A2RMNet: Adaptively Aspect Ratio Multi-Scale Network for object detection in remote sensing images

Heqian Qiu, Hongliang Li, Qingbo Wu, Fanman Meng, King Ngi Ngan, Hengcan Shi

Research output: Contribution to journalArticleResearchpeer-review

20 Citations (Scopus)

Abstract

Object detection is a significant and challenging problem in the study area of remote sensing and image analysis. However, most existing methods are easy to miss or incorrectly locate objects due to the various sizes and aspect ratios of objects. In this paper, we propose a novel end-to-end Adaptively Aspect Ratio Multi-Scale Network (A2RMNet) to solve this problem. On the one hand, we design a multi-scale feature gate fusion network to adaptively integrate the multi-scale features of objects. This network is composed of gate fusion modules, refine blocks and region proposal networks. On the other hand, an aspect ratio attention network is leveraged to preserve the aspect ratios of objects, which alleviates the excessive shape distortions of objects caused by aspect ratio changes during training. Experiments show that the proposed A2RMNet significantly outperforms the previous state of the arts on the DOTA dataset, NWPU VHR-10 dataset, RSOD dataset and UCAS-AOD dataset by 5.73%, 7.06%, 3.27% and 2.24%, respectively.

Original languageEnglish
Article number1594
Number of pages23
JournalRemote Sensing
Volume11
Issue number13
DOIs
Publication statusPublished - 4 Jul 2019
Externally publishedYes

Keywords

  • A multi-scale feature gate fusion network
  • Adaptively Aspect Ratio Multi-Scale Network (ARMNet)
  • An aspect ratio attention network
  • Object detection
  • Remote sensing and image analysis

Cite this