K均值算法实现遥感图像的非监督分类

作者：包 健，厉小润 日期：2008-04-17/span> 浏览：3597 查看PDF文档

K均值算法实现遥感图像的非监督分类
包 健，厉小润
（浙江大学 电气工程学院，浙江 杭州 310027）

摘 要：K均值算法在高光谱遥感影像的非监督分类中具有较强的实用性，表现出了良好的优点。首先采用了最大最小选心法确定初始类别中心，然后使用了K均值算法实现遥感影像的分类。在分类过程中采用了VC++2005作为开发平台，极大地提高了遥感影像的分类速度，同时还给出了实现K均值分类主要步骤的代码。最后在深入分析不同迭代次数下得到的不同分类图的基础上，研究了迭代次数值对最后分类结果的影响.
关键词：非监督分类;K均值算法;VC++2005

Unsupervised classification of remote images using Kmean algorithm
BAO Jian, LI Xiaorun
(College of Electrical Engineering, Zhejiang University, Hangzhou 310027, China)

Abstract: Kmean algorithm shows great utility and a lot of advantage in the area of unsupervised classification of remote images. The maximumminimum choosing center method was used to choose the clustering center, then the unsupervised classification of remote images was realized. At that step, choosing VC++2005 as the developing platform, the classification time was greatly decreased and the main procedure code for the realization of classification was given out. At last, the influence of iterating number in classified images was found by looking for the difference among different classified images which was got by using different iterating number.
Key words: unsupervised classification; Kmean algorithm; VC++2005

参考文献(Reference):
［1］张良培.高光谱遥感［M］.武汉:武汉大学出版社,2005.
［2］汤国安.遥感数字图象处理［M］.北京.科学出版社，2002.
［3］BEZDEK J C. Pattern Recognition with Fuzzy Objective Function Algorithms［M］. New York: Plenum Press,1981.
［4］KRUGLINSKI D J. Visual C++技术内幕：第4版［M］.北京:清华大学出版社，1998.
［5］PROSISE J. MFC Windows程序设计：第2版［M］.北京:清华大学出版社，1996.
［6］CONGALTON G. A review of accessing the accuracy of classifications of remotely sensed data［J］. Remote Sensing of Environment,1991,3（7）:35-46.