基于区域等分的改进K-Means算法研究
Study of improved K-Means algorithm based on equal region division
K-Means是对遥感图像在没有先验知识情况下进行无监督分类的重要算法之一，在遥感影像的分析中得到广泛的应用。但K-Means算法的随机性选择聚类中心的特点使得影像分类结果也存在一定的随机性。本文从提高聚类质量的角度出发，借鉴区域划分思想，结合传统的K-Means算法，提出了基于区域等分的K-Means改进算法。
K-Means is one of important algorithms of unsupervised classification without prior knowledge for remote sensing image,which has applied comprehensively in analysis of remote sensing image.But the character of selecting clustering center in randomness way makes there exists some randomness in the classification results.In this paper, base on the prospective of increasing clustering quality, referring the thought of region division,and combining the traditional K-Means algorithms.Then put forward improved K-Means algorithms based on equal region division.
陈国良、肖建铃
遥感技术
摄影测量与遥感K-Means划分算法
Photogrammetry and Remote SensingK-Meansclassificationalgorithm
陈国良,肖建铃.基于区域等分的改进K-Means算法研究[EB/OL].(2012-07-30)[2025-05-17].http://www.paper.edu.cn/releasepaper/content/201207-312.点此复制
评论