With the continuous development of UAV technology and target detection technology, Using UAVs to detect ground targets has become a current research hotspot. However, due to the altitude of the UAV, the size of the ground target in the aerial image of the UAV is small, and the feature information is not obvious, it is very easy to cause missed detection and false detection. Therefore, algorithms with higher precision are needed for UAVs. For this purpose, based on the YOLOv5 target detection algorithm, this paper improves the detection accuracy of small targets optimizing the prior anchor box, adding a small target detection layer, and adjusting the multi-scale feature fusion structure. Finally, experiments results verifies that the proposed method can effectively increase the detection accuracy for small targets to 10.4%.
An Improved YOLOv5-Based Small Target Detection Method for UAV Aerial Image
Lect. Notes Electrical Eng.
Chinese Intelligent Automation Conference ; 2023 ; Nanjing, China October 02, 2023 - October 05, 2023
Proceedings of 2023 Chinese Intelligent Automation Conference ; Kapitel : 30 ; 298-312
2023-09-23
15 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Research on Improved YOLOv5 Vehicle Target Detection Algorithm in Aerial Images
DOAJ | 2024
|STD-Yolov5: a ship-type detection model based on improved Yolov5
Taylor & Francis Verlag | 2024
|Unmanned aerial vehicle landing method based on improved YOLOv5 algorithm
Europäisches Patentamt | 2022
|Research on Object Detection and Recognition Method for UAV Aerial Images Based on Improved YOLOv5
DOAJ | 2023
|