justification, and it is difficult to see why this trick should be so 
effective. On the contrary, it is a simple matter to construct image 
pairs, where the wrap-around effects will cause an erroneous 
determination of the matching parameter. An example is when there is a 
trend in the grey levels in the direction of translation of the 
images. Due to wrap-around, the image borders will then generate a 
persistent discontinuity at right angles to the direction of the 
trend. In the presence of noise, the effect on the matching parameter 
u of factual details in the images is weakend. When the noise is 
strong enough, the method breaks down with the result that jy « O. 
An obvious approach is to eliminate a linear function from the images 
before performing the matching. However, while this elimination will 
remove the discontinuity at the image boundaries, it will also 
eliminate trends in the images which belong to the factual 
information. This procedure therefore appears to be rather arbitrary. 
Also, when the image pairs used here were rematched after eliminating 
a linear trend, the results where not convincing. The preliminary 
calculations using the maximum entropy spectrum are more promising. 
Work will therefore be concentrated on investigating the properties of 
this method. 
CONCLUSIONS 
All the tests in this investigation involve the phase information 
contained in the Fourier spectrum in one way or the other. The idea is 
of course, that there exists a unique functional dependence between 
phase differences and translation. In the absence of noise, this is a 
correct notion for images of infinite extent, and also for the 
unrealistic case of cyclic translation of finite images. It has been 
demonstrated that for image pairs consisting of aerial photographs of 
finite size, the wrap-around effects and effects of noise in the phase 
components are so strong that the functional dependence between phase 
differences and translation is almost completely spoilt. 
However, maximum entropy methods were found to be promising both from 
a theoretical point of view and from investigations on synthetic 
images. Work along these lines is therefore continued. 
REFERENCES 
1. Kuglin, C. D., Hines, D. C.(1975): The Phase Correlation Image 
Alignment Method. Proc. IEEE Int. Conf. on Cybernetics and Society, 
p163. 
2. Pease, M. (1965): Methods of Matrix Algebra. Academic Press, New 
York. 
3. Ulrych,.T. J., Bishop, T. N. (1975)::;; Maximum: Entropy Spectrum 
Analysis and Autoregressive Decomposition, Rev. Geophys. Space Phys., 
vol 13, p183. 
4. Ulrych, T. J., Jensen, O. G. (1974): Cross Spectral Analysis by 
using Maximum Entropy, Geophysics, 39, p353. 
- 652 -