The threshold can refer to different areas of the image: A global threshold uses the entire image for calculating the value while for a local threshold various image sections are defined and for each of them the calculation is executed individually. As the global threshold is very sensitive to changes of luminance it is mainly suitable for recordings with high contrasts like print samples or transmitted light images.
Furthermore there are different kinds of thresholds: fixed and dynamic thresholds. The latter should be used in situations with uneven illumination.
Thresholding has several steps: First the threshold is defined (e.g. by analysing the histogram). Next the grey value of each pixel is analysed and the pixels are grouped into segments and marked. In a simple process with only one threshold the pixels can, for example, be represented in white if they are above the threshold and in black if they are below the threshold, i.e. a binarization takes place. However it is possible to define several thresholds.
Based on analysis and marking it is easy to recognize which pixels belong to the object and which to the background.
Learn more about: