An Attention Feature Interaction Change Detection Method Based on Detail Enhancement for Dual-Temporal Hyperspectral Images

The application of hyperspectral image change detection (HSI-CD) in remote sensing is becoming increasingly widespread. However, due to the low spatial resolution of HSIs, conducting CD directly on the original HSIs does not effectively capture subtle changes. Therefore, this letter proposes an atte...

Full description

Saved in:
Bibliographic Details
Published in:IEEE geoscience and remote sensing letters 2024, Vol.21, p.1-5
Main Authors: Feng, Shou, Zhang, Jinghe, Huang, Fenghua, Peng, Ruihui, Zhao, Chunhui
Format: Article
Language:English
Subjects:
Attention feature interaction
Convolution
Correlation
detail enhancement
Feature extraction
HSIs
hyperspectral image change detection (HSI-CD)
Hyperspectral imaging
Image restoration
multiscale feature extraction
Rivers
Spatial resolution
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The application of hyperspectral image change detection (HSI-CD) in remote sensing is becoming increasingly widespread. However, due to the low spatial resolution of HSIs, conducting CD directly on the original HSIs does not effectively capture subtle changes. Therefore, this letter proposes an attention feature interaction CD method based on detail enhancement for dual-temporal HSIs (AIDECD). First, a detail enhancement module is designed to enhance the detail information of original HSIs. Second, considering the relationship between dual-temporal images, an attention interaction module is designed to achieve the interaction of temporal features between the dual-temporal images. Then, a multiscale feature extraction module is designed to capture features of different scales. The kappa coefficients obtained on three HSI datasets are 86.11%, 96.08%, and 97.54%, respectively. Compared with six other CD methods, this method has higher detection performance.
ISSN:1545-598X
1558-0571
DOI:10.1109/LGRS.2024.3435429