
mire reflection points is spherical and thus (for a four point 
keratometer) that the cornea is ellipsoidal, which may not 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996