基于改进YOLOv8的船舶目标检测算法
|
|
朵琳,殷瑜,段威,张芸,任勇
|
Ship target detection algorithm based on improved YOLOv8
|
|
Lin DUO,Yu YIN,Wei DUAN,Yun ZHANG,Yong REN
|
|
| 表 4 HRSID 数据集不同目标检测模型的比较 |
| Tab.4 Comparison of different object detection models in HRSID dataset |
|
| 方法 | p/% | r/% | mAP50/% | Np/
106 | FLOPs/
109 | | RetianNet[21] | 88.8 | 77.5 | 78.2 | 60.1 | 181.9 | | CenterNet[22] | 84.6 | 93.1 | 84.5 | 55.1 | 175.4 | | DAPN[23] | 88.9 | 77.1 | 86.1 | 63,8 | 266.1 | | CoAM+RFIM[24] | 91.9 | 84.5 | 92.0 | 27.7 | 123.6 | | PPA-Net[25] | 93.4 | 89.1 | 92.0 | 72.2 | 38.8 | | Context-aware network[26] | 93.8 | 88.8 | 92.1 | 70.4 | 144.5 | | DD-YOLO | 92.0 | 85.0 | 92.2 | 13.9 | 9.8 |
|
|
|