International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
Figure 12. Building region from 2D topographic map
Table 2. Accuracy assessment for building detection
Classified Data
Unit: pixel | Building | Non Building | Total
Building a=259685 | b-58176 e-317861
«| Non Building c=17505 d=954895 f=972460
Al Total g=277250 | h=1013071 1=1290321
9 Diagonal Total 1214580
5| Producer a/e=81% | d/f=98% 89.5%
| User a/g=93% | d/h=94% 93.5%
& Overall - - 94%
(a+d)/i
For each building region, the 3D planes are extracted by TIN-
based region growing. Meanwhile, initial building edges are
approximately detected in the DSM. The edges are then
projected into aerial image to mark the working area for straight
line detection. Through the Hough transform straight line
extraction, the precise straight lines which combine the 3D
planes can be projected into object space. After combining the
3D lines with SMS method, the building model can be
reconstructed. Comparing the coordinates roof corners in the
reconstructed models with the corners measured from stereo:
pairs, the root mean square errors are 0.45m, 0.56m, 0.70m in
the X, Y, Z directions, respectively. The results of building
models are shown in Figure 13.
Figure 13. 3D view of generated building model
6. CONCLUSIONS
In this investigation, we have presented a scheme for the
extraction of building regions and building modeling by
performing fusion of LIDAR data and optical imagery. The
results from the test show the potential of the automatic method
for building reconstruction. More than 81% buildings region
are correctly detected by our approach. The building models
generated by the proposed method have the merits of high
horizontal accuracy from aerial images and high vertical
accuracy from LIDAR data. Comparing the models
reconstructed by the proposed method with the reference data
from aerial stereo pairs, we achieved sub-meter accuracy.
However, in this investigation, we only consider flat roof
buildings. The improvements of the scheme for treating more
complex buildings are the major works in the future.
REFERENCES
Behan, A., 2000, On the matching accuracy rasterised scanning
laser altimeter data. IAPRS, Vol. XXXIII, Part B2, pp.75-82.
Briese, C., Pfeifer, N., and Dorninger, P., 2000, Application of
the Robust Interpolation for DTM Determination. IAPRS, vol.
XXXIILpp.55-61.
Canny, J., 1986, A computational approach to edge detection,
IEEE Transactions on Pattern Analysis and Machine
Intelligence, Vol. PAMI-8, No. 6, pp. 679-698.
Guo, T., 2003, 3D city modeling using high-resolution satellite
image and airborne laser scanning data. Doctoral dissertation,
Department of Civil Engineering, University of Tokyo, Tokyo.
Hofmann, A.D., Mass, H., Streilein, A., 2002, Knowledge-
based building detection based on laser scanner data and
topographic map information, IAPRS, Vol.34, Part 3A+B,
pp.163-169.
Hough, P.V.C. 1962, Methods and means for recognising
comples patterns, U.S. patent No. 306954
Lohmann, P., 2002, Segmentation and filtering of laser scanner
digital surface models, IAPRS, Vol. XXXIV, Part 2,, Xi'an,20-
23, Aug, 2002, pp311-316
736
Intern
Nakag
Sterec
3D Ut
Rau, .
buildi
Photo,
pp.18
Rotter
Buildi
Part 4,
Vosse
aerial
Geosc
Toron
Zhang
image:
filterir
Vol.54
