With the rapid development of artificial intelligence, visible-light object detection, as an important part of computer vision technology, has been widely used in the unmanned aerial vehicle (UAV) reconnaissance field. Using deep learning technology to deeply explore object features in complex battlefield environments and low-quality images can effectively solve the difficulties and challenges of visible-light object detection in UAV reconnaissance scenario, and further improve the accuracy of visible-light object detection. Therefore, a comprehensive survey is conducted on UAV visible-light object detection methods based on deep learning. First, various challenges of UAV visible-light object detection are introduced, such as small scale, arbitrary orientation, high camouflage, and motion blur. Second, main public datasets for visible-light object detection and image restoration are described. Then, combined with various challenges faced by UAV visible-light object detection, the application, advantages and disadvantages of deep learning methods in UAV visible-light object detection are summarized. Finally, the future possible research direction for UAV visible light object detection is discussed.
A Survey of UAV Visible-Light Object Detection Based on Deep Learning
2024
Article (Journal)
Electronic Resource
Unknown
Metadata by DOAJ is licensed under CC BY-SA 1.0
A Survey of Deep Learning Object Detection Algorithm
IEEE | 2022
|Object Detection Based on Deep Learning and Context Information
German Aerospace Center (DLR) | 2016
|Research on ship object detection based on deep learning
SPIE | 2024
|