Weighted Joint Sparse Representation Hyperspectral Image Classification Based on Spatial-Spectral Dictionary

Shanxue Chen; Yufeng He

doi:10.3788/AOS220854

Acta Optica Sinica, Volume. 43, Issue 1, 0110002(2023)

Weighted Joint Sparse Representation Hyperspectral Image Classification Based on Spatial-Spectral Dictionary

Shanxue Chen^1,2 and Yufeng He^1,2、*

¹Chongqing Key Laboratory of Mobile Communications Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China

²School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(15)

Fig. 1. Schematic diagram of proposed hyperspectral image classification method

Download full size

Fig. 2. Weighted joint sparse representation

Download full size

Fig. 3. Weighted joint sparse representation hyperspectral image classification based on spatial-spectral dictionary

Download full size

Fig. 4. Superpixel segmentation of Indian Pines dataset

Download full size

Fig. 5. Superpixel segmentation of Salinas dataset

Download full size

Fig. 6. Classification results of Indian Pines dataset. (a) Standard classification result; (b) SRC; (c) KNN; (d) LSRC; (e) SP-JSRC;(f) HybridSN; (g) SSD-WJSRC

Download full size

Fig. 7. Classification results of Salinas dataset. (a) Standard classification result; (b) SRC; (c) KNN; (d) LSRC; (e) SP-JSRC;(f) HybridSN; (g) SSD-WJSRC

Download full size

Fig. 8. Overall accuracies at different balance factors

Download full size

Fig. 9. Effect of number of superpixels on classification result

Download full size

Fig. 10. Effect of sparsity on classification result

Download full size

Table 1. Number of training samples and test samples in Indian Pines dataset

View table

Table 1. Number of training samples and test samples in Indian Pines dataset

Label	Class	Number of training samples	Number of test samples
Total		1018	9231
1	Alfalfa	4	42
2	Corn-notill	142	1286
3	Corn-mintill	83	747
4	Corn	23	214
5	Grass-pasture	48	435
6	Grass-trees	73	657
7	Grass-pasture-mowed	2	26
8	Hay-windrowed	47	431
9	Oats	2	18
10	Soybean-notill	97	875
11	Soybean-mintill	245	2210
12	Soybean-clean	59	534
13	Wheat	20	185
14	Woods	126	1139
15	Buildings-Grass-Trees-Drives	38	348
16	Stone-Steel-Towers	9	84

Table 2. Number of training samples and test samples in Salinas dataset

View table

Table 2. Number of training samples and test samples in Salinas dataset

Label	Class	Number of training samples	Number of test samples
Total		533	53596
1	Brocoli_green_weeds_1	20	1989
2	Brocoli_green_weeds_2	37	3689
3	Fallow	19	1957
4	Fallow_rough_plow	13	1381
5	Fallow_smooth	26	2652
6	Stubble	39	3920
7	Celery	35	3544
8	Grapes_untrained	112	11159
9	Soil_vinyard_develop	62	6141
10	Corn_senesced_green_weeds	32	3246
11	Lettuce_romaine_4wk	10	1058
12	Lettuce_romaine_5wk	19	1908
13	Lettuce_romaine_6wk	9	907
14	Lettuce_romaine_7wk	10	1060
15	Vinyard_untrained	72	7196
16	Vinyard_vertical_trellis	18	1789

Table 3. Classification accuracies of Indian Pines dataset under different algorithms

View table

Table 3. Classification accuracies of Indian Pines dataset under different algorithms

Class	SRC	KNN	LSRC	SP-JSRC	HybridSN	SSD-WJSRC
Alfalfa	21.43	7.14	0	61.90	100.00	97.62
Corn-notill	47.98	53.58	77.29	96.58	95.47	96.89
Corn-mintill	51.81	40.43	56.49	92.77	97.24	97.19
Corn	29.44	12.62	51.40	90.65	99.46	93.46
Grass-pasture	82.07	80.92	88.74	93.79	94.93	99.08
Grass-trees	87.82	97.56	95.43	98.17	89.09	100.00
Grass-pasture-mowed	61.54	65.38	3.85	84.62	100.00	96.15
Hay-windrowed	95.13	98.84	100.00	100.00	100.00	100.00
Oats	22.22	27.78	0	100.00	88.89	100.00
Soybean-notill	68.00	72.34	68.23	93.14	93.54	97.83
Soybean-mintill	72.04	79.59	87.87	97.06	95.65	98.73
Soybean-clean	40.82	35.39	83.71	79.59	95.98	88.95
Wheat	88.65	95.14	97.84	89.73	97.69	99.46
Woods	87.18	92.45	98.51	97.98	93.21	99.47
Buildings-Grass-Trees-Drives	35.92	12.64	40.80	99.14	96.71	94.83
Stone-Steel-Towers	83.33	80.95	75.00	82.14	84.75	88.10
OA	67.13	69.16	80.87	94.90	94.98	97.60
AA	60.96	59.55	64.07	91.08	89.73	97.26
Kappa	62.53	64.84	78.19	94.18	94.26	96.73

Table 4. Classification accuracies of Salinas dataset under different algorithms

View table

Table 4. Classification accuracies of Salinas dataset under different algorithms

Class	SRC	KNN	LSRC	SPJSRC	HybridSN	SSD-WJSRC
Brocoli_green_weeds_1	97.49	94.822	99.40	91.65	96.88	99.70
Brocoli_green_weeds_2	98.62	97.235	99.89	100.00	100.00	99.76
Fallow	89.37	80.838	97.04	99.18	100.00	99.74
Fallow_rough_plow	95.51	99.566	96.38	90.88	97.31	98.84
Fallow_smooth	88.08	86.237	98.68	95.36	97.10	97.17
Stubble	99.82	98.240	99.46	99.87	99.15	99.67
Celery	99.63	98.787	99.72	99.41	99.14	99.46
Grapes_untrained	77.96	70.132	86.18	96.59	99.16	97.33
Soil_vinyard_develop	98.76	96.320	99.98	98.65	98.95	100.00
Corn_senesced_green_weeds	87.80	84.104	95.90	97.41	99.56	97.87
Lettuce_romaine_4wk	89.79	72.779	90.45	90.64	100.00	97.07
Lettuce_romaine_5wk	86.01	96.960	99.90	95.28	99.53	98.74
Lettuce_romaine_6wk	88.31	97.905	77.40	69.90	98.22	97.35
Lettuce_romaine_7wk	90.38	90.660	93.68	94.15	97.90	95.75
Vinyard_untrained	44.89	63.396	67.37	86.27	89.96	94.23
Vinyard_vertical_trellis	97.26	66.797	98.66	96.25	99.94	98.88
OA	84.63	83.52	91.39	95.18	98.08	98.01
AA	89.36	84.17	93.76	93.84	97.72	98.22
Kappa	82.89	81.66	90.42	94.63	97.46	97.78

Table 5. Classification accuracies of ablation experiments without spatial-spectral dictionary

View table

Table 5. Classification accuracies of ablation experiments without spatial-spectral dictionary

Label	Indian Pines	Salinas
1	95.24	99.10
2	89.66	99.78
3	93.71	95.40
4	89.72	98.33
5	97.93	99.21
6	99.85	99.80
7	92.31	99.58
8	100.00	94.97
9	94.44	99.76
10	94.97	91.74
11	97.01	98.02
12	91.01	97.69
13	98.92	98.68
14	99.47	93.02
15	90.23	91.79
16	97.62	97.71
OA	95.46	96.24
AA	95.13	97.12
Kappa	94.83	96.24

Tools

Get Citation

Copy Citation Text

Shanxue Chen, Yufeng He. Weighted Joint Sparse Representation Hyperspectral Image Classification Based on Spatial-Spectral Dictionary[J]. Acta Optica Sinica, 2023, 43(1): 0110002

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Image Processing

Received: Mar. 24, 2022

Accepted: Jun. 20, 2022

Published Online: Jan. 6, 2023

The Author Email: He Yufeng (846320689@qq.com)

DOI:10.3788/AOS220854

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology