改进的K均值聚类算法彩色图像分割的研究

针对传统K均值聚类算法在彩色图像分割中受K值和初始聚类中心影响较大等问题。在基于图像子块划分的基础上给出了一种k值和初始聚类中心确定方法,并用区域生长算法对聚类后的子块进行块后处理,利用提出的算法对多幅自然图像进行了分割实验,并与相似的分割方法进行了比较实验,给出了详细的实验结果与分析。实验表明该方法分割速度快,效果好,具有较高的实用价值。

Color Image Segmentation Based on an Improved K-means Clustering Algorithm

For traditional k-means clustering algorithm for color image segmentation under the influence of k-value and the initial cluster centers in larger,and so on.On the basis of this article based on image block division gives a method to determine the k value and the initial cluster centers and region growing algorithm for clustering of child block after block,made using natural image segmentation algorithm on a number of experiments,and compared with the similar segmentation experiments,give details of experimental results and analysis.