Performs thresholding on input images using Niblack's technique or some of the popular variations it inspired

dst = cv.niBlackThreshold(src, k)
dst = cv.niBlackThreshold(src, k, 'OptionName',optionValue, ...)

Input

• src Source 8-bit single-channel image.
• k The user-adjustable parameter used by Niblack and inspired techniques. For Niblack, this is normally a value between 0 and 1 (its absolute value) that is multiplied with the standard deviation and subtracted from the mean (mean + k * std).

Output

• dst Destination image of the same size and the same type as src.

Options

• MaxValue Non-zero value assigned to the pixels for which the condition is satisfied, used with the Binary and BinaryInv thresholding types. default 255
• Type Thresholding type, default 'Binary'. One of:
  • Binary dst(x,y) = (src(x,y) > thresh) ? maxValue : 0
  • BinaryInv dst(x,y) = (src(x,y) > thresh) ? 0 : maxValue
  • Trunc dst(x,y) = (src(x,y) > thresh) ? thresh : src(x,y)
  • ToZero dst(x,y) = (src(x,y) > thresh) ? src(x,y) : 0
  • ToZeroInv dst(x,y) = (src(x,y) > thresh) ? 0 : src(x,y)
• BlockSize Size of a pixel neighborhood that is used to calculate a threshold value for the pixel: 3, 5, 7, and so on. default 5
• Method Binarization method to use. By default, Niblack's technique is used. Other techniques can be specified:
  • Niblack Classic Niblack binarization. See [Niblack1985].
  • Sauvola Sauvola's technique. See [Sauvola1997].
  • Wolf Wolf's technique. See [Wolf2004].
  • Nick NICK technique. See [Khurshid2009].

The function transforms a grayscale image to a binary image according to the formulae:

• Binary

             | maxValue, if src(x,y) > T(x,y)
  dst(x,y) = |
             | 0, otherwise
  

• BinaryInv

             | 0, if src(x,y) > T(x,y)
  dst(x,y) = |
             | maxValue, otherwise
  

• Trunc

             | T(x,y), if src(x,y) > T(x,y)
  dst(x,y) = |
             | src(x,y), otherwise
  

• ToZero

             | src(x,y), if src(x,y) > T(x,y)
  dst(x,y) = |
             | 0, otherwise
  

• ToZeroInv

             | 0, if src(x,y) > T(x,y)
  dst(x,y) = |
             | src(x,y), otherwise
  

where T(x,y) is a local threshold calculated individually for each pixel.

The threshold value T(x,y) is determined based on the binarization method chosen. For classic Niblack, it is the mean minus k times standard deviation of BlockSize x BlockSize neighborhood of (x,y).

References

[Niblack1985]:

Wayne Niblack. "An introduction to digital image processing". Strandberg Publishing Company, 1985.

[Sauvola1997]:

Jaakko Sauvola, Tapio Seppanen, Sami Haapakoski, and Matti Pietikainen. "Adaptive document binarization". In Document Analysis and Recognition, Proceedings of the Fourth International Conference on, volume 1, pages 147-152. IEEE, 1997.

[Wolf2004]:

Christian Wolf and J-M Jolion. "Extraction and recognition of artificial text in multimedia documents". Pattern Analysis & Applications, 6(4):309-326, 2004.

[Khurshid2009]:

Khurram Khurshid, Imran Siddiqi, Claudie Faure, and Nicole Vincent. "Comparison of niblack inspired binarization methods for ancient documents". In IS&T/SPIE Electronic Imaging, pages 72470U-72470U. International Society for Optics and Photonics, 2009.