Figure 1.
YOLOv5s network structure diagram
Figure 2.
Slicing operation
Figure 3.
FPN + PAN structure
Figure 4.
Dataset partial image
Figure 5.
Experimental flow chart
Figure 6.
Mosaic data enhancement
TABLE I.
Experimental environment configuration
| Parameter | Configuration |
|---|---|
| CPU | Intel(R) Xeon(R) Silver 4210R CPU @ 2.40GHz |
| GPU | RTX 3090 |
| Memory | 60G |
| Display memory | 24G |
| System environment | Ubuntu18.04 |
| Experimental platform | PyTorch1.6.0, Python3.8 |
| Accelerated environment | CUDA10.1 |
Figure 7.
Learning rate curve
Figure 8.
Experimental training process
TABLE II.
Test results of each model on the test set
| Model | Precision/% | Recall/% | mAP@0.5/% |
|---|---|---|---|
| YOLOv5s | 94.8 | 89 | 93.5 |
| YOLOv4 | 76.2 | 85.4 | 51.2 |
| YOLOv3 | 73.6 | 82.3 | 48.9 |
Figure 9.
Experimental test effect