摘要
针对提升高光谱遥感影像的分类表现,提出了基于EMAPs的高光谱遥感多分类器集成算法。该算法首先提取扩展多属性剖面(EMAPs)特征,然后选取极限学习机、协同表示分类器和支持向量机作为基分类器,基于提取的EMAPs特征参与集成分类。选取Purdue Campus和Indian Pines两组实验数据分析评价所提出算法的有效性,结果表明,与单分类器相比,基于EMAPs的多分类器集成算法可以取得更优异的分类表现。
In order to further improve the classification performance of hyperspectral remote sensing images,this paper proposes a hyperspectral remote sensing multi-classifier ensemble algorithm based on EMAPs.The algorithm first extracts Extended Multi-attribute Profile(EMAPs)features,selects extreme learning machine,collaborative representation classifier,and support vector machine as base classifiers,and participates in ensemble classification based on EMAPs features.By selecting the experimental data of Purdue Campus and Indian Pines to analyze and evaluate the proposed algorithm,the results show that the multi-classifier ensemble algorithm based on EMAPs can achieve better classification performance than the single classifier.
作者
虞瑶
沈泉飞
吴越
YU Yao;SHEN Quanfei;WU Yue(Basic Geographic Information Center of Jiangsu Province,Nanjing 210013,China)
出处
《测绘与空间地理信息》
2025年第2期170-173,共4页
Geomatics & Spatial Information Technology
关键词
EMAPs
极限学习机
协同表示分类器
支持向量机
高光谱影像分类
EMAPs
extreme learning machine
collaborative representation classification
support vector machine
hyperspectral image classification
