Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection
Band selection (BS) is a critical topic in hyperspectral image dimensionality reduction, focusing on identifying representative bands that can convey the essential information of the full bands without significant loss. Recently, BS based on multiobjective optimization (MO) has become the predominan...
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/10771661/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841557091698868224 |
|---|---|
| author | Shihui Liu Bing Xue Meiping Song Haimo Bao Mengjie Zhang |
| author_facet | Shihui Liu Bing Xue Meiping Song Haimo Bao Mengjie Zhang |
| author_sort | Shihui Liu |
| collection | DOAJ |
| description | Band selection (BS) is a critical topic in hyperspectral image dimensionality reduction, focusing on identifying representative bands that can convey the essential information of the full bands without significant loss. Recently, BS based on multiobjective optimization (MO) has become the predominant method. However, there are still two problems that need to be solved. First, the majority of multiobjective BS methods predominantly optimize accuracy in classification tasks, neglecting the emphasis on anomaly detection. Second, in the process of addressing the combinatorial optimization problem of multiobjective BS using evolutionary algorithms, insufficient consideration is given to the impact of bands on the recognition capability of an anomaly when devising solution strategies and determining optimal solutions. To handle the above-mentioned problems, a novel anomaly-oriented multiobjective optimization band selection (AOMOBS) is developed to better suppress background and identify anomalies. Specifically, for the first problem, by calculating the degree of deviation of the band, the noise estimates of the band, and the degree of redundancy between bands, an unsupervised BS algorithm for anomaly detection tasks, based on MO, is designed. For the second problem, an MO band similarity sorting strategy for anomaly detection tasks is designed for nondominated sorting, and for decision makers to choose the most appropriate solution from the tradeoff solutions. Experiments conducted on real hyperspectral datasets demonstrate that the algorithm effectively identifies a subset of bands with high representational power for anomaly detection. Moreover, AOMOBS outperforms current state-of-the-art methods in both effectiveness and robustness. |
| format | Article |
| id | doaj-art-86cca326df91496a9bfd57029ebf4a14 |
| 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-86cca326df91496a9bfd57029ebf4a142025-01-07T00:00:42ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01182813282810.1109/JSTARS.2024.350882310771661Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly DetectionShihui Liu0https://orcid.org/0000-0002-6697-7899Bing Xue1https://orcid.org/0000-0002-4865-8026Meiping Song2https://orcid.org/0000-0002-4489-5470Haimo Bao3https://orcid.org/0000-0001-5706-2155Mengjie Zhang4https://orcid.org/0000-0003-4463-9538School of Information and Control Engineering, Qingdao University of Technology, Qingdao, ChinaCenter for Data Science and Artificial Intelligence and the School of Engineering and Computer Science, Victoria University of Wellington, Wellington, New ZealandSchool of Innovation Design, Guangzhou Academy of Fine Arts, Guangzhou, ChinaSchool of Innovation Design, Guangzhou Academy of Fine Arts, Guangzhou, ChinaCenter for Data Science and Artificial Intelligence and the School of Engineering and Computer Science, Victoria University of Wellington, Wellington, New ZealandBand selection (BS) is a critical topic in hyperspectral image dimensionality reduction, focusing on identifying representative bands that can convey the essential information of the full bands without significant loss. Recently, BS based on multiobjective optimization (MO) has become the predominant method. However, there are still two problems that need to be solved. First, the majority of multiobjective BS methods predominantly optimize accuracy in classification tasks, neglecting the emphasis on anomaly detection. Second, in the process of addressing the combinatorial optimization problem of multiobjective BS using evolutionary algorithms, insufficient consideration is given to the impact of bands on the recognition capability of an anomaly when devising solution strategies and determining optimal solutions. To handle the above-mentioned problems, a novel anomaly-oriented multiobjective optimization band selection (AOMOBS) is developed to better suppress background and identify anomalies. Specifically, for the first problem, by calculating the degree of deviation of the band, the noise estimates of the band, and the degree of redundancy between bands, an unsupervised BS algorithm for anomaly detection tasks, based on MO, is designed. For the second problem, an MO band similarity sorting strategy for anomaly detection tasks is designed for nondominated sorting, and for decision makers to choose the most appropriate solution from the tradeoff solutions. Experiments conducted on real hyperspectral datasets demonstrate that the algorithm effectively identifies a subset of bands with high representational power for anomaly detection. Moreover, AOMOBS outperforms current state-of-the-art methods in both effectiveness and robustness.https://ieeexplore.ieee.org/document/10771661/Anomaly detectionband selection (BS)hyperspectral images (HSI)multiobjective optimization (MO) |
| spellingShingle | Shihui Liu Bing Xue Meiping Song Haimo Bao Mengjie Zhang Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Anomaly detection band selection (BS) hyperspectral images (HSI) multiobjective optimization (MO) |
| title | Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection |
| title_full | Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection |
| title_fullStr | Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection |
| title_full_unstemmed | Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection |
| title_short | Multiobjective Optimization-Based Hyperspectral Unsupervised Band Selection for Anomaly Detection |
| title_sort | multiobjective optimization based hyperspectral unsupervised band selection for anomaly detection |
| topic | Anomaly detection band selection (BS) hyperspectral images (HSI) multiobjective optimization (MO) |
| url | https://ieeexplore.ieee.org/document/10771661/ |
| work_keys_str_mv | AT shihuiliu multiobjectiveoptimizationbasedhyperspectralunsupervisedbandselectionforanomalydetection AT bingxue multiobjectiveoptimizationbasedhyperspectralunsupervisedbandselectionforanomalydetection AT meipingsong multiobjectiveoptimizationbasedhyperspectralunsupervisedbandselectionforanomalydetection AT haimobao multiobjectiveoptimizationbasedhyperspectralunsupervisedbandselectionforanomalydetection AT mengjiezhang multiobjectiveoptimizationbasedhyperspectralunsupervisedbandselectionforanomalydetection |