SVM结合模糊方法在遥感图像分类中的应用

Application of Support Vector Machines and Fuzzy Method for Remote Sensing Image Classification

提出一种支持向量机(SVM)结合模糊方法的遥感图像分类算法。首先介绍了SVM基本算法及其在遥感图像分类中应用情况,然后针对SVM多类判别存在混分和漏分的缺陷,对混分和漏分样本采用模糊方法判决分类。实验证明该方法的分类精度优于单一的SVM方法、模糊方法或神经网络方法。

A mixed classify algorithm that support vector machines combine fuzzy method is proposed for remote sensing image classification.First,the basic algorithm of Support Vector Machines and its application in classification of remote sensing images is introduced.Then this paper aims at the case that the missed samples and the error-separate samples via multi-class SVM classifier,uses fuzzy method to classify these examples.Experiment results prove the accuracy of this method is improved than several methods such as using single SVM method,using single Fuzzy method and using single Artificial Neural Networks method.