Visual Form pp 249-258 | Cite as

Fast Thinning Algorithms for Large Binary Images

  • Christian Evers

Abstract

Image resolutions of 100002 pixels or more are required if we scan large engineering drawings or maps to convert them to CAD-data. An important preprocessing step for the vectorization of the scanned lines is thinning them to minimal width for line-following. Most thinning algorithms are suitable for small resolutions only. We present two fast algorithms for thinning large binary images. The first one is a variation of an iterative parallel thinning algorithm. For most images it requires only one scan through the image matrix. The second method requires two forward scans through the image matrix, regardless of the thickness of the scanned lines.

Keywords

Extractor 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    C. Arcelli, G. Sanniti di Baja, “A Width-Independent Fast Thinning Algorithm”, Trans. PAMI, Vol. 7, No. 4, 1985, pp. 463–474.CrossRefGoogle Scholar
  2. [2]
    C. Arcelli, G. Sanniti di Baja, “A One-Pass Two-Operation Process to Detect the Skeletal Pixels on the 4-Distance Transform”, IEEE Trans. PAMI, Vol. 11, No. 4, 1989, pp. 411–414.CrossRefGoogle Scholar
  3. [3]
    C. Arcelli, G. Sanniti di Baja, “A Fast Algorithm to Convert Thin Digital Sets into Line-Drawings”, Proc. Int. Symp. Comp. Arch.and Dig. Sign. Proc., 1989, Hong Kong.Google Scholar
  4. [4]
    U. Eckhardt, G. Maderlechner, “Thinning of Binary Images”, Hamburger Beiträge zur Angewandten Mathematik, Universität Hamburg, 1989.Google Scholar
  5. [5]
    C. Evers, “Analyse von Exoskeletten”, Proc. 10. DAGM-Symp. Mustererkennung, Zürich, 1988, pp. 219-225.Google Scholar
  6. [6]
    C. Evers, “Thinning Binary Images with Two Forward Scans”, Internal Siemens Report ZFE IS INF 1-24/90/Ev, München, 1990.Google Scholar
  7. [7]
    M. Ejiri, S. Kakumoto, T. Miyatake, S. Shimada, K. Iwamura, “Automatic Recognition of Engeneering Drawings and Maps”, in Image Analysis Applications, Kasturi et al. (editors), Marcel Dekker Verlag, 1990, pp. 73-126.Google Scholar
  8. [8]
    J. Hofer-Alfeis, “Automated Conversion of Existing Mechanical-Engeneering Drawings to CAD-Data: State of the Art”, in Computer Applications in Production and Engineering, K. Bo, L. Estensen, P. Falster, E.A. Warman (editors), Elsevier Science Publ. B.V., North-Holland.Google Scholar
  9. [9]
    M.P. Martínez-Perez, J. Jiminéz, J.L. Navalón, “A Thinning Algorithm Based on Contours”, Comp. Vis., Graph. Im. Proc., Vol. 39, 1987, pp. 186–201.CrossRefGoogle Scholar
  10. [10]
    V. Nagasamy, N. Langrana, “Engeneering Drawing Processing and Vectorization System”, Comp.Vision Graph. Im. Proc, Vol. 49, 1990, pp. 379–397.CrossRefGoogle Scholar
  11. [11]
    T. Pavlidis, “An Asynchronous Thinning Algorithm”, Comp. Graph. Im. Proc, Vol. 20, 1982, pp. 133–157.MATHCrossRefGoogle Scholar
  12. [12]
    T. Pavlidis, “A Vectorizer and Feature Extractor for Document Recognition”, Comp. Vision, Graph. Im. Proc., Vol. 35, No. 1, 1986, pp. 111–127.CrossRefGoogle Scholar
  13. [13]
    P. T. Speck, “Übersetzung von Linien-und Flächenstrukturen in kombinatorisch-relationale Datenstrukturen zur automatischen Mustererkennung in Digitalbildern”, Dissertation, ETH Zürich, 1984.Google Scholar
  14. [14]
    S. Suzuki, T. Yamada, “MARIS: Map Recognition Input System”, Pattern Recognition, Vol. 23, No. 8, 1990, pp. 919–933.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 1992

Authors and Affiliations

  • Christian Evers
    • 1
  1. 1.Corporate Research and Development, ZFE IS INF 11Siemens AGMünchen 83Germany

Personalised recommendations