Segmentation is basically a process of pixel classification, the picture is segmented into subsets by assigning the individual pixel to classes. In this case when segmentation by thresholding it have to do classifying the pixels into " dark" and "light" classes, in an attempt to distinguish dark object from their light background.Similarly in edge detection pixels is classified into "edge" and "not edge" by thresholding the response of some difference operator that has high values when the rate of change of gray level is high.This thesis shows segmentation techniques base on gray level values, and some experiments process correspond to segmentation techniques.