Research on UAV Instrusion Dectection Based on Deep Learning
In view of the insufficient capability of existing detection system to detect UAV targets with low altitude,slow speed,low RCS and weak infrared radiation,and difficulties of detecting drones ef-fectively through images,the method based on YOLOv5 and ResNet50 models was proposed,and the deep learning method was used to detect and identify UAV images.Comparing with original YOLOv5model,accuracy,recall and mAP of identifying aerial targets in Drone-detection-dataset are higher based on the proposed model.Besides,comparing with Faster-RCNN and SSD,YOLOv5+Res-Net50 could reach better results of detecting UAV images,especially for distant UAV targets.As a re-sult,YOLOv5+ResNet50 can be used to detect UAV efficiently.
NETL
NSTL
万方数据
