The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
913 
Most part of the track has been surveyed in both ways of traffic 
direction. Therefore an “internal” quality check consists in the 
coordinates comparison taken in the two directions. This 
operation permits to evaluate the effects of GPS/IMU bias in 
coordinates determination due to different boundary conditions 
such as geometry constellation and typology of the track. This 
kind of control is a standard quality control and permits to point 
out where the trajectory is out of tolerance or where there could 
be an error in position and orientation determination. About 50 
control points have been measured, in both way, and the graph 
in Figure 8. shows the result, ordered in N, E and h. 
-0,20 -0,15 -0,10 -0,05 0,00 0,05 0,10 0,15 0,20 
(m) 
Figure 8. Comparison of 50 control points surveyed on forward 
and backward trajectories (internal comparison) 
Like mentioned before, most part of the track has been 
surveyed in both way of traffic direction, therefore an “internal” 
quality check consist in the coordinates comparison taken in the 
two directions. This operation permit to evaluate the effects of 
GPS/INS bias in coordinates determination due to different 
boundary conditions such as geometry constellation and 
typology of the track. This kind of control is a standard quality 
control and permits to point out where the trajectory is out of 
tolerance or where there could be an error in position and 
orientation determination. About 50 control point have been 
measured, in both way, and the graph in Figure 8. shows the 
result, ordered in N, E and h. 
As the figure 8 shows the obtained results are really good; the 
discrepancies between the forward and backward determination 
are quite the same, both in planimetry and in height. The 
repeatability is contained within 20 centimeters for the greatest 
part of the coordinates. 
3.3 Comparison between MMS and RTK 
After checking the internal accuracy it is possible to evaluate 
the so called “external” comparison. This phase consists in the 
comparison between points surveyed both by Road-scanner and 
using another approach generally more accurate such as classic 
survey or GPS/RTK survey. This methodology was used to 
survey some points identified on the images. For this 
comparison 62 points located in four different areas have been 
used. 
Zone 
Component 
Min 
Max 
Mean 
RMS 
1 
DN 
-0,09 
-0,39 
-0,15 
0,09 
DE 
0,01 
0,10 
0,01 
0,09 
Dh 
0,54 
0,64 
0,59 
0,03 
2 
DN 
0,00 
0,40 
0,09 
0,17 
DE 
0,00 
0,21 
0,04 
0,07 
Dh 
0,10 
-0,06 
-0,01 
0,03 
3 
DN 
0,00 
0,04 
0,00 
0,02 
DE 
0,00 
0,23 
0,00 
0,12 
Dh 
0,12 
0,44 
0,05 
0,05 
4 
DN 
0,01 
-0,32 
-0,02 
0,12 
DE 
0,01 
-0,26 
-0,02 
0,14 
Dh 
0,02 
-0,12 
-0,03 
0,06 
Table 3. Results of comparison between Photogrammetric 
determination and RTK for Control points (External 
Comparison) 
The Table 3. shows, for each zone, the statistical parameters for 
the control points. The best results can be found in the zone n°3 
where the max value, for the difference with RTK, for N, E and 
h is respectively 4 cm, 23 cm and 40 cm, but the mean value of 
the differences are, for the same components, 0 cm, 0 cm and 5 
cm. The accuracy in the height component is comparable with 
the planimetric one, but in zone 1 a shift of about 60 cm has 
been detected, associated to a small value of r.m.s. 
The interpretations of these results are in progress but 
apparently no different boundary conditions are evident for the 
different areas 
4. OPEN ISSUES; LEAST SQUARE ADJUSTMENT 
APPROACH USING A POSTERIORI SURVEY. 
Considering the obtained results, some issues are still open. 
The GPS/RTK survey in this work has been used to check the 
quality of the obtained product and to guarantee to the 
purchaser the quality of the performed work. 
The performed comparison constitutes an indicator that the final 
product is good at decimetre level. 
Furthermore this result was checked and reached in areas where 
GPS, both in RTK survey and in MMS system, fixes ambiguity. 
It is necessary to stress as in some areas along the track the sky 
visibility is strongly reduced due to the presence of close and 
high buildings (such as the historical city centre) and in these 
area the number of common satellites can decrease dramatically. 
In figure 9 the number of common satellite tracked by the MMS 
along the track during the survey is represented. In the figure 
the red points represent the areas where the number of satellites 
is lower than 4, the yellow ones represent the areas where the 
satellite’s number a good condition correspondent to a satellite 
number bigger than 5. From these picture we can deduce that in 
the red areas the navigation was charged completely to the IMU 
sensor, that, as well known, is affected by drift (mostly time 
dependent). The drift decreases the precision of positioning and 
attitude along the track and can be very significant when the 
window time is consistent.