http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Chushi Yu(유초시),Yoan Shin(신요안) 대한전자공학회 2024 대한전자공학회 학술대회 Vol.2024 No.6
With the advances of drone technology, object detection using unmanned aerial vehicle (UAV) images has become an essential task. Detecting targets in UAV images pose challenges including varying size, shapes, occlusions, and lighting conditions. Despite significant progress with deep learning-based object detection, issues like missed detections and false alarms persist. We propose the method based on real-time detection transformer (RT-DETR) with dual convolutional kernels (DualConv) and context guided downsampling in UAV images. Experiments on the VisDrone dataset demonstrate our proposed method improves detection accuracy and capability in complex environments.