基于改进YOLOv10的全天候无人机航拍图像检测算法OA

In view of the single scene and low detection accuracy of the existing UAV aerial image detection algorithms,this paper proposes an all-weather UAV aerial image detection algorithm FEMFF-YOLO based on the improved YOLOv10.Firstly,an all-weather UAV aerial image dataset DNV(day-night-visible)is constructed to evaluate the robustness and generalization ability of the object detection model during multiple time periods and multiple weather conditions.Secondly,the original model backbone is replaced with the FasterNet structure that can extract features more efficiently.Finally,RepHMS,a re-parameterized heterogeneous multi-scale module,is used to enhance the multi-scale feature fusion ability of the model.Experiments show that the accuracy of FEMFF-YOLO algorithm is improved by 72.8%,its mean average precision mAP@0.5 reaches 63.8%,and its recall rate reaches 57.9%on the dataset DNV.In comparison with those of the basic YOLOv10 algorithm,its accuracy is increased by 3.9%,its mAP@0.5 by 3.3%,and its recall rate by 2.3%.The effectiveness of UAV aerial target detection is verified.