PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 12, 2026Remote Sensing3 citationsOpen Access

A Dual-Branch Perception Network for High-Precision Oriented Object Detection in Remote Sensing

View Full Paper
QWQi WangWSWei Sun

Key Points

  • The research aims to enhance high-precision oriented object detection in remote sensing, focusing on small objects in complex backgrounds.
  • Proposal of MDCA-YOLO algorithm for oriented object detection.
  • Design of Dual-Branch Perception Module for long-range dependencies.
  • Integration of Multi-Adaptive Selection Fusion to improve feature response.
  • Development of CoordAttOBB detection head for better angle regression accuracy.
  • MDCA-YOLO outperforms YOLO11s in object detection tasks.
  • Achieved improvements of 2.5% in mAP50 and 2.7% in mAP50:95 metrics on DIOR-R dataset.
  • Demonstrated superior performance in detecting small objects under challenging conditions.

Abstract

With the rapid evolution of remote sensing earth observation technology, high-resolution object detection is crucial in military and civilian domains but faces challenges from expansive views and complex backgrounds. Small objects are particularly challenging due to their low pixel coverage, poor textures, and susceptibility to drastic illumination changes and background clutter. To address these problems, this paper proposes MDCA-YOLO for oriented object detection. A Dual-Branch Perception Module (DBPM) is designed utilizing a synergistic mechanism of large-kernel and strip convolutions to establish long-range dependencies, accurately capturing geometric features of tiny objects even in the absence of local details; Multi-Adaptive Selection Fusion (MASF) is proposed to address cross-scale feature loss by adaptively enhancing feature response while suppressing background noise; furthermore, a reconstructed decoupled detection head, CoordAttOBB, significantly improves angle regression accuracy while reducing complexity. Experimental results on the DIOR-R dataset show MDCA-YOLO surpasses YOLO11s, improving mAP50 and mAP50:95 by 2.5% and 2.7%, respectively, effectively proving the algorithm’s superiority in remote sensing tasks.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Wang et al. (2026) studied this question.

synapsesocial.com/papers/69b257cd96eeacc4fcec6d86https://doi.org/10.3390/rs18050839
Ask AI
Helpful
Bookmark
Share
View Full Paper