1 2004 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
  
  
  
hnique difference is not a displacement but it is caused by the random 
measurement errors (Erol and Ayan, 2003). 
Northern Track - Levelling+GPS 
Leica The conventional deformation analysis method was applied to 
veling all three data sets (levelling derived height differences, GPS S00 
vel. derived height differences and the combination of both) as is E 400 
going to be mentioned below. And the results that obtained = 00 (p "UB ns dise e da de de diem conn HP — 
ict and from evaluating these three data sets are shown in the graphics s 4s Hb 221211201 191 181174 161 15% 141131121111 6 3 1 
ict and (see Figure 2, Figure 3 and Figure 4). 80.0 
cted to Points 
Stochastic information involves the number of instrument setup @dh 1-2 Odh 2-3 Mdh 3-4 
while evaluating levelling derived height differences in the 
deformation analysis algorithm. EN T. D ET | 
zy T Figure 4: Combined derived height differences between 
; : s : consecutive epo { int i 2 t 
On the other hand, in the algorithm of the second evaluation, Secutive epoctisfopsachipointindhe nonhenmdrack 
the ellipsoidal height differences (dh;), which were derived 
di 8. 
from the GPS baseline solutions, are used as measurements, 8D DEFORMATION ANALYSIS WITH 
: see : ids : S-TRANSFORMATION USING GPS DATA 
Also, the stochastic information of measurements comes from 
the GPS baseline solutions. It can be accomplished the datum consistency between different 
: : : e : E epochs by employing the S-Transformation. Also, the moving 
ern In the third Sine height RIEN ie Ron both points are determined by applying this transformation 
measuring "tec Sues anevuse A ané deformation  consecutively. S-transformation is an operation that is used for 
analysis is applied according to results of evaluation of | transition from one datum to another datum without using a 
combined data groups. A very important point that has to be new adjustment computation With: another word: :S- 
IT considered in the first step of this evaluation is that the both | transformation is a transformation computation of the unknown 
neering 
measurements groups derived from both techniques do not have 
the same accuracy naturally. And so, the stochastic information 
between these measurement groups (relative to each other) has 
  
  
  
  
  
  
  
  
  
parameters, which were determined in a datum, and their 
cofactor matrix, from the current datum to the new datum. The 
equations that imply the transition from the datum i to datum k 
c. ithe to be derived. In this study, for computing the weights of both are given below (Demirel, 1987; Baskaya, 1995). 
d out in measurement groups, Helmert Variance Component Estimation 
lis way (HVCE) Technique has been used. k 
'erified. X =8.2 
bserved In here, only the graphics, which shows the height differences Qk =8 Q! ST 
of northern-track points, are given. Because of that the —x 0 Tkxxck 
deformations in both tracks points have the same character, so it S al ofa” E gj. TE 
ce steps has been found sufficient to give results belongs to only one Sp =U ER} OS Er 
ements, track of the viaduct. 
nes 5.1 Global Test Using S-Transformation 
nethod; F 
d in this NortherniTrack »Levelling A control network is composed by the datum points and 
out and deformation points. By the help of the datum points, the 
ch Were control network, which were measured in t; and t; epochs, are 
d ume 2 we transformed to the same datum. 
ustment s dde mn O0 -- 
ed free © mo |] 21 211 201 191 181 am 164 154 wr wet az mg 3 702 00 While inspecting the significant movements of the points, 
(Ayan, 9 continuous datum transformation is needed. Because of this, 
points first of all, the networks, which are going to be compared each 
0 dh1-2 Qair23 N dh3-4 other, are adjusted in any datum such as using free adjustment. 
esult of 
izing of Figure 2: Levelling derived height differences between In the result of free adjustment, the coordinates of the control 
lated for consecutive epochs for each point in the northern track network points, which were measured in any epoch, are divided 
they are into two groups as f (datum points) and n (deformation points). 
> Fisher re ee vas ee 
Northern Track - GPS Then, datum i and datum j were able to be transformed into the 
same datum k by the help of the datum points. As the result, 
909 the vectors of coordinate unknowns and also their cofactor 
pa En matrix are founded for the datum points in the same k datum. 
g 00 |np UB JL qu Jy mp By on flo cre Ty pf om ow — = ifm 
S Jools M 221 211 201 191 181 171 161 151 141 131 121 111 6 3 1 By the global test, it is determined if there is any significant 
|f 200 : : LA movements in the datum points or not. In the result of global 
Points test, if it is decided that there is deformation significantly in one 
Bdh 1-2 Odh 2-3 Wdh 3-4 part of the datum points, determining the significant point 
he point movements using S-transformation (localization of the 
d height 
Figure 3: GPS derived height differences between consecutive 
epochs for each point in the northern track 
deformations) step is started. In this step, as is assumed that 
each of the datum points might change its position, for each 
625 
DE