Edge Intelligence-Based OCS Fault Detection in Rail Transit Systems

Table 5

The results of 9 elements based on YOLOv5s.

Class	Labels			mAP@.5	@.5:.95
All	1317	92.6%	93.7%	94.5%	0.548
Guanmao	200	96%	95.3%	95.4%	0.513
Xuanzhuanshuanger	289	91.8%	94.8%	95%	0.463
Chenglisuozuo	97	94.9%	95.7%	97.4%	0.646
Dingweixianja	147	95.5%	95.9%	94.4%	0.409
Dingweizhizuo	97	97%	96.9%	98.5%	0.7
Ubaogu	58	97.5%	96.6%	97.2%	0.669
Jueyuanzi	185	87.9%	91.9%	93%	0.751
Wanbidizuo	163	82.6%	87.1%	87.2%	0.308
Diaoxuan	81	90%	88.9%	92%	0.472