欢迎访问《图学学报》 分享到:

图学学报

• 图像与视频处理 • 上一篇    下一篇

基于改进 K 均值聚类算法的星点聚类研究

  

  1. 1. 郑州轻工业学院计算机与通信工程学院,河南 郑州 450000; 
    2. 格里菲斯大学工程信息技术学院,昆士兰 布里斯班 4000
  • 出版日期:2019-04-30 发布日期:2019-05-10
  • 基金资助:
    国家自然科学基金项目(81501547);河南省科技攻关项目(172102410080)

Star Point Clustering Based on Improved K-Means Clustering Algorithm

  1. 1. School of Computer and Communication Engineering, Zhengzhou University of Light Industry, Zhengzhou Henan 450000, China; 
    2. Faculty of Engineering and Information Technology, Griffith University, Brisbane Queensland 4000, Australia
  • Online:2019-04-30 Published:2019-05-10

摘要: 针对高分辨率天文图像中的星点聚类研究中存在的 2 个问题:①天文图像的分辨率 较高,且图像处理速度较慢;②选取何种聚类算法对天文图像中的星点进行聚类分析效果较好。 在研究中,问题 1 采用图像分块的方法提高图像的处理速度;问题 2 提出了一种改进的 K 均值聚 类算法,以解决传统的 K 均值聚类算法的聚类结果易受到 k 值和初始聚类中心随机选择影响的问 题。该算法首先在用 K 均值聚类算法对数据初步聚类的基础上确定合适的 k 值,其次用层次聚类 对数据聚类确定初始聚类中心,最后在此基础上再采用 K 均值聚类算法进行聚类。通过 MATLAB 仿真实验的结果表明,该算法的聚类结果与效率优于其他聚类算法。

关键词: k 值, 初始聚类中心, K 均值聚类算法, 层次聚类

Abstract: Two problems in the study of star point clustering in high resolution astronomical images: ① The resolution of the astronomical image is higher, and the image processing speed is slower. ② Which clustering algorithm is selected to cluster the star points in the astronomical image is better. In the research, problem 1 uses image segmentation method to improve image processing speed. problem 2 proposes an improved K-means clustering algorithm to solve the traditional K-means clustering algorithm clustering results are susceptible to k-value and The initial clustering center randomly selects the problem of impact. Firstly, the K-means clustering algorithm is used to determine the appropriate k-value based on the preliminary clustering of data. Secondly, the clustering is used to determine the initial clustering center by data clustering. Finally, K-means clustering is used. The algorithm performs clustering. The simulation results of MATLAB show that the clustering results and efficiency of the algorithm are better than other clustering algorithms.

Key words: k-value, initial cluster center, K-means clustering algorithm, hierarchical clustering