Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection
Change detection is an important task in geospatial analysis that aims to identify noticeable variations in geographic elements between images captured at different periods. However, existing methods often overlook the distribution discrepancies across images caused by changes in imaging time. Meanw...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10748394/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841533450282074112 |
|---|---|
| author | Xuexiang Qin Yuxiang Zhang Yanni Dong |
| author_facet | Xuexiang Qin Yuxiang Zhang Yanni Dong |
| author_sort | Xuexiang Qin |
| collection | DOAJ |
| description | Change detection is an important task in geospatial analysis that aims to identify noticeable variations in geographic elements between images captured at different periods. However, existing methods often overlook the distribution discrepancies across images caused by changes in imaging time. Meanwhile, the spectral and spatial features of hyperspectral images still have great potential for further development in extracting and detecting changes. To mitigate these challenges, we propose a novel approach called domain alignment dynamic spectral and spatial feature fusion (DADSSFF) for hyperspectral change detection. First, DADSSFF uses the main network to optimize the alignment of the mean (first-order statistics) and correlation (variance, second-order statistics) of the bitemporal images, coordinating features across both levels to alleviate the issue of inconsistent feature distribution. Second, the Kullback–Leibler divergence is employed to increase the interaction between the two auxiliary networks and the main network, enhancing the extraction of spectral and spatial attention features from bitemporal hyperspectral images. Finally, the cosine similarity is applied to measure the weights of the spectral and spatial features, enabling a dynamic evaluation of their importance. The effectiveness of DADSSFF is demonstrated by experimental results on three classical hyperspectral change detection datasets. |
| format | Article |
| id | doaj-art-ed3b506c07eb4110bd9fbc5728b60715 |
| institution | Kabale University |
| issn | 1939-1404 2151-1535 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj-art-ed3b506c07eb4110bd9fbc5728b607152025-01-16T00:00:29ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-011855756810.1109/JSTARS.2024.349521710748394Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change DetectionXuexiang Qin0Yuxiang Zhang1https://orcid.org/0000-0002-2913-3515Yanni Dong2https://orcid.org/0000-0003-0592-7887School of Resource and Environmental Sciences, Wuhan University, Wuhan, ChinaSchool of Geophysics and Geomatics, China University of Geosciences, Wuhan, ChinaSchool of Resource and Environmental Sciences, Wuhan University, Wuhan, ChinaChange detection is an important task in geospatial analysis that aims to identify noticeable variations in geographic elements between images captured at different periods. However, existing methods often overlook the distribution discrepancies across images caused by changes in imaging time. Meanwhile, the spectral and spatial features of hyperspectral images still have great potential for further development in extracting and detecting changes. To mitigate these challenges, we propose a novel approach called domain alignment dynamic spectral and spatial feature fusion (DADSSFF) for hyperspectral change detection. First, DADSSFF uses the main network to optimize the alignment of the mean (first-order statistics) and correlation (variance, second-order statistics) of the bitemporal images, coordinating features across both levels to alleviate the issue of inconsistent feature distribution. Second, the Kullback–Leibler divergence is employed to increase the interaction between the two auxiliary networks and the main network, enhancing the extraction of spectral and spatial attention features from bitemporal hyperspectral images. Finally, the cosine similarity is applied to measure the weights of the spectral and spatial features, enabling a dynamic evaluation of their importance. The effectiveness of DADSSFF is demonstrated by experimental results on three classical hyperspectral change detection datasets.https://ieeexplore.ieee.org/document/10748394/Cosine similaritydomain alignmenthyperspectral change detectionKullback–Leibler divergence (KLD)spectral and spatial attention |
| spellingShingle | Xuexiang Qin Yuxiang Zhang Yanni Dong Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Cosine similarity domain alignment hyperspectral change detection Kullback–Leibler divergence (KLD) spectral and spatial attention |
| title | Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection |
| title_full | Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection |
| title_fullStr | Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection |
| title_full_unstemmed | Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection |
| title_short | Domain Alignment Dynamic Spectral and Spatial Feature Fusion for Hyperspectral Change Detection |
| title_sort | domain alignment dynamic spectral and spatial feature fusion for hyperspectral change detection |
| topic | Cosine similarity domain alignment hyperspectral change detection Kullback–Leibler divergence (KLD) spectral and spatial attention |
| url | https://ieeexplore.ieee.org/document/10748394/ |
| work_keys_str_mv | AT xuexiangqin domainalignmentdynamicspectralandspatialfeaturefusionforhyperspectralchangedetection AT yuxiangzhang domainalignmentdynamicspectralandspatialfeaturefusionforhyperspectralchangedetection AT yannidong domainalignmentdynamicspectralandspatialfeaturefusionforhyperspectralchangedetection |