The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
Measurements in the vicinity and outside the edges of the 
whiteboard are likewise omitted. Figure 6 presents the 
distributions of all distance residuals of the examined image 
sequence, grouped by the integration time they were captured 
with. Obviously, they all exhibit a negative offset from zero in 
the range of centimetres. 
Figure 5. Sample frames from the image sequence used for the 
integrated calibration of both the interior orientation and the 
ranging system, with detected targets indicated in red. 
[pixel! 
1: 4x4.2ms 
2: 4x4.2ms 
3: 4x4.2ms 
Xo ± O x o 
84.298 ±.041 
84.700 ± .028 
85.327 ±.047 
yo± Ovo 
-54.404 ± .033 
-54.500 ± .030 
-54.357 ±.050 
f ± Of 
200.459 ± .031 
200.743 ± .035 
200.884 ± .048 
a 3 ± Oa3 
-0.382 ±.001 
-0.386 ±.001 
-0.373 ±.001 
4: 4x2.2ms 
5: 4x10.2ms 
6: 4x20.2ms 
Xo ± 0x0 
84.695 ± .039 
84.653 ± .039 
84.860 ±.038 
yo ± OyO 
-54.521 ±.043 
-54.202 ± .037 
-54.631 ±.036 
f ± Of 
200.740 ± .035 
200.656 ± .042 
200.935 ± .040 
a 3 ± Oa3 
-0.387 ±.001 
-0.389 ±.001 
-0.378 ±.001 
Table 2. Interior orientation and radial distortion parameter, 
adjusted for groups of 1000 frames featuring uniform 
integration times that are indicated in the table header. 
Due to the limited extents of the used whiteboard, the whole 
distance measurement range cannot be covered. Nevertheless, 
the confrontation of the residuals with their reference values 
reveals parts of the periodic non-linearities reported in literature. 
See Figure 7 for details, bearing in mind that mean residuals for 
reference values above 2.5m are weakly determined. 
Figure 6. Distribution of distance deviations, grouped by 
integration time. 
Figure 7. Histogram of derived distances and residuals 
confronted with them. 
Figure 8. Histogram of observed amplitudes and their relation 
to range residuals. 
The comparison of residuals to the corresponding amplitude 
observations discovers a very clear relation, supported by an 
almost equal distribution of residuals over the encoding range. 
While range measurements with amplitudes above 2><10 4 do not 
appear to be sensitive to changes in the signal strength, there is 
a very strong effect on values below, see Figure 8. 
Examining the influence of the angle of incidence on the 
residuals does not provide much information. The seeming 
strong relation above 75gon is supported by few residuals only, 
see Figure 9. 
As a consequence of using only a subset of the range 
observations, the utilised residuals are not equally distributed 
on the sensor, which may be verified on Figure 10. Still, each 
pixel is related to at least 4000 residuals, which gives 
significance to their mean: with respect to the distances derived 
from the exterior orientation, the range observations are 
generally too large, still getting worse towards the principal 
point, roughly. 
Histogram of Incidence Angles 
Figure 9. Histogram of incidence angles and their 
confrontation with ranging residuals. 
Figure 11 presents the distribution of pixel footprints on the 
whiteboard. The precisely circular holes prove the high quality 
of target centroiding. As large-scale images have mainly been 
captured in the centre, the density of residuals decreases 
towards the borders of the test field. For that reason, the 
apparent effect of radially increasing mean residuals is rather 
little supported. The absence of scattering is proved again, as 
there are no circular structures around the voids of the target 
areas.