Research On Color Image Segmentation Algorithms Based On FCM

Image segmentation is a key step from image processing to image analysis,and its quality directly affects the subsequent processing results.For decades,scholars at home and abroad have been attaching great importance to the study of image segmentation and have proposed thousands of related algorithms.Among them,fuzzy C-means clustering algorithm(FCM)is a segmentation algorithm that combines fuzzy theory and cluster analysis.It not only can represent the fuzzy characteristics of pixel membership,but also can achieve unsupervised segmentation.Therefore,it is widely used in the field of image segmentation.This thesis studies the color image segmentation algorithm based on FCM.Firstly,the related theories and methods of image segmentation are summarized.Then,the research trends of fuzzy C-means clustering algorithm in image segmentation are analyzed in detail.To solve some problems in existing FCM algorithms,two new color image segmentation algorithms are proposed in this thesis.The main work is as follows:Aiming at the problems that the existing fuzzy C-means clustering(FCM)algorithm is sensitive to initial conditions and have poor anti-noise performance,an adaptive color image segmentation algorithm based on FCM clustering is proposed.Firstly,the thresholds of R,G and B histogram of color images are obtained by histogram thresholding method respectively.Then the number of clusters and the initial cluster centers are obtained by using the region splitting and merging method.Finally,the fuzzy C-mean clustering algorithm based on spatial neighborhood pixels is used to cluster the image,and the final segmentation result is obtained.Experiments show that the proposed algorithm has a good segmentation effect on color images.Compared with the existing similar algorithms,the segmentation effect and noise resistance performance are improved obviously.In order to solve the problems of high computational complexity and unsatisfactory segmentation effect in the traditional FCM clustering algorithms,a color image segmentation algorithm based on superpixel and FCM clustering is proposed in this thesis.Firstly,a simple linear iterative clustering(SLIC)algorithm is used to obtain the superpixels of the color image.Then the similarity between each superpixel and the surrounding superpixels is calculated according to the color,border,and spatial position relationship,and the color information is weighted to obtain the super-pixel comprehensive color information according to the similarity.Finally,the superpixels are clustered by the FCM algorithm to get the final segmentation results.Experimental results show that the proposed algorithm not only has better segmentation effect than existing similar algorithms,but also greatly reduces the computational complexity.