93
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008
Target
Number
Day
Model
Number
MSE of
Intersectio
n [pm]
RMS at Target
Points [m]
X
Y
Z
43142
Ma
y 23
501
59.8
1.20
-0.09
0.80
43143
Ma
y 23
506
30.8
0.39
0.68
2.0
43144
Ma
y 23
505
29.6
0.85
1.257
-2.5
47803
Ma
y 24
457
48.4
2.44
2.89
4.89
47804
Ma
y 24
457
48.4
2.13
2.59
4.28
46465
Ma
y 26
199
30.9
0.24
1.54
0.71
46466
Ma
y 26
324
53.8
0.56
-0.59
5.82
Table 3. The results of performance check in-situ and camera
calibration with LiDAR specific target points
calibration process was tested by using a limited number of
target points which were originally used for the validation of
the LiDAR accuracy. Figure 9 shows the LiDAR-specific
targets, made of trampolines with 2m diameter, and GPS-
positioned at cm-level accuracy.
(a) (b)
Figure 8. The geo-registered LiDAR intensity image and
orthoimages based on reference adjustment (a) and
direct georeferencing (b).
Figure 9. Target points 43142 (a) and 46466 (b)
The image coordinates of check points were measured and the
object coordinates were computed by combined intersection
(direct sensor orientation). The overall results of the in-situ
camera and boresight calibration performance check as a root
means square errors at target points are shown in Table 3. The
obtained root mean square errors at target points and mean
square error of intersection confirmed that the determined in-
situ camera and boresight calibration is optimal and stable, as
there is not change day to day and from one location to another.
5. CONCLUSIONS
The system calibration of multi sensor airborne systems,
including the boresight misalignment and geometric calibration
of the imaging sensor has vital importance for achieving
accurate geospatial data extraction performance. Any
discrepancies between the assumed mathematical model and
physical condition during the data acquisition can cause an
error in object space. Because of this, the determination of the
displacement vector and attitude difference between the camera
and IMU body frame (boresight misalignment) and geometric
calibration of camera are a critical issue for direct
georeferencing.
In this paper, in-situ sensor calibration was performed to
integrate a small format camera into a high-performance
LiDAR system. The investigation clearly showed that the
Redlake MS 4100 digital camera could be successfully
calibrated and boresighted to the LiDAR system, using only
control information derived from the LiDAR data. Using a
block of 21 images and 25 LiDAR-derived control points,
bundle block adjustments with various parameters were used to
perform an in-situ camera calibration. The focal length, the
radial symmetric lens distortions and systematic image errors
were accurately estimated (with respect to the camera quality).
In addition, the boresight misalignment was simultaneously
estimated using the BLUH and Applanix software products.
The effect of orientation discrepancies were checked by
computing y-parallaxes for each model at the reference block.
In addition, an independent performance check was performed
using LiDAR-specific ground targets. The combined results
clearly proved that the determined in-situ calibration parameters
were optimal and stable; in fact, the achieved accuracy appears
to be quite good compared to the camera quality.
ACKNOWLEDGEMENTS
The B4 project was funded by the National Science Foundation
and the ALTM 3100 system was generously provided by
Optech International. The authors wish to acknowledge the
support from the National Science Foundation and the
contribution of many researchers from OSU, USGS, NCALM,
UNAVCO and Optech International who participated in the B4
mapping project, as well as, the many OSU students who did
much of the GPS observations. The authors acknowledge the
financial support given by The Scientific and Technological
Council of Turkey (TUBITAK) provided for the post doctoral
research studies of Naci Yastikli at The Ohio State University.
REFERENCES
Cramer M., 2005. EuroSDR: Digital Camera Calibration and
Validation, Geolnformatic 2(8), March 2005, pp. 16-19 .