Infrared Image Detection of Conveyor Belt Rollers Based on Improved YOLOv5

Table 1. Comparison results of two backbone networks
View table
Table 1. Comparison results of two backbone networks
Model Layer Precision /% Recall /% mAP /% FPS /（frame·s^-1）
YOLOv5 157 96.2 92.2 96.5 119
YOLOv5-198 198 94.4 94.4 97.4 100
YOLOv5-234 234 93.1 94.6 97.7 81

Table 2. Results of ablation experiments
View table
Table 2. Results of ablation experiments
Model Precision /% Recall /% mAP /% FPS /（frame·s^-1） Size /MB
YOLOv5 96.2 92.2 96.5 119 13.7
YOLOv5-W 95.2 93.6 97.6 123 13.7
YOLOv5-D 93.1 94.6 97.7 81 15.0
YOLOv5-DW 95.5 94.6 98.0 80 15.0

Table 3. Comparison results of different algorithms
View table
Table 3. Comparison results of different algorithms
Model Size /MB Precision /% Recall /% mAP /% FPS /（frame·s^-1）
YOLOv5 13.7 96.2 92.2 96.5 119
YOLOv7 71.3 96.3 93.1 97.8 24
YOLOv8s 127.0 92.8 95.1 98.1 74
Improved model 15.0 95.5 94.6 98.0 80

Tools

Get Citation

Copy Citation Text

Zhichao Zhou, Jianping Zhou, Xiaodong Yang, Xiaojing Wan, Binbin Gui. Infrared Image Detection of Conveyor Belt Rollers Based on Improved YOLOv5[J]. Laser & Optoelectronics Progress, 2025, 62(4): 0437006

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites