Digital image processing technique is an important part of computer vision, and is the base for the image recognition, including noise filtering, pixel point interpolation, image enhancement, image segmentation and so on. Image segmentation is the key part of digital image processing. Thousands of methods have been put forward. By now, there is no one method that is fit for all images.This thesis was mainly involved in the research of image segmentation, and proposed some algorithm and application for image segmentation. The main contributions include:1. Review the development of image segmentation algorithm researches, put these methods into four parts as the basic segmentation principle: on the base of threshold, on the base of edge detection, on the base of region, on the base of using particular tool.2. On the base of the Fisher criterion image segmentation method, in order to put the image into three parts, put forward the double thresh image segmentation method, and put forward the many thresh image segmentation method.3. Put forward the Fisher criterion 2D projection image segmentation based on 2D histogram, and more, put forward the Fisher criterion image segmentation based on 3D histogram. Compared with the Fisher criterion lineal image segmentation method based on 2D histogram, the new method is better in image segmentation for gaussian noise image.4. Summarize the cell image segmentation methods. Apply the Fisher criterion image segmentation method based on 3D histogram into the blood cell image. The method is good for the gaussian noise blood cell image.
|