Existing object detection models struggle to effectively capture cross-scale contextual information

The gap

Evidence profile

Sourced from the stated research gap and limitations section of the source papers, classified as general, drawn from work published between 2025 and 2026, spanning 2 journals. Those papers have been cited 32 times in total.

Supporting evidence — 4 representative gaps

• WCDB-YOLO: Wavelet-Enhanced Contextual Dual-Backbone Network for Small Object Detection in UAV Aerial Imagery (2026) · Drones · cited 2× · doi

  Existing object detection models struggle to effectively capture cross-scale contextual information and multi-level semantic representations of objects. General-purpose object detection models are not suitable for aerial imagery due to its distinctive characteristics. There is a need for a more accurate and robust solution for small object detection from UAV platforms.

  generalstated research gapevidence 5/5
  Keywords: existing object detection models struggle effectively capture cross-scale
• The Survey of Evolutionary Deep Learning-Based UAV Intelligent Power Inspection (2026) · Drones · cited 5× · doi

  Mainstream deep learning models have complex structures and require expert domain knowledge. The images collected during the power inspection process by UAVs have problems such as complex backgrounds, uneven lighting, and significant differences in object sizes. There is a need for a technology that can independently design lightweight and high-precision deep learning models.

  generalstated research gapevidence 5/5
  Keywords: mainstream deep learning models have complex structures require
• RTUAV-YOLO: A Family of Efficient and Lightweight Models for Real-Time Object Detection in UAV Aerial Imagery (2025) · Sensors · cited 10× · doi

  Multi-scale and small objects detection deficiencies: Successive downsampling operations in deep networks result in substantial information loss for diminutive objects and inadequate exploitation of fine-grained shallow features. Background interference susceptibility: Conventional feature extraction mechanisms and detection head architectures exhibit limited robustness against the heterogeneous backgrounds characteristic of aerial imagery. Computational resource incompatibility: High-precision deep architectures impose substantial computational burdens that exceed the processing capabilities of resource-constrained UAV platforms, thereby precluding actual deployment.

  generallimitations sectionevidence 5/5
  Keywords: multi-scale small objects detection deficiencies successive downsampling operations
• I-YOLOv11n: A Lightweight and Efficient Small Target Detection Framework for UAV Aerial Images (2025) · Sensors · cited 15× · doi

  The detection of small objects in aerial imagery remains challenging due to complex backgrounds, congested multi-object distribution, blurred object boundaries, and drastic scale variations. The traditional single-stage model faces challenges in detecting small targets with UAVs, including feature weakening and IoU's excessive sensitivity. The model size and computational resources are limited on UAV platforms.

  generallimitations sectionevidence 5/5
  Keywords: detection small objects aerial imagery remains challenging due

Questions about this gap