Acta Photonica Sinica, Volume. 51, Issue 12, 1210001(2022)
Hyperspectral Image Denoising Based on Hybrid Space-spectral Total Variation and Double Domain Low-rank Constraint
Figures & Tables(11)
Fig. 2. Denoising results on band 50 of simulated data in noise case 1
Fig. 3. Denoising results on band 120 of simulated data in noise case 5
Fig. 4. Comparison diagram of PSNR and SSIM values of each band in the case of simulated data
Fig. 7. Band 109 of HYDICE dataset vertical mean profile before and after denoising via different methods
Fig. 8. Band 207 of HYDICE dataset vertical mean profile before and after denoising via different methods
Fig. 9. The change diagram of MPSNR and MSSIM value as the value of
Fig. 10. The change diagram of MPSNR and MSSIM value as the number of iterations increases
Table 1. Qualitative evaluation results of various methods on simulated data under different noise cases
View tableView in ArticleTable 1. Qualitative evaluation results of various methods on simulated data under different noise cases
Case Level Index Noise WNNM LRMR BM4D TDL WSNM LRTV LRTDTV LRHSSTV Case1 G=0.1 MPSNR/dB
MSSIM
ERGAS
20.002 7
0.3674
233.861 0
32.693 3
0.852 0
55.655 3
36.445 3
0.941 1
35.805 3
38.743 6
0.976 3
28.040 5
38.15
0.966 4
30.52
37.557 6
0.948 3
32.304 5
38.787 6
0.986 3
27.474 5
40.909 9
0.980 4
23.121 3
47.714 9 0.998 6 11.351 2 Case2 G=0.1
+ Deadline
MPSNR/dB
MSSIM
ERGAS
19.392 4
0.360 1
255.100 1
32.880 6
0.851 3
58.394 4
35.575 6
0.919 1
38.844 4
35.105 3
0.939 1
110.403 9
33.228 9
0.874 3
111.356 6
36.171 5
0.941 2
43.897 6
38.048 8
0.982 8
49.101 2
40.545 6
0.988 5
23.547 1
46.610 4 0.991 8 12.621 8 Case3 G=0.75
P=0.15
MPSNR/dB
MSSIM
ERGAS
13.090 6
0.178 3
519.682 4
32.365 3
0.878 6
59.712 3
36.403 3
0.934 5
36.046 6
28.596 6
0.840 1
90.298 3
27.504 5
0.907 6
102.085 5
38.146 4
0.955 1
29.523 7
39.533 7
0.986 6
35.025 4
41.080 9
0.991 0
21.883 3
46.428 9 0.997 9 13.834 7 Case4 G=0.75
P=0.15
+ Deadline
MPSNR/dB
MSSIM
ERGAS
12.923 9
0.175 5
529.5936
32.396 3
0.887 7
59.103 4
35.767 7
0.931 6
39.990 5
27.026 3
0.817 3
128.111 3
26.546 7
0.846 5
120.545 3
36.637 5
0.948 8
46.317 8
38.744 7
0.982 6
55.346 3
40.725 6
0.990 6
22.908 1
45.442 5 0.997 8 14.560 8 Case5 P+G
Random
+ Deadline
MPSNR/dB
MSSIM
ERGAS
13.810 5
0.204 4
500.610 2
31.432 2
0.845 5
66.423 1
33.725 4
0.895 1
50.161 3
27.947 3
0.820 1
120.954 7
24.345 7
0.593 1
157.757 5
35.027 9
0.928 5
51.334 5
36.543 3
0.974 2
72.525 7
38.835 5
0.985 9
28.663 4
43.567 2 0.996 9 18.339 4 Case6 P+G
Random
+ Deadline
+ Stripe
MPSNR/dB
MSSIM
ERGAS
13.736 9
0.202 1
503.698 6
29.9873
0.843 1
82.433 2
33.425 6
0.891 8
52.624 3
27.539 3
0.806 0
126.285 7
23.345 6
0.558 3
174.287 9
33.886 6
0.926 1
53.294 6
36.357 5
0.973 6
72.056 3
38.636 5
0.985 2
29.823 3
43.151 9 0.996 6 19.716 9
Tools
Get Citation
Copy Citation Text
Pengdan ZHANG, Jifeng NING. Hyperspectral Image Denoising Based on Hybrid Space-spectral Total Variation and Double Domain Low-rank Constraint[J]. Acta Photonica Sinica, 2022, 51(12): 1210001
Paper Information
Category:
Received: Mar. 11, 2022
Accepted: Jun. 7, 2022
Published Online: Feb. 6, 2023
The Author Email: NING Jifeng (njf@nwafu.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20