Research On Technology Of Image Segmentation And Its Application

Image segmentation is one of the most challenging tasks of low level fields of computer vision; it extracts the features, such as edges and regions. It converts the original image into a more compact form, which makes it possible to understand images. Results of image segmentation determine the quality of image analysis and understanding; it occupies an important place in the image engineering. Image segmentation algorithms are dependent on the applications’requirements, different application may segment an image based on distinctive regional consistency, so there is no general segmentation algorithm suitable for all applications, and there isn’t a common criteria for judging segmentation success. This paper analyzes the unique requirements of the computer aided irregular needling embroidery system and the3D stratum model reconstructing system, and points out the weakness of the existing image segmentation algorithms, and three image segmentation algorithms are proposed to solve these problems and generate better results for these two application. The main wok includes:(1) When applications require the clustering centers must be chosen from a specified color set the existing clustering algorithms can’t meet it. An image segmentation algorithm based on clustering is presented to solve this problem, which adopts a genetic algorithm to guarantee that the clustering centers are chosen from the specified color set; and a novel mutation operator which can adaptively adjust the mutation probabilities is proposed to avoid the prematurity of the genetic algorithm. This algorithm solves the problem of image segmentation and the best color subset selection for the computer aided irregular needling embroidery system.(2) Traditional segmentation algorithms based on color feature clustering cannot get desirable results when the image contains fuzzy boundaries and noises. A new image segmentation method is proposed to solve this problem, it adopts an interactive image segmentation algorithm based on Random Walks to improve the computing method of the membership functions of fuzzy C-mean(FCM) incorporating spatial information, it decreases the noise sensitivity of FCM, and can effectively segment the structure physical modeling images. This algorithm is applied to segment the stratum image containing colors mixture near the layers’boundaries and heavy noise, and it gets better results.(3) A semi-supervised image segmentation algorithm is presented, which exploits some label data to segment the stratum image better or faster. The algorithm is based on integration of semi-supervised fuzzy c-means clustering algorithms with random walks. It models the image’s color feature through semi-supervied c-means clustering algorithm(SSFCM) based label data, then it defines a reliability function based on the membership calculated by SSFCM, and the pixels are classified into two types that are considered as labeled and unlabeled pixels of Random Walks, then it performs Random Walks to classified the pixels more accurately. The algorithm not only reduces the noise sensitivity of SSFCM but also avoids cumbersome operations that the user labels the seed points of all objects for Random Walks.