Brain is a complex structure of human body .In brain there may be abnormalities such as tumor, hemorrhage etc. There are different types of modalities available to identify them like .CT, MRI, PET. In CT scan images we can analyze the traumatic brain injuries due to better contrast, low cost, better segmentation and abnormality location in brain. With CT image, tumor location can be identified using morphological operations and watershed algorithm . In this work, to identify brain tumor, preprocessing is done i.e converting to gray scale , for finding abnormality region histogram is applied. Then morphological operation are performed and watershed algorithm is applied to segment the image. After segmentation process we will get different regions in image, to identify all region correctly edge detection is used. From edge detection region boundary can be identified .There are different edge detection techniques such as Canny edge detection, Robert edge detection , Prewit edge detection, in this project sobel edge detection is used. Finally the segmented region will be sorted and compared for the segmented region will be sorted and compare for identifying abnormality. On the abnormality region ellipse is drawn. The different feature used to for identification are sobel edge detection The main intension of this project to identify the tumor location accurately and if small abnormality region are there then it will identify correctly. The small abnormal region its called here as binary holes.