基于进化人工免疫网络的遥感影像分类算法  

Classification of remote sensing image based on evolutionary artificial immune network

作　　者：许仕珍[1] 吴云东[1] 

机构地区：[1]信息工程大学测绘学院,河南郑州450052

出　　处：《测绘工程》2009年第1期22-25,共4页Engineering of Surveying and Mapping

摘　　要：人工免疫网络是受免疫网络理论的启发建立的用于数据处理的一种人工免疫系统。同时免疫系统是进化的,通过不断调节系统内细胞的数量和种类适应外界环境。提出的进化人工免疫网络借鉴免疫网络和进化的思想,是用于遥感影像分类的监督算法。算法通过一个学习过程,得到能够表示训练数据特征的网络细胞,然后利用这些网络细胞进行分类。实验表明进化人工免疫网络是遥感影像分类的有效工具。Artificial immune network is one kind of artificial immune system for data processing taking inspiration from immune network. Meanwhile the immune network is an evolutionary system for that it can adjust the number and type of immune cells to adapt exoteric environment. The proposed evolutionary artificial immune network based on idea of immune network and evolution is a supervised classification algorithm for classification of remote sensing image. The algorithm is able to obtain ceils of network which can represent feature of training data and then use these cell to finish classification. The experiment result shows it is an effective tool for classification of remote sensing and it is a superior method when compared with some traditional methods for its high precision.

关 键 词：人工免疫系统 人工免疫网络 遥感影像分类 监督分类 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]