Research Article
Environmental Fault Diagnosis of Solar Panels Using Solar Thermal Images in Multiple Convolutional Neural Networks
Table 8
Statistics of thermal images in SN.
| | Class | Accuracy | Specificity | Sensitivity | Precision | F1 score |
| | Normal panel | 0.98 | 0.971 | 1 | 0.938 | 0.968 | | Completely soil covered | 1 | 1 | 1 | 1 | 1 | | Partially soil covered | 1 | 1 | 1 | 1 | 1 | | Temperature rise | 0.98 | 1 | 0.857 | 1 | 0.923 | | Fully shaded | 1 | 1 | 1 | 1 | 1 | | Partially shaded | 1 | 1 | 1 | 1 | 1 | | Average | 0.993 | 0.995 | 0.976 | 0.990 | 0.982 |
|
|
