tests refer in either case to identical geometry of the block and to 
the same measuring instrument (C 100). Thus the comparison shows di- 
rectly the error effects of digital point transfer. 
The results can be summarized on 3 levels: 
- The internal precision of digital point transfer, as indicated by the 
correlation algorithm, proved to be highly consistent, on the ex- 
pected level, with r.m.s. transfer parallaxes of 1.2 um and 1.1 um 
(see also fig. 1 and 2). 
- The Oo values obtained after the bundle block adjustment amount for 
signalized and digitized natural transfer points, respectively, to 
2.7 um and 4.0 pm (Schnürpflingen) and 3.3 um and 4.0 um (Appen- 
weier). It can be deduced from those values that the contribution of 
digital point transfer to the image coordinate errors of tie points 
amounts to 3.1 yum and 2.4 ym, respectively. 
- Both test blocks contained check points. They allowed assessment of 
the absolute accuracy of the adjusted blocks which refers, of course, 
to the respective geometric configuration of the olocks. The r.m.s. 
differences of adjusted points against planimetric check points ob- 
tained with signalized tie points and digitally transferred natural 
tie points, respectively, were AML, 3.3 um and 3.7 um (Schnürpflin- 
gen) and u = 6.6 um and 7.3 am (Appenweier). It can be deduced 
therefrom that the error contribution of digital point transfer to 
the absolute planimetric accuracy of the adjusted blocks amounts to 
1.7 um and 3.1 um, respectively. The absolute vertical accuracy of 
the block Appenweier with digital point transfer was 9.8 um = 
0.06 %h, as estimated from (only) 10 vertical check points. 
The experimental results demonstrate clearly the high precison of digi- 
tal transfer of natural tie points by digital image correlation. The 
transfer errors of image coordinates are only in the order of 3 um or 
better. Thus, it is shown that digital transfer of natural tie points 
can raise conventional aerial triangulation to the high precision level 
which so far was only obtainable with signalized tie points. The 2 ex- 
periments refer to large scala photography (1:4 000 and 1:7 800). How- 
ever, similar high precision of digital transfer of natural tie points 
is expected also with medium scale and small scale photography. 
References 
/1/ Ackermann, F.; Bettin, R. (1980): An Empirical Accuracy Test on 
Point Transfer in Aerial Triangulation, IAP Vol 23-B9, Hamburg, 
1980, pp. 302-312 
/2/ Ackermann, F.; (1984): Digital Image Correlation: Performance and 
Potential Application in Photogrammetry, Photogrammetric Record, 
Vol. 11, No. 64, 1984, pp. 429 - 435 
/3/ Pert], A. (1984): Digital Image Correlation with the Analytical 
Plotter Planicomp.C 100, IAP Vol. 25-A3b, Rio de Janeiro, 1984, 
pp.878-882 
/^/ Ackermann, F.; Schneider W. (1986,: Empirical Investigation into 
the Precision of Digital Image Correlation, pres. paper, ISPRS 
Symposium Com. III, Rovaniemi, 1986 
-326.-