pbinarization

Synopsis

Description

pbinarization classifies pixels of the input image im_in into 2 clusters. Pixels values that are lower than low or greater than high are set to 0; other values are set to 255:

   if im_in[p] ≥ low and im_in[p] ≤ high
   then im_out[p]=255;
   else im_out[p]=0;

If high is lower than low then pbinarization performs an inverse thresholding:

   if im_in[p] < high or im_in[p] > low
   then im_out[p]=255;
   else im_out[p]=0;

For multispectral and color images, the threshold is compared with the min or max pixel value of each band. For graph, pbinarization operates on graph nodes.

Parameters

• low and high specify the gray level bounds. Values are in the gray level domain of the input image (eg. [0..255] for byte image, [-2147483648..+2147483648] for long integer image).
  
  If high is lower than low then pbinarization performs an inverse thresholding.
  If high is greater than the maximum gray level then high is set with the maximum value (respectively for low).

Inputs

• im_in: an image or a graph.

Outputs

• im_out: an image of bytes (Img2duc, Img3duc) or a graph.

Result

Returns SUCCESS or FAILURE.

Examples

Segments the tangram.pan image in 2 classes, background and tangram pieces:

   pbinarization 100 255 examples/tangram.pan out.pan

• Same result with a high bound greater than the maximum value:

     pbinarization 100 1e30 examples/tangram.pan out.pan
  

• Same result with a high bound lower than the low bound:

     pbinarization 100 -1 examples/tangram.pan out.pan
  

See also

C++ prototype