Table 6. Assessment of deep learning models with respect to sampling intensity was conducted by comparing them with the Subdivision Land Cover Map. (Unit : %)
| Sampling intensity | OA | Kappa | F-score |
| Forest land | Cropland | Grassland | Wetlands | Settlements |
| S1 | 83.8 | 67.7 | 93.6 | 66.7 | 30.8 | 92.3 | 57.1 |
| S2 | 77.4 | 53.5 | 90.6 | 60.0 | 41.0 | 62.5 | 35.3 |
| S3 | 79.1 | 57.4 | 89.8 | 68.6 | 37.8 | 84.8 | 43.5 |
| S4 | 78.3 | 55.2 | 89.7 | 60.4 | 32.4 | 92.8 | 41.8 |
| S5 | 76.7 | 53.3 | 90.1 | 45.5 | 35.1 | 78.3 | 48.5 |
| S6 | 78.4 | 54.2 | 90.6 | 50.7 | 32.7 | 81.0 | 40.9 |
| S7 | 79.5 | 55.0 | 91.5 | 48.2 | 31.2 | 81.6 | 44.4 |