Full text: Proceedings, XXth congress (Part 1)

   
  
| 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B1. Istanbul 2004 
ctor in 
object 
itively 
en the 
dles at 
cantly 
errain 
S. EXPERIMENTAL RESULTS 
A total of nine digital cameras were calibrated and evaluated for 
stability over an eight-month period. For each camera, with the 
exception of the four Nikon 4500 cameras, image datasets were 
acquired in two or more months. For the four Nikon 4500 
cameras, two image datasets were captured on the same day by 
  
  
  
  
  
  
  
  
  
  
  
  
EOP, de Y 
ilarity simply switching the camera off and on between datasets. 
a 
ust be As mentioned in Section 3.2, to check the stability of a camera, 
object the estimated standard deviation (o,) component resulting from 
the adjustment procedure represents the spatial offset between 
the two bundles along the image plane. If this value is not 
eias significantly larger than the expected image coordinate 
, € . J 
sional measurement accuracy, which can be considered to be 
for approximately two-thirds of a pixel, then the two IOP sets are 
similar. 
kd: 3:5 deemed 
, dark The stability results for the nine digital cameras (denoted by 
id the their experiment names) are listed below. The comparison 
s used method implemented is the ROT method with K1 estimated and 
or the four distances measured in the calibration procedure. 
ites of : Fa 
y the i. CanonOne - For the Canon camera, the standard deviation 
ereing of the spatial offset (o,) must be less than 0.0077 mm, 
at are which corresponds to two-thirds of a pixel, for the IOP sets 
on the to be considered similar. From Table 2, it can be seen that 
image all IOP set comparisons passed this test of similarity. 
| Date Simila 
| ID | IOP Set IOP Set a, (mm) ; 
| I Il 
| 1 | Jul03 Oct 03 0.00385 Yes 
| 2 Jul 03 Jan 04 0.00142 Yes 
^ 3 | Oct03 Jan 04 0.00469 Yes 
Table 2 — Stability Comparison of IOP sets for CanonOne 
(Note: If 6, « 0.0077, IOP sets considered similar) 
ii. CanonTwo — From observing Table 3, all IOP set 
comparisons reveal that the IOP of this camera is stable 
8 over a short and long period of time since o, is less than 
e 0.0077 mm. 
1alysis Date Simila 
USD. ID | JOP Set 10P Set a, (mm) 
arged- H 
NS the 1 Jul 03 Oct 03 0.00553 Yes 
ides 2 Jul 03 Jan 04 0.00480 Yes 
fective 
Pixels 3 Oct 03 Jan 04 0.00614 Yes 
Pixels) 
Table 3 — Stability Comparison of IOP sets for CanonTwo 
4.15 (Note: If a, « 0.0077, IOP sets considered similar) 
3.87 iii. Rollei — The results in Tab/e 4 demonstrate that the IOP sets 
of the Rollei are similar for the six comparisons since o, is 
4.90 within the acceptable image coordinate measurement 
accuracy of 0.003 mm (% of the Rollei pixel size). 
Vi. 
  
  
Date A 
m Pi E x 6, Sintila 
DE Det (mm) r 
I II 
1 Jul 03 Oct 03 0.00290 Yes 
2 Jul 03 Jan 04 0.00147 Yes 
3 Oct 03 Jan 04 0.00146 Yes 
4 Jul 03 Feb 04 0.00201 Yes 
S Oct 03 Feb 04 0.00180 Yes 
6 Jan 04 Feb 04 0.00103 Yes 
  
  
  
Table 4 — Stability Comparison of IOP sets for Rollei 
(Note: If à, « 0.003, IOP sets considered similar) 
iv. SonyF707 — For the SonyF707 experiment results given in 
Table 5, all comparisons of IOP sets indicate that the IOP of 
this camera remains stable since o, is within the acceptable 
image coordinate measurement accuracy of 0.003 mm. 
  
  
Date : 
De De (mm) r 
I IT 
1 Jul 03 Oct 03 0.00206 Yes 
2 Jul 03 Feb 04 0.00176 Yes 
3 Oct 03 Feb 04 0.00291 Yes 
  
  
  
Table 5 — Stability Comparison of IOP sets for SonyF707 
(Note: If o; < 0.003, IOP sets considered similar) 
. SonyP9 — Image datasets were acquired only in the months 
of July and January, and as shown in Table 6, the IOP 
comparison of the SonyP9 gave a standard deviation 
component (c,) well below the required image coordinate 
measurement accuracy. However, more comparisons need 
to be made to determine how stable the camera really is. 
  
  
  
Date s 
ID a, Simila 
IOP Set IOP Set (mm) r 
I II 
1 Jul 03 Jan 04 0.00145 Yes 
  
  
  
Table 6 — Stability Comparison of IOP sets for SonyP9 
(Note: If à, < 0.003, IOP sets considered similar) 
Four Nikon Cameras — As mentioned earlier, the Nikon 
datasets are acquired on the same day by switching the 
camera off and on between dataset acquisitions. The results 
in Table 7 indicate that all four Nikon cameras do not 
maintain the same IOP values and are considered unstable. 
  
   
  
  
   
      
  
   
    
   
  
    
  
  
  
   
  
     
  
    
    
   
  
   
    
  
   
  
   
  
   
   
    
  
  
   
  
  
  
  
  
  
   
  
  
  
  
   
  
  
   
   
   
   
  
  
  
   
   
  
  
  
  
  
  
  
 
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.