基于改进YOLOv8的露天矿区行车障碍物检测
顾清华,周琼,王丹
Traveling Obstacle Detection in Open-Pit Mine Area Based on Improved YOLOv8
Qinghua GU,Qiong ZHOU,Dan WANG
表3
损失函数改进后性能对比
Table 3
Performance comparison after loss function improvement
模型
计算量
/GFLOPs
参数量/M
mAP
50/%
白天
夜间
YOLOv8n（Baseline）
8.2
3.01
95.7
92.5
+WIoU
8.2
3.01
96.8
94.2
+2 C2fCA+WIoU
8.4
3.13
97.2
95.0
+new-Neck+WIoU
7.4
2.80
97.0
94.6