International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
target is rejected from the sequence (see the abnormal position 
and velocity vector in Fig. 9, at the bottom). 
  
800 
600 T - T. s “4200 
  
Figure 9: 3D position and velocity vectors of the targets at 
different times 
In order to assess the repeatability of the test under identical 
condition, the trajectories of target 79 (2nd column, middle) in 
four test have been superimposed on the same graph (see Fig. 
10 and 11). 
Confronto dellaposizione in x nel tempo 
1500 - 
| prova del 04-09-03 
1400 1 prova del 09-08-03 
| prova del 28-09-03 
| — prova del 29-09-03 de 
1300 re ee re rs 
1200 
1100 
= 1000 oA 
= yt 
900 
800 
700 
600- 
500 
0 0.2 0.4 0.6 0.8 1 12 14 16 
Figure 10: position of the target 79 at different times 
It can be seen that the agreement is quite good, especially for 
velocities (the differences in the final position are up to 20 cm, 
this depends on a different height of the sand specimen in these 
tests). 
Pallino79 
3000 ---- s : 
prova del 04-07-03 
prova del 09-07-03 
2500 prova del 28-07-03 
prova del 39-07-03 
2000 - 
r3 , 
E FR 
E 1500: FEN 
= N 
> « 
1000 X 
s. / N 
j 
0 05 + 15 8 259 3 335 4 45 
tempo (s) 
  
Figure 10: position of the target 79 at different times 
4. CONCLUSIONS 
Overall, both cases were solved to a high degree of automation, 
and with accuracy level that, perhaps of not-so-high quality in 
the second case, but still more than enough for the application 
at hand. Using a prediction model to label the target along the 
sequence seems to be a feasible and flexible technique, but its 
reliability with only two cameras and a low frame rate cannot 
be guaranteed and demands additional efforts to control the 
results. 
References 
Montrasio, L., and Nova, R., 1989. Assestamenti di una 
fondazione modello sotto carico inclinato: risultati sperimentali 
e modellazione matematica, Rivista Italiana di Geotecnica, 
pp.35-49. 
Hampel, U., and Maas, H.-G., 2003. Application of Digital 
Photogrammetry for Measuring Deformation and Cracks during 
Load Tests in Civil Engineering Material Testing. In Proc. of 
Optical 3D Meas. Tech. VI, Zurich, pp. 80-88. 
Gruen, A., 1985. Adaptive least squares correlation: a powerful 
image matching tecnique. South African Journal of Photog., 
Remote Sensing and Cartography, 14(3), pp. 175-187. 
Gruen, A., 1996. Least squares matching: a fundamental 
measurement algorithm. In: K. Atkinson (ed.), Close Range 
Photogrammetry & Machine Vision, Whittles, pp. 217-255. 
Baltsavias, E.P., (1991). Geometrically Constrained Multiphoto 
Matching. Mitteilungen No. 49, Inst. of Geodesy and 
Photogrammetry, ETH, Zurich. 
Gordon, S.J., Lichti, D.D., Chandler, I., Stewart, M.P., Franke, 
J., 2003. Precision Measurement of Structural Deformation 
using Terrestrial Laser Scanners. In Proc. of Optical 3D 
Measurement Techniques VI , Zurich, pp. 322-329. 
References from websites 
www.cimne.upc.es, accessed at 1 May 2004. 
      
  
  
  
  
   
   
   
  
   
    
  
   
  
   
  
  
   
   
   
   
   
   
   
  
   
  
  
  
  
  
  
  
    
  
    
   
  
   
  
   
  
   
    
   
  
   
  
  
    
    
   
   
  
   
  
   
CAMI 
KEY WOR 
ABSTRAC 
Single imag 
pal point cai 
ted vanishin 
constant (as 
of only two 
tions and th: 
image syste 
absolute cor 
rameters (in 
too, at the e: 
Apart from 
lengths, len; 
imaged scer 
expected to 
quired here; 
implemente 
grammetric 
Recovering 
grammetry : 
are generall 
the prior kn 
object geom 
3D object, p 
ct al. 2002: 
hand, exploi 
ported in pl 
Voss, 2001: 
as well as in 
et al., 1999; 
type of appr 
ty among ol 
environmen: 
Indeed, it is 
tion (camer: 
ed — along s 
gonal vanish 
is the orthoc 
points (Mert 
in this conte 
& Voss (20( 
the latter au 
is close to ir 
justed line o 
calibration. | 
recently repc 
a single step 
cluding radi 
(Grammatik