2025/07/01 by Hongxing Peng, L Lin-Lin Chen, Peng, Hongxing +5 · 1 citation
Computer Science · Engineering · #Advanced Neural Network Applications #Aerial image #Aerial imagery #Boosting (machine learning) #Bottleneck #Computer Vision and Pattern Recognition (cs.CV) #Discriminative model #FOS: Computer and information sciences #I.2.10 #I.4.8 #I.5.1 #Object detection #Prior probability #Robustness (evolution) #Transformer #UAV Applications and Optimization #Video Surveillance and Tracking Methods
paper · pdf · doi:10.48550/arxiv.2507.00825
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2025/07/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
Object detection in Unmanned Aerial Vehicle (UAV) imagery is fundamentally challenged by a prevalence of small, densely packed, and occluded objects within cluttered backgrounds. Conventional detectors struggle with this domain, as they rely on hand-crafted components like pre-defined anchors and heuristic-based Non-Maximum Suppression (NMS), creating a well-known performance bottleneck in dense scenes. Even recent end-to-end frameworks have not been purpose-built to overcome these specific aerial challenges, resulting in a persistent performance gap. To bridge this gap, we introduce HEDS-DETR, a holistically enhanced real-time Detection Transformer tailored for aerial scenes. Our framework features three key innovations. First, we propose a novel High-Frequency Enhanced Semantics Network (HFESNet) backbone, which yields highly discriminative features by preserving critical high-frequency details alongside robust semantic context. Second, our Efficient Small Object Pyramid (ESOP) counteracts information loss by efficiently fusing high-resolution features, significantly boosting small object detection. Finally, we enhance decoder stability and localization precision with two synergistic components: Selective Query Recollection (SQR) and Geometry-Aware Positional Encoding (GAPE), which stabilize optimization and provide explicit spatial priors for dense object arrangements. On the VisDrone dataset, HEDS-DETR achieves a +3.8% AP and +5.1% AP50 gain over its baseline while reducing parameters by 4M and maintaining real-time speeds. This demonstrates a highly competitive accuracy-efficiency balance, especially for detecting dense and small objects in aerial scenes.