! B2. [Istanbul 2004 ; ; : ol AT ; ; . ap YYYV 
BZ Istanbul 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
  
dinates and ground 
venient to evaluate 
adjustment. Table | 
Table 2. Accuracy of bundle block adjustment 
  
GCPs Check points Exterior orientation elements 
| | 
Ou | | 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
| set of simulating | 
magery| | | ae X 7 animetrv | Hejeht | € c 7 7 
| full | height | (um) | Number | Unconformity x ; Paten Holo | dl a | K X, X, Z, 
| (m) (m) (m) | (m) | (13-10 | (G) (m) (m) (m) 
| minimum value 0.000 | 0.000 0.003 | 0.001 | © D; 0 0.001 | -0.007 0.000 
——M— : 3 maximum value | -0.410 | 0.350 0408 | 0.945 | -56 | 53 | 33 | 0948  -0.803 | 0.368 
| 161 30) 55-4083 : | | | | 
practical accuracy | 0.014 | 0.090 0166 | 0.255 | 20 19 | 11 0.341 0.304 0.123 
agery-5 m TRA | | | 
nager theoretic accuracy | 0.082 [ 0.089 9121 | o211| 20] i8 |. e [ 0321. 0310 |]. 0126 
G3 50i ai | minimum value | 0.000 | 0.000 0.012 | -0.001 | 0| 0j 0 -0053 -0029 | -0.004 
3x7 55 8 = Pi maximum value | -0.790 | 0.530 QU DIS | 47} 501-42 1.468 1.560 | -0.473 
a 3 22 5.7 | | | i 
36-66 practical accuracy | 0.280 |‘ 0.160 #324 | 0443 1 20 16 | 12 | 0648 | 0542 | 0.194 
629 theoretic accuracy | 0.164 | 0.177 0242 | 0415 | 18 18.16 0.591 |. 0.619 0.251 
Rio d Note: 1). Unconformity 4; is the difference between computed value and theoretic value. 
mmm 2). Practical accuracy ist; = 2 4 /n.G - x. v. 2) and Ho EAL ral. 
3). Theoretic accuracy is m, = Go ftr(Q, )/n.u - x.v.zy and Mans Zr Km. 
Table 3. Accuracy of bundle block adjüstment by using fixed imageries without any GCP 
i | | Check points- : Exterior orientation elements 
m. It accords with Method! Zo. Y Y Z 
SOS in imaoe | : . X A i y ig w Ÿ À. 5 ; 
M CITOIS IN tmage (um) Number | Unconformity N ( Y Pee Heicht 2 : E 3 3 s 
isted coordinates of | R (qw m) um e | (9 | (GO ] O3 (m) (m) (m) 
theoretic accuracy. | minimum value | 0.000 | 0.000 0.003 0.000 0 0 | 0 | -0.004 | -0.601 | 0.004 
ulated photography A. | 5.0 | 667 | maximum value | 0:440 | 0.280 0446 | 0946 | -47 | -57 | 20 | 0784 | -0.764 | -0.438 
ed ur the following | : re eee : : ; ; s 
| practical accuracy | 0.116 | 0.090 0.143 0.228 17 16 6 0.292 0.266 }- 0.163 
: | minimum value | 0.000-| 0.000 0000 | 0.006 | 0 0 | 0 | -0043 | -0037 | -0.019 
ANALYSIS B 40 | 242 maximum value. | -0.800 | -0.500 0.830 1.144 | -37 -34 32 1.803 1.240 | 0.694 
io. | practical accuracy | 0.290 0.170 0.332 0.372 15 13 14 0.688 0.424 0.268 
sets of simulated | = T 5: | 0 
cation points are | minimum value 0.006 1 0.000 10.008 -0.011 0 0 0 0.004 0.005 0.147 
justment with the JA 5.0 | 667 maximum value 0.470 -0.360 0.474 -1.417 | -30 -31 -44 0.876 1.104 0.983 
nent methods are | practical accuracy | 0.120 | 0.090 0.145 0.621 9 11 10 0.2800 | 0.368 0.605 
minimum value 0.000 0.000 0.006 1.331 0 0 0 -0.006 | -0.002 | -1.694 
D | S31 697 maximum value | -1.240 0.540 1.238 3.396 69 103 -34 1.290 | -2.123 | -2.889 
the onentation | | practical accuracy | 0.300 0.150 0.335 2.461 22 29 12 0.486 0.561 | 2.238 
je coordinates of : 
dis orientation From Table 3 the following conclusions can be drawn: . CL à | P3 
le coordinates of -- : . 3) The resu ts of met 10d C shows that the accuracy of 3 
1) Table 3 shows that the accuracy of the orientation coordinates of all photogrammetric points is satisfying 
parameters of new imageries and 3D coordinates of when a bundle adjustment is conducted by combining a 
the orientation ground objects is decided by the fixed imageries, the large-scale fixed imageries with a small-scale new 
€ coordinates of larger the scale, the higher the accuracy; the smaller the imageries in the same area. Comparing to the results of 
scale, the lower the accuracy. And the accuracy of the conventional bundle block adjustment in table 2, we can 
the orientation unknowns relating to planimetry is almost identical find that the accuracy of the exterior orientation elements 
€ coordinates of between the mentioned adjustment and the conventional of the new imageries and the 3D coordinates of ground 
bundle block adjustment, but the accuracy of the objects has no substantial difference. 
unknowns relating to height is fluctuated. 
: = ari Se : ^ > ) H e adiuste 
nt methods. These ea = : 4) Comp wing the results of method D with the adjusted 
| by combine the 2) The results of method A and B shows that the accuracy of results of imagery-5 in table 2, we can see that the 
on elements of the 
tric practice, the 
calculated by the 
e second section, 
zeries. And image 
e 5 
€ Sum measuring 
je 3 
the combined bundle block adjustment is highest when the 
two periods of imageries are at the same scale. Comparing 
table 3 with table 2, we can learn that the exterior 
orientation elements of new imageries obtained in this 
paper are close to the ones calculated by the conventional 
bundle block adjustment with GCPs. And the accuracy of 
3D coordinates of all photogrammcetric points is identical 
between the two methods. 
accuracy will decrease apparently when small-scale fixed 
imageries are used to calculate the exterior orientation 
elements of large-scale imageries in the same area. The 
planimetric accuracy of densification points doesn't 
change much, but the height accuracy decreases terribly. 
In spite of the fact that the adjusted results still meet the 
specification for 1:1,000 topographic mapping in 
mountainous area. The coordinate residuals of check 
points are less than 5.0m in planimetry and 3.0m in height