Sander Oude Elberink 
  
classification of buildings, discrimination can be made in several sorts of buildings by using the first and second height 
derivative. The unsupervised classification proved to be superior to supervised classification techniques, because the 
selection of homogeneous training areas proved to be a very time consuming activity, while the Supervise 
classification does not show improved results comparing to the unsupervised classification. 
5 CONCLUSIONS 
The achieved results show the potential of image processing techniques applied for the segmentation and classification 
of laser scanner data. In very dense laser scanner data the anisotropic height texture measure can be used for an 
accurate, automatic detection of trees. If a laser scanner system is able to register first and last pulse simultaneously, the 
detection of trees can be done by extracting the last pulse from the first pulse. At non-tree pixels one can easily dete 
buildings by its height in the normalised DSM. The poor quality of reflectance measurements results in poor 
classification results of roads, grasslands and agriculture fields. 
Obviously, the low-level vision techniques described here can only depict a first step in the procedure of 
segmentation of laser scanner data. Low- and high-level techniques using knowledge, e.g. on size and shape, on the 
objects to be extracted have to be used to improve the reliability of the results and get to products that can be used in 
GIS-related applications. 
ACKNOWLEDGEMENTS 
The authors would like to thank the Survey Department of Rijkswaterstaat in Delft/NL for providing the FLI-MAP laser 
scanner data used in this study, and Geodelta (Delft/NL) and Fotonor (Norway) for providing the Optech data set. 
REFERENCES 
Haala, N., 1999. Combining multiple data sources for urban data acquisition. Photogrammetric week, Wichmann 
Verlag, Heidelberg, Germany. 
Haala, N., Brenner, C., Anders, K.-H., 1998. 3D urban GIS from laser altimeter and 2D map data. IAPRS 32, pp. 339- 
346. 
Haralick, R., Shapiro, L., 1992. Computer and Robot Vision, volume 1. Addison-Wesley Publishing Company. 
Lemmens, M. Deijkers H., Looman, P., 1997. Building detection by fusing airborne laser-altimeter DEMs and 2D 
digital maps. [APRS 32, pp 29-42. 
Maas, H.-G., 1999. The potentional of height texture measures for the segmentation of airborne laser scanner data. 
Presented at the Fourth International Airborne Remote Sensing Conference and Exhibition / 21* Canadian Symposium 
on Remote Sensing, Ottawa, Ontario, Canada. 
Maas, H.-G., Vosselman G., 1999. Two algorithms for extracting building models from raw laser altimery data. ISPRS 
Journal of Photogrammetry and Remote Sensing 54, pp 153-163. 
Oude Elberink, S., 2000. The use of height texture measures for the classification of laser scanner data (in Dutch). Final 
thesis: Section of Photogrammetry and Remote Sensing, Department of Geodetic Engineering, Delft University of 
Technology. 
Weidner, U., Fórstner, W. 1995. Towards automatic building reconstruction from high-resolution digital elevation 
models. ISPRS Journal of Photogrammetry and Remote Sensing 50 (4), pp. 38-49. 
Zhang, Y. Optimisation of building detection in satellite images by combining multispectral classification and texture 
filtering. ISPRS Journal of Photogrammetry and Remote Sensing 54, pp 50-60. 
  
684 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000.