In global thresholding for Image Processing, we choose a single threshold value for the whole document. So its a good strategy to use when most part of the document have similar properties (as in one part isnt darker than others, for example). One of the most prominent method for determining global threshold is otsu threshold selection. Otsu’s method exhaustively searches for the threshold that minimizes the intra-class variance. The goal of thresholding is to divide the pixels of a given image into two classes Black and White (binarization). If we consider a gray scaled image as layers of gray scales, then we need to determine a level Gl of gray above which the pixels will be black, and the rest will be white.
Jun 12, 13