Table 4 Results of classification models for healthy and diseased leaves using fused feature data
From: Intelligent identification on cotton verticillium wilt based on spectral and image feature fusion
Feature band extraction algorithm | Fusion method | SVM classification accuracy (%) | BPNN classification accuracy (%) | ||
|---|---|---|---|---|---|
Training set | Test set | Training set | Test set | ||
PCA | Original data-FF | 97.80 | 97.00 | 97.60 | 97.19 |
SG-MSC-FF | 98.00 | 97.59 | 99.20 | 98.59 | |
SG-DT-FF | 97.19 | 96.79 | 98.60 | 97.39 | |
SG-MN-FF | 98.20 | 97.99 | 98.59 | 98.39 | |
SPA | Original data-FF | 96.99 | 96.59 | 98.60 | 97.79 |
SG-MSC-FF | 98.20 | 97.60 | 98.80 | 98.19 | |
SG-DT-FF | 96.59 | 96.19 | 97.40 | 96.59 | |
SG-MN-FF | 98.39 | 97.79 | 99.00 | 98.99 | |