House and Street Network Detection in Residential Regions

  • Cem Ünsalan
  • Kim L. Boyer
Part of the Advances in Computer Vision and Pattern Recognition book series (ACVPR)


In the previous chapters, we introduced several methods to detect residential regions starting from land use classification. In this chapter, we introduce a novel subsystem (of our multispectral satellite image understanding system) to detect houses and the street network in residential regions. Detecting houses is far more challenging than detecting larger buildings for several reasons. First, their relatively small size (in one meter resolution Ikonos images) makes their detection difficult. Second, occlusion by nearby trees is common. Third, in some neighborhoods, houses may come in fairly complex shapes. Analogous problems (small cross-section, overhanging trees, and winding curves) present challenges for street detection in residential regions. Our house and street network detection subsystem comprises four main parts. We first introduce measures on multispectral images to detect regions of possible human activity. On these measures, we introduce a variation of the k-means clustering algorithm to extract possible houses and the street network by combining both spatial and spectral features. This combination of information improves the final clustering results. From clustering, we obtain a binary image containing possible street network fragments and houses. We then decompose this binary image using a balloon algorithm based on binary mathematical morphology. Having obtained the decomposition, we represent them in a graph-theoretical framework. Balloons serve as vertices, and their neighborhood information is encoded as edges in the graph. The street network is extracted from the graph by using the unary and binary constrains. The remaining vertices (balloons) are assigned as possible houses in the region.


False Alarm Street Network False Alarm Rate Street Segment Seed Point 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    T.E. Avery, G.L. Berlin,Fundamentals of Remote Sensing and Airphoto Interpretation, 5th edn. (Macmillan Publishing, Indianapolis, 1992) Google Scholar
  2. 2.
    R.O. Duda, P.E. Hart, D.G. Stork,Pattern Classification, 2nd edn. (Wiley Interscience, New York, 2001) MATHGoogle Scholar
  3. 3.
    C. Ünsalan, K.L. Boyer, Comput. Vis. Image Underst.98(3), 432 (2005) CrossRefGoogle Scholar
  4. 4.
    J.A. Hartigan,Clustering Algorithms (Wiley, New York, 1975) MATHGoogle Scholar
  5. 5.
    A.K. Jain, R.P.W. Duin, J. Mao, IEEE Trans. Pattern Anal. Mach. Intell.22(1), 4 (2000) CrossRefGoogle Scholar
  6. 6.
    N. Otsu, IEEE Trans. Syst. Man Cybern.9(1), 62 (1979) MathSciNetCrossRefGoogle Scholar
  7. 7.
    R. Adams, L. Bischof, IEEE Trans. Pattern Anal. Mach. Intell.16(6), 641 (1994) CrossRefGoogle Scholar
  8. 8.
    C. Godsil, G. Boyle,Algebraic Graph Theory (Springer, New York, 2001) MATHCrossRefGoogle Scholar
  9. 9.
    M. Kass, A. Witkin, D. Terzopoulos, Inter. J. Comput. Vis., 321–331 (1988) Google Scholar
  10. 10.
    R.C. Gonzales, R.E. Woods,Digital Image Processing (Addison Wesley, Mass., 1993) Google Scholar
  11. 11.
    R.A. Brualdi,Introductory Combinatorics, 3rd edn. (Prentice Hall, New Jersey, 1999) MATHGoogle Scholar
  12. 12.
    M.P. Do Carmo,Differential Geometry of Curves and Surfaces (Prentice Hall, New Jersey, 1976) MATHGoogle Scholar
  13. 13.
    C. Berge,The Theory of Graphs (Dover, New York, 2001) MATHGoogle Scholar
  14. 14.
    J.R. Evans, E. Minieka,Optimization Algorithms for Networks and Graphs, 2nd edn. (Dekker Inc., New York, 1992) Google Scholar
  15. 15.
    D.M. McKeown, J.L. Denlinger, inProceedings of CVPR1988 (1988), pp. 662–672 Google Scholar
  16. 16.
    C. Lin, R. Nevatia, Comput. Vis. Image Underst.72(2), 101 (1998) CrossRefGoogle Scholar
  17. 17.
    E. Riseman, X. Wang, B. Hanechak, H. Schultz, A. Hanson, inFifth International Conference on Computer Graphics and Visualization, 1997 Google Scholar

Copyright information

© Springer-Verlag London Limited 2011

Authors and Affiliations

  1. 1.Electrical and Electronics EngineeringYeditepe UniversityKayisdagiTurkey
  2. 2.Dept. Electrical, Comp. & Systems Eng.Rensselaer Polytechnic InstituteTroyUSA

Personalised recommendations