摘要
遥感图像的分类方法包括统计模式识别、句法模式识别、以及神经网络、遗传算法、模拟退火算法等等。本文分析了统计模式识别方法的优缺点,提出了使用SVM的方法进行遥感图像分类的设想,通过实验证明该方法是有效和稳健的。
The traditional classification methods of remote sensing images include statistical pattern recognition, syntax pattern recognition, neural network, genetic algorithm, simulated anneal algorithm etc. This paper analyses the advantages and disadvantages of statistical pattern recognition and proposes a new algorithm . The algorithm is based on SVM( Support Vector Machine). The experiment proved that the algorithm is efficient and robust.
出处
《遥感信息》
CSCD
2003年第2期14-18,T001,共6页
Remote Sensing Information
