基于遗传聚类算法和小波变换特征的自动分类  

Auto-classification Based on Genetic Algorithm and Cluster Analysis and Orthogonal Wavelet Transform Feature

作　　者：李霆[1] 陈学佺[1] 邹晓涛[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,合肥230027

出　　处：《计算机工程》2003年第2期153-154,254,共3页Computer Engineering

摘　　要：研究了遥感影像的特点，提出了一种基于图像的正交小波变换构造特征，同时引入遗传聚类的分类方法，实现了遥感影像的自动分类。实验表明，将遗传聚类算法与小波特征相结合的分类方法具有自动、快速、自适应的优点，能克服传统分类方法中由于样本选择而产生的局限性，提高了识别的正确率。The problem of remote sensing texture images is studied by means of orthogonal wavelet transform features and genetic algorithm and cluster analysis and the images are classified with auto-classification .The experiment results show that automaticity,speediness and auto-adapting could be obtained in genetic cluster algorithm with orthogonal wavelet transform features. And this method overcomes the limitation of traditional classification by selecting samples and gets higher accuracy.

关 键 词：遗传聚类算法 小波变换 特征 自动分类 遥感影像 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]