| With the development of the disciplines and industries of science, technology, agriculture and medicine in recent decades, there are more and more signals and images needing for processing in production and our lives, and the amount of work is increasing rapidly . Nowadays ,digital image processing technology has been used in many fields of industry and agriculture,medical image processing technology included.The thesis discusses the problem of noise removing firstly. Although many image de-noising methods do well in reducing the influence of noise, some edge information is weakened as well, which is not beneficial to the work of edge-based image segmentation. Using nonnasality image information, the method of improving vector median filtering to preprocessing was carried out for image and its result was compared with Ideal Filter's .Secondly,The discrimination of object from background is the chief content of image segmentation; it is also the weight-bearing point in this article. Because of the inadequacy of traditional region-based image segmentation method in the field of nonnasality image segmentation, the paper build up a suit of theory and algorithm of edge detection and processing based on which the paper reaches the aim of preliminary image segmentation . The paper compares the algorithm of parallel regional segmentation with the algorithm of serial regional segmentation and gets thresholding image through adopting a best threshold value of image segmentation method .The influence of holes that seats in the center of the nonnasality to segmentation and measurement is well avoided.On the basis of inquiry and research on the medical image segment at home and abroad, and the author's study on image processing and pattern recognition, expatiates the method for thresholding segment using the best threshold value of image segmentation method and pattern character operation which will be help to computer repalacing man for finishing observation and analysis in nonnasality disease. |
PDF Full Text Request