The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Voi. XXXVII. Part B5. Beijing 2008 
We compared the tie point coordinates of the Version 21 
(February 2007) and Version 61 (September 2007) results. The 
means of the differences (February’07 - September’07) are 
+0.043, +0.002 and +0.004 mm for the X, Y and Z axes, 
respectively. The standard deviations of the coordinate 
differences (0.089, 0.169 and 0.106 mm for X, Y and Z axes, 
respectively) are considerably smaller than the empirical 
accuracy numbers. 
- ~j - -_-=J - ; — • - 
N93 - image: 10 (version 61) N93 - image: 13 (version 61) 
(a) (b) 
Figure 8. Systematic residual patterns of images 10 (a) and 13 
(b) of version 61 after the self-calibrating bundle adjustment. 
the best results with respect to randomness and magnitude of 
the averaged residual vectors (Figure 9f). 
4. ANALYSIS OF RESULTS 
In spite of giving the worst results in the test, the K750i still can 
offer sub-millimeters accuracy in object space. Both block- 
invariant 10 and 44 additional parameter sets cannot 
compensate the systematic errors fully. 
The cameras K750i, N93 and W100 give identical standard 
deviation values for the image observations (between 1/4 - 1/5 
pixel). They all apply a chip level image enhancement for 
sharpening the images. This effect is visible in Figures 5a, 5b 
and 5c. This low level image enhancement, while improving the 
visual quality, is probably reducing the geometric quality of the 
cameras. They show noticeably block-variant systematic errors 
after the self-calibrating bundle adjustment with block-invariant 
additional parameters. 
The change of the principal point locations (x 0 and y 0 ) and the 
focal length (c) between Version 21 (of February 2007) and 
Version 61 (of September 2007) are only -1.2, +1.0 and -1.8 
microns, respectively. The corresponding standard deviations of 
the differences (calculated according to the law of error 
propagation without considering the correlations) are ±1.1, ±0.8 
and ±0.6 microns, respectively. 
3.7 Image residual analysis 
In an image residual analysis we average the directions and the 
magnitudes of the residual vectors at pre-defmed grid locations 
(in our case at 24 x 18 locations). This shows the nature of the 
systematic errors remaining after the self-calibrating bundle 
adjustment (Figure 9). The systematic error pattern of the 
twelve image version (Figure 9b) of the K50i is similar to the 
eighteen image version (Figure 9a). We also note that the image 
residual analysis results of the two epochs’ calibration results of 
the N93 are almost identical (Figure 9c and 9d). The F828 has 
The N93 and W100 have the same lens systems (Zeiss, Vario- 
Tessar). The W100 has a CCD sensor of larger size with 8 
Mpixels. It is 2.5 times larger than the CMOS sensor of N93. 
According to theoretical expectations, the N93 should give an 
accuracy of factor 1.6 (square root of 2.5) worse compared to 
the W100. The N93 almost strictly meets this expectation by 
giving 1.7-1.9 times worse numbers than the W100. On the 
other hand, there is a large difference between those two 
cameras, considering the size of the imaging system and the 
cost of the materials used in the construction. In this respect, the 
accuracy performance of the N93, as compared to the W100, is 
noteworthy. Although the W100 and F828 have the same image 
format with 8 Mpixels, the expectation of equal accuracy does 
not hold here. The W100 gives substantially worse accuracy 
numbers (almost 3 times) than the F828. This is mainly due to a 
better lens system of the F828 and (possibly) the degrading chip 
level image enhancement operation of the W100. 
K750Ì-version 1Ì 
N93 - version 61 
F828 - version 41 
(b) 
(d) 
(f) 
Figure 9. Graphical results of the image residual analysis, based on residual averaging over all images of a particular test 
configuration. 
734