ISPRS Commission III, Vol.34, Part 3A „Photogrammetric Computer Vision“, Graz, 2002
finding the “model keypoints”. If there is a large elevation
difference in elevation, which is the case with discrepancy
between two strips, more points are needed. Figure 10 shows
the effect on model keypoints before and after laser strip
adjustment.
Figure 10. “Model keypoints” in the Gävle project. In the left
image are model key points before and in the right after the
strip adjustment.
3.6 Summary of results
A summary of the results is presented in table 7. As one can
see, the standard error of unit weight presents
so [m] | Observation | Time [s] No of
density obs.
[n:th point]
Calibration 1 | 0.0465 5 1033.6 5 000
Calibration 2 | 0.0350 1 56.4 19 000
Svinesund 0.0992 100 2335.5 5 600
Toensberg | 0.1093 100 563.4 20 000
Gávle 0.0960 200 409.0 10 000
Table 7. List of result in practical test sights
The strip adjustment method used in this case depends on
measured elevation differences between strips and difference
towards known points. There are often many observations and
a few unknowns. In most cases this is a favourable situation
but in this case it can cause slow convergence. Assume there is
an area of flat terrain with a couple of ditches. Planimetric
discrepancies between laser strips will only cause discrepancy
in the position of the ditch. Assume that you make observations
in all laser points, most of them will show no elevation
difference while just a few, e.g. 1-5 % will show elevation
discrepancies. This means that you only wish to make
observations where you have discrepancies, i.e. in the areas
with undulated terrain. Elevation and roll errors are on the
other hand easier to solve for if the observations are spread
evenly over the terrain.
4. SUMMARY AND CONCLUSIONS
Orientation errors cause systematic errors that in some cases
can be modelled and corrected for in strip adjustment. There is
however often a strong correlation between unknowns, which
limits the possibility to compensate for all errors in all cases.
In the tests presented in this paper, elevation differences and
roll offsets were easiest to solve for, as they only need
elevation difference measurements. Planimetry (X and Y),
pitch and heading are dependent on gradients in different
directions, i.e. undulated terrain. In addition to this, unknowns
are strongly correlated and needs certain fly pattern and/or
control information to be solved.
Issues of future improvement in the laser strip adjustment
procedure can be derived from the practical tests:
: The procedure of selecting areas of interest for
matching should be improved — this will speed up the
convergence
Further investigation of how to solve for different
orientation unknowns should be made — this will
increase the reliability of the method
The error model should be extended to include
modelling of strip deformation (might be roll mirror
scale factor) — this can improve the result in many
cases
= More effort should be put to matching laser
reflectance intensity — this will add information in
flat areas and make it possible to use e.g. painted
crossroads as ground control
In all practical tests presented here there was an improvement
of the result by doing a strip adjustment. There are still
investigations to be made for further development of laser strip
adjustment. The method is necessary for laser data calibration
and accuracy verification.
5. ACKNOWLEDGEMENTS
Many thanks to TopEye AB, Fotonor AS and Terrasolid OY for
their cooperation, without their help there would not be any
paper. Thanks also to the National Land Survey for letting us
take part of their test.
6. REFERENCES
Axelsson, P., 2000. DEM Generation from Laser Scanner Data
using Adaptive TIN Models, IAPRS Amsterdam, 2000
Burman, H., 2000a. Adjustment of Laser Scanner Data for
Correction of Orientation Errors. IAPRS Vol. XXXIII
Amsterdam 2000.
Burman, H., 2000b. Calibration and Orientation of Airborne
Image and Laser Scanner Data Using GPS and INS. PhD-
Thesis, Photogrammetry Reports No 69, 2000.
Crombaghs, M.J.E., R. Brugelman, E.J. de Min, 2000. On the
Adjustment of Overlapping Strips of Laseraltimeter Height
Data. IAPRS Vol XXXIII. Part B3, Amsterdam 2000, pp 230-
237.
Maas, H-G, 2000. Least-Squares Matching with Airborne
Laser Scanner Data in a TIN Structure. IAPRS, Vol XXXIII,
Part B3, Amsterdam 2000, pp 548-555.
Maas, H-G, 2001. On the Use of Pulse Reflectance Data for
Laserscanner Strip Adjustment. IAPRS, Vol XXXIV
Annapolis, 2001.