基于LFDA和GA-ELM的高光谱图像地物识别方法研究

李宝芸,范玉刚,杨明莉

Research on Feature Recognition Method of Hyperspectral Image based on LFDA and GA-ELM

Baoyun Li,Yugang Fan,Mingli Yang

表2 Pavia University数据集实验结果

Table 2 Experimental results of Pavia University dataset

降维方式	分类方式	AA/%	OA/%	Kappa系数
不进行降维	SVM	96.17	95.56	0.965
	ELM	53.48	53.92	0.545
	GA-ELM	86.37	86.67	0.862
PCA	SVM	94.37	93.67	0.945
	ELM	92.03	91.04	0.922
	GA-ELM	95.89	94.13	0.944
LDA	SVM	93.33	96.44	0.945
	ELM	93.12	95.11	0.947
	GA-ELM	95.92	95.42	0.956
LFDA	SVM	94.50	96.60	0.930
	ELM	95.56	96.00	0.946
	GA-ELM	96.23	97.11	0.964