Remote Sensing Technology and Application

Fusion of Multiscale Low-rank Representation and Two Way Recursive Filtering for Hyperspectral Image Classification

Mei LU, Jiatian LI, Wen LI, Mihong HU, Jiaxin YANG

Remote Sensing Technology and Application, 2024, 39(2): 393-404. DOI: 10.11873/j.issn.1004-0323.2024.2.0393

数据集	光谱范围 /nm	空间分辨率/m	图像尺寸	波段数	类别数	带标签样本
Indian Pines	400~2 500	20×20	145×145	200	16	10 249
PaviaU	430~860	1.3×1.3	610×340	103	9	42 776
Salinas	400~2 500	3.7×3.7	512×217	204	16	54 129

数据集

光谱范围

/nm

空间分辨

率/m

图像尺寸

波段数

类别数

带标签样本

Indian Pines

400~2 500

20×20

145×145

200

10 249

PaviaU

430~860

1.3×1.3

610×340

103

42 776

Salinas

400~2 500

3.7×3.7

512×217

204

54 129

Table 1 Dataset Description

Other figure/table from this article

Fig.1 The flowchart to MSLRR_TWRF
Fig.2 Influence of parameter C on classification accuracy of MSLRR_TWRF method
Fig.3 Influence of parameter S_f on classification accuracy of MSLRR_TWRF method
Fig.4 Influence of parameter δ_r on classification accuracy of MSLRR_TWRF method
Table 2 Classification accuracy of different methods for Indian Pines
Table 3 Classification accuracy of different methods for PaviaU
Table 4 Classification accuracy of different methods for Salinas
Fig.5 Indian Pines image classification results obtained by different methods
Fig.6 PaviaU image classification results obtained by different methods
Fig.7 Salinas image classification results obtained by different methods
Table 5 Classification accuracy of different training samples for Indian Pines
Table 6 Classification accuracy of different training samples for PaviaU
Table 7 Classification accuracy of different training samples for Salinas
Table 8 Running time of different compared methods