A novel attention-based deep learning method for post-disaster building damage classification

作者：

Highlights：

• Four building damage levels were classified rather than only collapse or intact.

• A SE attention added dual HRNet deep learning model was proposed.

• SE-PRE performed best among four SE added options.

• A Haiti Earthquake image dataset including four building damage levels was created.

摘要

•Four building damage levels were classified rather than only collapse or intact.•A SE attention added dual HRNet deep learning model was proposed.•SE-PRE performed best among four SE added options.•A Haiti Earthquake image dataset including four building damage levels was created.

论文关键词：Building damage classification,Deep learning,Channel attention,Remote sensing,Natural disaster,High resolution imagery

论文评审过程：Received 13 September 2021, Revised 26 March 2022, Accepted 15 April 2022, Available online 20 April 2022, Version of Record 29 April 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117268