ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression

Most current hyperspectral image compression methods rely on well-designed modules to capture image structural information and long-range dependencies. However, these modules tend to increase computational complexity exponentially with the number of bands, which limits their performance under constr...

Full description

Saved in:
Bibliographic Details
Main Authors: Qizhi Fang, Zixuan Wang, Jingang Wang, Lili Zhang
Format: Article
Language:English
Published: MDPI AG 2025-03-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/6/1843
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850088254548738048
author Qizhi Fang
Zixuan Wang
Jingang Wang
Lili Zhang
author_facet Qizhi Fang
Zixuan Wang
Jingang Wang
Lili Zhang
author_sort Qizhi Fang
collection DOAJ
description Most current hyperspectral image compression methods rely on well-designed modules to capture image structural information and long-range dependencies. However, these modules tend to increase computational complexity exponentially with the number of bands, which limits their performance under constrained resources. To address these challenges, this paper proposes a novel triple-phase hybrid framework for hyperspectral image compression. The first stage utilizes an adaptive band selection technique to sample the raw hyperspectral image, which mitigates the computational burden. The second stage concentrates on high-fidelity compression, efficiently encoding both spatial and spectral information within the sampled band clusters. In the final stage, a reconstruction network compensates for sampling-induced losses to precisely restore the original spectral details. The proposed framework, known as ARM-Net, is evaluated on seven mixed hyperspectral datasets. Compared to state-of-the-art methods, ARM-Net achieves an overall improvement of approximately 1–2 dB in both the peak signal-to-noise ratio and multiscale structural similarity index measure, as well as a reduction in the average spectral angle mapper of approximately 0.1.
format Article
id doaj-art-6040e80a5c484f98b5e4dee9018becbe
institution DOAJ
issn 1424-8220
language English
publishDate 2025-03-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-6040e80a5c484f98b5e4dee9018becbe2025-08-20T02:43:03ZengMDPI AGSensors1424-82202025-03-01256184310.3390/s25061843ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image CompressionQizhi Fang0Zixuan Wang1Jingang Wang2Lili Zhang3Liaoning General Aviation Academy, Shenyang 110136, ChinaSchool of Electronical and Information Engineering, Shenyang Aerospace University, Shenyang 110136, ChinaSchool of Electronical and Information Engineering, Shenyang Aerospace University, Shenyang 110136, ChinaSchool of Electronical and Information Engineering, Shenyang Aerospace University, Shenyang 110136, ChinaMost current hyperspectral image compression methods rely on well-designed modules to capture image structural information and long-range dependencies. However, these modules tend to increase computational complexity exponentially with the number of bands, which limits their performance under constrained resources. To address these challenges, this paper proposes a novel triple-phase hybrid framework for hyperspectral image compression. The first stage utilizes an adaptive band selection technique to sample the raw hyperspectral image, which mitigates the computational burden. The second stage concentrates on high-fidelity compression, efficiently encoding both spatial and spectral information within the sampled band clusters. In the final stage, a reconstruction network compensates for sampling-induced losses to precisely restore the original spectral details. The proposed framework, known as ARM-Net, is evaluated on seven mixed hyperspectral datasets. Compared to state-of-the-art methods, ARM-Net achieves an overall improvement of approximately 1–2 dB in both the peak signal-to-noise ratio and multiscale structural similarity index measure, as well as a reduction in the average spectral angle mapper of approximately 0.1.https://www.mdpi.com/1424-8220/25/6/1843hyperspectral image compressionspectral reconstructionspatial-spectral attention mechanismrecurrent spectral attention mechanism
spellingShingle Qizhi Fang
Zixuan Wang
Jingang Wang
Lili Zhang
ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression
Sensors
hyperspectral image compression
spectral reconstruction
spatial-spectral attention mechanism
recurrent spectral attention mechanism
title ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression
title_full ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression
title_fullStr ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression
title_full_unstemmed ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression
title_short ARM-Net: A Tri-Phase Integrated Network for Hyperspectral Image Compression
title_sort arm net a tri phase integrated network for hyperspectral image compression
topic hyperspectral image compression
spectral reconstruction
spatial-spectral attention mechanism
recurrent spectral attention mechanism
url https://www.mdpi.com/1424-8220/25/6/1843
work_keys_str_mv AT qizhifang armnetatriphaseintegratednetworkforhyperspectralimagecompression
AT zixuanwang armnetatriphaseintegratednetworkforhyperspectralimagecompression
AT jingangwang armnetatriphaseintegratednetworkforhyperspectralimagecompression
AT lilizhang armnetatriphaseintegratednetworkforhyperspectralimagecompression