Discrimination and Thresholding

  • John C. Russ

Abstract

There have been several references in the preceding chapters to the use of brightness discrimination to select pixels belonging to features of interest. This is a widely used method of converting a grey scale image to a binary (black and white) one, illustrated in Figure 5-1. Discrimination with threshold values is much more efficient than any edge following or region growing method (as discussed in the previous chapter) because it works on the entire image at once. Hence the time required is fixed, regardless of the complexity of the image, and very short. Also, the resulting binary image is a pixel-based representation of features of interest, and is easier for most measurement operations than the boundary representation that results from the location and identification of edges.

Keywords

Entropy Graphite Sandstone Hexagonal Remotely Sense 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. D. S. Bright, E. B. Steel (1986) Bright-field image correction with various imageprocessing tools Microbeam Analysis 1986 (A.D. Romig, Jr., W.F. Chambers, eds.) San Francisco Press, 517-520Google Scholar
  2. D. S. Bright (1987) An object finder based on multiple thresholds, connectivity and internal structure Microbeam Analysis 1987 (R.H. Geiss, ed.) San Francisco Press 1987, 290-292Google Scholar
  3. K. R. Castleman (1979) Digital Image Processing, Prentice Hall, Englewood Cliffs, NJGoogle Scholar
  4. K. R. Castleman, J. Melnyk (1976) An Automated System for Chromosome Analysis: Final Report, Document 5040-30, Jet Propulsion Lab., Pasadena, CAGoogle Scholar
  5. R. M. Haralick, K. Shanmugam, I. Dinstein (1913) Textural Features for Image Classification IEEE Trans. Syst. Man. Cybern., SMC- 3 610–621CrossRefGoogle Scholar
  6. J. N. Kanpur, P. K. Sahoo, A. K. C. Wong (1985) A new method for grey-level picture thresholding using the entropy of the histogram CVGIP 29, 273–285Google Scholar
  7. J. G. Moik (1980) Digital Processing of Remotely Sensed Images NASA publication SP - 431, 277Google Scholar
  8. J. F. O’Callaghan (1974) Computing the Perceptual Boundaries of Dot Patterns Computer Graphics and Image Processing 3 # 2,141–162CrossRefGoogle Scholar
  9. T. Pavlidis (1982) Algorithms for Graphics and Image Processing, Computer Science Press, Rockville MDGoogle Scholar
  10. W. K. Pratt (1978) Digital Image Processing Wiley, New YorkGoogle Scholar
  11. J. Prewitt, M. Mendelsohn (1966) The Analysis of Cell Images Annals of the N.Y. Academy of Sciences 128,1035–1053CrossRefGoogle Scholar
  12. J. P. Rigaut (1988) Automated image segmentation by mathematical morphology and fractal geometry Journal of Microscopy 150 21–30CrossRefGoogle Scholar
  13. A. Rosenfeld (1979) Some experiments on variable thresholding Pattern Recognition 11, 191CrossRefGoogle Scholar
  14. A. Rosenfeld, A. C. Kak (1982) Digital Picture Processing Academic Press, LondonGoogle Scholar
  15. J. C. Russ, J. Ch. Russ (1984) Image processing in a general purpose microcomputer J. Microscopy 135, 89CrossRefGoogle Scholar
  16. J. C. Russ (1986) Practical Stereology Plenum Press, New YorkGoogle Scholar
  17. J. C. Russ, J. Ch. Russ (1988) Automatic discrimination of features in grey-scale images Journal of Microscopy 148 263–277CrossRefGoogle Scholar
  18. A. W. M. Smeulders, A. D. Beckers (1989) Accurate image measurement methods (applied to 3D length and distance measurements) Proc. 1st International Conf. on Confocal Microscopy, Academisch Medisch Centrum, AmsterdamGoogle Scholar
  19. R. J. Wall, A. Klinger, K. R. Castleman (1974) Analysis of Image Histograms, Proc 2nd Joint Int’l Conference on Patt. Recog., IEEE 74CH-0885-4C. 341-344Google Scholar
  20. J. Weszka (1978) A Survey of Threshold Selection Techniques Comp. Graph. & Image Proc. 7, 259–265CrossRefGoogle Scholar

Copyright information

© Plenum Press, New York 1990

Authors and Affiliations

  • John C. Russ
    • 1
  1. 1.North Carolina State UniversityRaleighUSA

Personalised recommendations