基于改进YOLOv5s的无人机小目标检测算法
宋耀莲,王粲,李大焱,刘欣怡

UAV small target detection algorithm based on improved YOLOv5s
Yaolian SONG,Can WANG,Dayan LI,Xinyi LIU
表 3 不同算法在Visdrone2019数据集上的检测精度与速度
Tab.3 Detection accuracy and speed of different algorithms on Visdrone2019 data set
模型mAP50/%mAP50:95/%FPS/帧
YOLOv3[11]20.412.231
YOLOv4[12]33.317.636
YOLOv5s34.118.8116
YOLOv6[21]33.818.448
YOLOv7[22]35.219.887
YOLOv839.422.182
Faster R-CNN[2]22.515.115
RetinaNet[4]30.118.5
本研究模型39.922.253