基于距离相似性K-means的红外图像聚类算法研究

A K-means Clustering Algorithm Applied to the Infrared Images Based on Distance Similarity

提出了一种基于距离相似性K-means的红外图像聚类算法。该算法对通过Isomap算法降维后的空间点,进一步进行聚类;算法中引入了密度因素,通过距离相似性的差异进一步排除孤立点和选取初始聚类中心,使数据内部的紧凑性得到加强。经过实验证明,改进后的方法比原方法更有效,时间复杂度也大幅度降低。

An infrared image clustering algorithm based on K-means distance similarity was proposed. Firstly, it re-clusters the space points which are gained by Isomap dimension reduction algorithm. Secondly, by introducing density factor, some isolated points can be further eliminated and the initial clustering center can be selected by the difference of distance similarity, making the compactness within the data be strengthened. The experimental results show that the improved method is more effective and can also reduce the time complexity.