International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
By several ground control points have been observed and by a 
is | final adjustment. 
the 
em A number of previously surveyed points were signalised with 
the reflecting material, which allows a good identification of the 
the | ground control points in the laser intensity image (see second 
ere image on figure 6). Several scenes were captured while the 
ine, van was stationary and the GPS/IMU sensors were collecting 
the data. Afterwards, an adjustment was done for computing the 
nt. offset and the misalignment matrix. The translation. and 
  
rotation matrix (XGpsAaMU: Y GPsaMU. ZGpsimu) and M; were 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
A derived form the GPS/IMU trajectory and used as 
the observations in the adjustment. The measurements of the 
nce signalised points in the intensity image were also used as 
the observations. In order to improve the determination of the 
the parameters and to decouple their correlation four scenes 
lone taken with different azimuth angles were used in the 
rom adjustment. Figure 7 shows the distribution of the adjustment 
tion with the red ellipses indicating the position and formal error 
step of the IMU orientation, the blue ellipses indicating the 
| the position and formal error of the laser orientation and the very 
the small black ellipses showing the position and formal error of 
ence the 15 signalised points used (notice that some of the points 
nent differ only in the vertical coordinates and therefore, are 
Cms shown superimposed in figure 7). The black lines link the 
laser position with the ground control points and symbolise 
ation the laser image measurements of the signalised points. 
ue to 
stant The results and the formal accuracies of the adjustment are 
su presented in table 1. 
the 
wi value o 
N 
RIS X -2.169 m| 0.006 m 
LE = Figure 6: Calibration scene. Combination of intensity and 
ihe e Y 0.007m| 0.006 m distance image (above), intensity image (centre) 
o and RGB image (below) (signalised points are 
7 0.462 m| 0.006m easily identified in the intensity image) 
€ e -892 59° 3° 34" 3 
E 
DOT © 09 1* 32" j2* 
= K -89° 59° 49” 12.4 A - a M E t 
tesian S i = e 
tesian Table 1: Determination of the offset and misalignment matrix 
in the adjustment e 
t laser 4. DYNAMIC LASER SCANNING us uw 
en the had NUR 2 
{frame Once the laser was synchronized to GPS time and the v 
nized), constant offset and misalignment matrix computed by using Figure 7: Laser mounting calibration: adjustment of four 
| mode the method described in the previous section, it was possible A laser static scenes! Bed' position of the GPS/IMU 
d asa to use the terrestrial laser as a pushbroom sensor by fixing sensors. Blue: position of the terrestrial laser. 
eeping the scan angle and sweeping the scene with profiles while the Black: control point measured in the laser scenes 
Yt. vehicle is in movement. Every single point measured by the 
laser is transformed to the mapping reference frame by using 
Its corresponding orientation and applying the formulas 
described in equation 4. Figures 8 and 9 show the laser RGB images collected in a 
kinematic mode. The laser was vertically mounted on the 
Kam integration platform and the rotating mirror was performing 
w 
vertical scans looking at the right side of the van while it was 
moving. In figure 8 the exterior buildings of a roundabout 
993