The alternative computation programs applied to these data gave very 
useful comparisons of their relative advantages, and also revealed the need 
for further improvements in the computation. Much of this improvement remains 
to be done, but sufficient was achieved to judge the suitability of the type 
of camera used. In particular, the radial distortion was rather large; 
equivalent to about 250pm variation in the principal distance (0.17%). The 
method of correction of this distortion, and the stability of the effective 
perspective centres, are both open to some doubt in this camera. A better 
corrected optical system would be expected to make possible a much more 
effective correction for distortion. Experience and mechanical tests with 
several cameras have shown that there is also some room for concern over the 
mechanical stability of the lens/plate relation after the camera has been set 
up. This may not matter if the camera is used only for single short 
exposures, and is not moved between exposures, as is the normal practice. 
Unfortunately some of our specialized techniques cannot conform to the 
intended mode of operation of the cameras. 
Further consideration of these requirements (and others), and evaluation 
of the available smaller format cameras, both metric and non-metric, 
convinced us that it would be necessary to choose the most stable and 
versatile system to be able to establish a fair assessment of the precision 
available at the present state of the art, and to have the best chance of 
justifying the incorporation of some of the optical improvements available to 
us to attempt to satisfy our specialized requirements. The choice eventually 
fell on the Zeiss Jena UMK 1318 10NF, and further trials were undertaken, 
mainly to gain handling experience, and to try out one or two immediately 
available techniques expected to give some improvement. 
ial with UM FP camera 
Since there was only one UMK 10NF camera (the version with minimum 
distortion at close range) in the U.K., further experience was obtained with 
the temporary use of a pair of UMK 10FP cameras. Although the distortion at 
close range (2.1m) is nominally some 10 times greater than the UMK 10NF, it 
is still 4 or 5 times less than the camera used in the previous test. The 
camera is equipped with illuminated fiducial marks, enabling operation with a 
dark field to be carried out easily. This mode of use is necessary for making 
multiple records of self luminous sources that are not illuminated 
simultaneously, and it is in this way that the verification of multi-axis 
measuring machines and other metrological structures can best be carried out. 
Targetti it lf-lumino source 
Photographic recording with self luminous sources has considerable 
advantages over conventional diffusely reflecting targets. The definition of 
the loeation ean be made as precise as may be required, limited only by the 
constraints imposed by the wave nature of light itself, the camera aperture, 
the lens aberrations, and the resolution of the photosensitive material. The 
ultimate limit of the size of the effective source. is the diffraction 
limitation (Gates, 1976). A single transverse mode laser oan provide 
milliwatts of power from such a diffraction limited source so that there is 
no difficulty in recording such an image with a short exposure. Trials showed 
that hundreds of exposures from such a source could be made on a single plate 
without the background being recorded, However, it can be very difficult to 
find the very small image generated by such a source (20-30pm in diameter for 
an f/22 lens aperture) so that an auxiliary marking pattern is desirable. One 
such system which has been used is shown in Fig. 1. The use of a high 
numerical aperture lens system provides illumination over a very wide 
188 
————