面向光伏电站建设的移动端人体跌倒检测方法
|
|
李彬彬,张超,覃涛,陈昌盛,刘兴艳,杨靖
|
Mobile-based human fall detection method for photovoltaic power plant construction
|
|
Binbin LI,Chao ZHANG,Tao QIN,Changsheng CHEN,Xingyan LIU,Jing YANG
|
|
| 表 4 各模型的整体性能测试结果 |
| Tab.4 Overall performance test result of each model |
|
| 模型 | mAP50/% | P/% | R/% | NP/106 | FLOPs/109 | F/(帧·s−1) | | Nanodet | 76.8 | 78.5 | 58.2 | 0.93 | 1.4 | 140.3 | | YOLOv5 | 85.5 | 87.8 | 79.3 | 2.50 | 7.1 | 82.1 | | YOLOv8n | 85.7 | 86.9 | 80.8 | 3.00 | 8.1 | 88.0 | | YOLOv9t | 85.7 | 86.3 | 78.1 | 1.73 | 6.4 | 59.2 | | YOLOv11n | 85.1 | 87.7 | 81.4 | 2.58 | 6.3 | 86.2 | | YOLOv8-D | 86.3 | 87.7 | 81.9 | 2.70 | 7.4 | 97.6 | | YOLOv8-C | 86.0 | 82.3 | 82.7 | 1.96 | 6.6 | 100.7 | | YOLOv8-CD | 86.6 | 86.4 | 81.8 | 1.78 | 6.1 | 98.6 | | YOLOv8-CM | 86.6 | 86.7 | 83.0 | 2.24 | 6.8 | 91.6 | | CMD-YOLO | 88.6 | 87.8 | 84.5 | 2.06 | 6.3 | 101.9 |
|
|
|