SICEM: A Generation Approach of Band Combination for Hyperspectral Imagery Reconstitution Based on Space and Information Analyses

Table 2

Classification accuracies (%) achieved using 40 selected bands on IndianP with KNN classifier.

Class	Algorithm
	MVPCA	WaLuDi	DBSCAN	I-FDPC	SICEM
Corn-notill	66.5	71.3	69.6	71	73.7
Corn-mintill	55.2	64.4	69.4	70.8	70.8
Grass-pasture	82.6	88.7	86.2	85	84.4
Grass-trees	93.1	94.5	95	94.3	96
Hay-windrowed	97.6	96.3	96.6	95.8	96.6
Soybean-notill	62.5	76.2	73.7	68.3	75.7
Soybean-mintill	77.4	79.3	81.5	82.6	84.2
Soybean-clean	57.1	71.6	74	76.2	76
Woods	93.6	94.4	93.3	93.5	95.8
Buildings-Grass-Trees-Drives	54.7	61.7	63.9	57.4	63.6
AA	74.1	79.8	80.3	79.5	81.7
OA	80.2	83.1	82.5	83.3	86
Variance	287.9	166.5	141.4	165.9	135.3