基于分类器组合的遥感影像分类
首发时间:2011-03-10
摘要:将多分类器集合应用土地覆盖分类,首先构建分类器集合,应用支持向量机、径向基神经网络、J48决策树、简单贝叶斯和REPtree分类等进行土地覆盖分类,然后利用Boosting、投票法、证据理论和模糊积分法等分类器集成方法,得到综合不同分类器输出的最终分类结果。试验表明,多分类器集成能够有效提高土地覆盖分类的精度,具有广泛的应用前景。
关键词: Boosting算法 多分类器集成 土地覆盖分类 遥感影像
For information in English, please click here
Remote Sensing Image Classification Based on the Classifier Ensemble
Abstract:In this paper, we applied multiple classifier ensemble method to land cover classification. Firstly, Support Vector Machine (SVM), Radius Basic Function Neural Network (RBFNN), J48 decision tree, Na?ve Bayes and Reptree were selected as the base classifiers. Then, some classifier ensemble approaches such as Boosting, Voting, D-S and fuzzy integral were chosen to combine the classification results derived from the base classifiers. The experiment results show that the ensemble learning methods can improve classification accuracy efficiently and has extensive application prospect.
Keywords: Boosting Multiple classifier ensembles land cover classification remote sensing
基金:
论文图表:
引用
No.4414036553912129****
同行评议
共计0人参与
勘误表
基于分类器组合的遥感影像分类
评论
全部评论0/1000