Share Email Print

Proceedings Paper

Multi-scale oriented object detection in aerial images based on convolutional neural networks with global attention
Author(s): Jingjing Fei; Zhicheng Wang; Zhaohui Yu; Xi Gu; Gang Wei
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Object detection is a fundamental yet challenging problem in natural scenes and aerial scenes. Although region based deep convolutional neural networks (CNNs) have brought impressive improvements for object detection in natural scenes, detecting oriented objects in aerial images still remains challenging, due to the complexity of the aerial image backgrounds and the large degree of freedom in scale, orientation, and density. To tackle these problems, we propose a novel network, composed of backbone structure with global attention module, multi-scale object proposal network and final oriented object detector, which can efficiently detect small objects, arbitrary direction objects, and dense objects in aerial images. We utilize pyramid pooling blocks as a global attention module on the top of the backbone structure to generate discriminative feature representations, which provide diverse context information and complementary receptive field for the detector. The global attention module can help the model reduce false alarms and incorrect classifications in the complex aerial image backgrounds. The multi-scale object proposal network aims to generate object-like regions at different scales through several intermediate layers. After that, these regions are sent to the detector for refined classification and regression, which can alleviate the problem of variant scales in aerial images. The oriented object detector is designed to generate predictions for inclined box. The quantitative comparison results on the challenging DOTA dataset show that our proposed method is more accurate than baseline algorithms and is effective for objection detection in aerial images. The results demonstrate that the proposed method significantly improves the performance.

Paper Details

Date Published: 14 February 2020
PDF: 8 pages
Proc. SPIE 11432, MIPPR 2019: Remote Sensing Image Processing, Geographic Information Systems, and Other Applications, 114320K (14 February 2020); doi: 10.1117/12.2541855
Show Author Affiliations
Jingjing Fei, Tongji Univ. (China)
Zhicheng Wang, Tongji Univ. (China)
Zhaohui Yu, Tongji Univ. (China)
Xi Gu, Tongji Univ. (China)
Gang Wei, Tongji Univ. (China)

Published in SPIE Proceedings Vol. 11432:
MIPPR 2019: Remote Sensing Image Processing, Geographic Information Systems, and Other Applications
Zhiguo Cao; Jie Ma; Zhong Chen; Yu Shi, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?