Table 3 Image fusion quantitative metrics for spatiotemporal problem for 12-bit MSI and OLI at two different acquisition times over different object classes for the Tehran dataset. No object class with the farmland label was detected in this area.
From: An object-based sparse representation model for spatiotemporal image fusion
Object class | Approach | RMSE | ERGAS | SAM | UIQI | PSNR | AG | SAMRG index | ||
|---|---|---|---|---|---|---|---|---|---|---|
G | R | RG | ||||||||
Impervious surface | Fit-FC | 92.619 | 2.601 | 25.96 | 0.8544 | 200.66 | 167.31 | 3.35 | 20.64 | 12.29 |
FSDAF | 81.236 | 2.159 | 21.45 | 0.9539 | 201.81 | 244.06 | 2.77 | 17.19 | 10.53 | |
STSR | 30.492 | 1.186 | 14.37 | 0.9896 | 210.32 | 208.25 | 1.88 | 7.38 | 3.71 | |
Farmland | Fit-FC | – | – | – | – | – | – | – | – | – |
FSDAF | – | – | – | – | – | – | – | – | – | |
STSR | – | – | – | – | – | – | – | – | – | |
Vegetated surface | Fit-FC | 19.568 | 0.307 | 5.36 | 0.9661 | 214.17 | 153.29 | 3.63 | 4.39 | 4.01 |
FSDAF | 23.755 | 0.372 | 9.23 | 0.9913 | 212.48 | 170.98 | 4.52 | 8.07 | 4.84 | |
STSR | 12.776 | 0.200 | 2.95 | 0.9983 | 217.87 | 162.73 | 2.03 | 3.10 | 2.59 | |
Water body and wetland | Fit-FC | 0.909 | 0.104 | 1.97 | 0.9904 | 240.83 | 92.95 | 0.74 | 1.68 | 1.27 |
FSDAF | 0.987 | 0.113 | 2.65 | 0.9972 | 240.11 | 119.06 | 1.52 | 2.23 | 1.95 | |
STSR | 0.813 | 0.093 | 1.44 | 0.9990 | 241.80 | 130.14 | 0.49 | 0.91 | 0.82 | |
Overall | Fit-FC | 37.699 | 1.004 | 11.10 | 0.9370 | 218.55 | 137.85 | 2.57 | 8.90 | 5.86 |
FSDAF | 35.326 | 0.881 | 11.11 | 0.9808 | 218.13 | 178.03 | 2.94 | 9.16 | 5.77 | |
STSR | 14.694 | 0.493 | 6.25 | 0.9956 | 223.33 | 167.04 | 1.47 | 3.80 | 2.37 | |
