Table 8 Comparison of the EC-EY and other algorithms.
From: A novel edge crop method and enhanced YOLOv5 for efficient wind turbine blade damage detection
Method | mAP0.5[%] | mAP0.5[%] | Model size (M) | FPS | ||||
|---|---|---|---|---|---|---|---|---|
Crackle | Lacquer peels off | Lightning strike | Scratch | Stains | ||||
Faster R-CNN | 30.6 | 31.6 | 65 | 67.8 | 68.8 | 52.7 | 108.3 | 12 |
SSD | 30.4 | 41.2 | 74.8 | 78 | 70.1 | 58.9 | 92.6 | 31 |
YOLOX | 49.7 | 32.2 | 99.5 | 90.9 | 67.4 | 68.0 | 68.52 | 21 |
YOLOv3-tiny | 49.7 | 36.6 | 99.5 | 87.5 | 76.9 | 70.0 | 17.4 | 150 |
YOLOv5s | 53.1 | 48.8 | 99.5 | 90.9 | 74.8 | 73.4 | 13.8 | 128 |
YOLOv8s | 39.1 | 35.4 | 99.5 | 99.5 | 78.2 | 70.3 | 21.4 | 125 |
74.1 | 47 | 99.5 | 99.5 | 73.9 | 78.8 | – | – | |
RT-DETR-L | 85.7 | 46.5 | 99.5 | 95.5 | 80.6 | 81.6 | 66.2 | 53 |
EC-EY | 88.0 | 79.8 | 99.5 | 99.5 | 84.8 | 90.3 | 21.5 | 82 |
