Skip to main content

Table 2 Information classes, number of training samples, and classification accuracy for the PaviaU

From: A multiscale modified minimum spanning forest method for spatial-spectral hyperspectral images classification

 

No. of samples

No. of training samples

SVM

SSEPF-MMSF

MSEPF-MMSF

Asphalt

6631

548

82.17

99.14

99.2

Meadows

18,649

540

87.25

97.28

99.23

Gravel

2099

392

81.23

95.95

95.43

Trees

3064

524

93.21

98.66

96.28

Metal sheets

1345

265

99.78

99.93

100

Bare soil

5029

532

91.05

99.54

99.8

Bitumen

1330

375

89.4

99.25

100

Bricks

3682

514

85.17

97.66

98.18

Shadows

947

231

99.89

99.79

97.36

AA (%)

–

–

89.9

98.57

98.36

OA (%)

–

–

87.6

99.18

98.8

Kappa coefficient (%)

–

–

85.23

98.81

98.41