Research Article
A Two-Stage Optimization Strategy for Fuzzy Object-Based Analysis Using Airborne LiDAR and High-Resolution Orthophotos for Urban Road Extraction
Table 6
Classification assessment (confusion matrix, per class accuracies, and total accuracies) on the test site.
| User class/sample | Tree | Road | Building | Grass | Bare | Shadow | Sum |
| Confusion matrix |
| Tree | 58 | 0 | 1 | 7 | 0 | 18 | 84 | Road | 3 | 54 | 1 | 0 | 2 | 3 | 63 | Building | 2 | 6 | 116 | 15 | 4 | 22 | 165 | Grass | 2 | 1 | 1 | 53 | 0 | 2 | 59 | Bare | 5 | 10 | 5 | 3 | 57 | 3 | 83 | Shadow | 0 | 0 | 0 | 1 | 0 | 37 | 38 | Sum | 70 | 71 | 124 | 79 | 63 | 85 | |
| Accuracy |
| Producer | 0.82 | 0.76 | 0.93 | 0.67 | 0.90 | 0.43 | | User | 0.69 | 0.85 | 0.70 | 0.89 | 0.68 | 0.97 | | KIA per class | 0.79 | 0.72 | 0.90 | 0.62 | 0.88 | 0.38 | |
| Overall accuracy (%) | 76 | Kappa index of agreement (KIA) | 0.70 |
|
|