{"title":"基于全局局部空间感知的高光谱图像降维保留投影","authors":"Tao Zhang , Fang Wang , Limin Zeng","doi":"10.1016/j.infrared.2025.106014","DOIUrl":null,"url":null,"abstract":"<div><div>Hyperspectral remote sensing images, constrained by noise interference and acquisition equipment limitations, often suffer from severe phenomena of ‘spectral confusion between heterogeneous targets (SCHT)’ and ‘spectral variability within homogeneous targets (SVHT)’. Conventional dimensionality reduction methods based on locality preserving projection (LPP) fail to effectively capture complex spatial–spectral features due to their neglect of spatial context structures. To address this challenge, this paper proposes the global-local spatially aware preserving projection (GLSAPP) framework, which achieves dual optimizations through spatial–spectral joint modeling. On the one hand, a composite neighborhood structure is constructed by integrating spatial distance neighborhoods and spectral similarity neighborhoods, and a dual-weighting mechanism for spectral consistency and spatial coherence is designed. This mechanism adaptively enhances intra-class compactness and inter-class discriminability during manifold learning. On the other hand, a novel local feature divergence metrics and a global class separation indices are proposed to measure the phenomena of SCHT and SVHT. A joint optimization model incorporating local geometric preservation terms and global repulsive regularization terms is then established. Experiments conducted on three publicly available hyperspectral datasets demonstrate that compared with state-of-the-art dimensionality reduction methods, GLSAPP significantly reduces both SCHT rate and SVHT rate. These results validate the effectiveness of GLSAPP in suppressing spectral variability and confusion through spatial–spectral collaborative optimization, providing a more accurate feature representation method for hyperspectral image classification.</div></div>","PeriodicalId":13549,"journal":{"name":"Infrared Physics & Technology","volume":"150 ","pages":"Article 106014"},"PeriodicalIF":3.4000,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Global-local spatially aware preserving projection for dimensionality reduction of hyperspectral images\",\"authors\":\"Tao Zhang , Fang Wang , Limin Zeng\",\"doi\":\"10.1016/j.infrared.2025.106014\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Hyperspectral remote sensing images, constrained by noise interference and acquisition equipment limitations, often suffer from severe phenomena of ‘spectral confusion between heterogeneous targets (SCHT)’ and ‘spectral variability within homogeneous targets (SVHT)’. Conventional dimensionality reduction methods based on locality preserving projection (LPP) fail to effectively capture complex spatial–spectral features due to their neglect of spatial context structures. To address this challenge, this paper proposes the global-local spatially aware preserving projection (GLSAPP) framework, which achieves dual optimizations through spatial–spectral joint modeling. On the one hand, a composite neighborhood structure is constructed by integrating spatial distance neighborhoods and spectral similarity neighborhoods, and a dual-weighting mechanism for spectral consistency and spatial coherence is designed. This mechanism adaptively enhances intra-class compactness and inter-class discriminability during manifold learning. On the other hand, a novel local feature divergence metrics and a global class separation indices are proposed to measure the phenomena of SCHT and SVHT. A joint optimization model incorporating local geometric preservation terms and global repulsive regularization terms is then established. Experiments conducted on three publicly available hyperspectral datasets demonstrate that compared with state-of-the-art dimensionality reduction methods, GLSAPP significantly reduces both SCHT rate and SVHT rate. These results validate the effectiveness of GLSAPP in suppressing spectral variability and confusion through spatial–spectral collaborative optimization, providing a more accurate feature representation method for hyperspectral image classification.</div></div>\",\"PeriodicalId\":13549,\"journal\":{\"name\":\"Infrared Physics & Technology\",\"volume\":\"150 \",\"pages\":\"Article 106014\"},\"PeriodicalIF\":3.4000,\"publicationDate\":\"2025-07-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Infrared Physics & Technology\",\"FirstCategoryId\":\"101\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S135044952500307X\",\"RegionNum\":3,\"RegionCategory\":\"物理与天体物理\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"INSTRUMENTS & INSTRUMENTATION\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Infrared Physics & Technology","FirstCategoryId":"101","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S135044952500307X","RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"INSTRUMENTS & INSTRUMENTATION","Score":null,"Total":0}
Global-local spatially aware preserving projection for dimensionality reduction of hyperspectral images
Hyperspectral remote sensing images, constrained by noise interference and acquisition equipment limitations, often suffer from severe phenomena of ‘spectral confusion between heterogeneous targets (SCHT)’ and ‘spectral variability within homogeneous targets (SVHT)’. Conventional dimensionality reduction methods based on locality preserving projection (LPP) fail to effectively capture complex spatial–spectral features due to their neglect of spatial context structures. To address this challenge, this paper proposes the global-local spatially aware preserving projection (GLSAPP) framework, which achieves dual optimizations through spatial–spectral joint modeling. On the one hand, a composite neighborhood structure is constructed by integrating spatial distance neighborhoods and spectral similarity neighborhoods, and a dual-weighting mechanism for spectral consistency and spatial coherence is designed. This mechanism adaptively enhances intra-class compactness and inter-class discriminability during manifold learning. On the other hand, a novel local feature divergence metrics and a global class separation indices are proposed to measure the phenomena of SCHT and SVHT. A joint optimization model incorporating local geometric preservation terms and global repulsive regularization terms is then established. Experiments conducted on three publicly available hyperspectral datasets demonstrate that compared with state-of-the-art dimensionality reduction methods, GLSAPP significantly reduces both SCHT rate and SVHT rate. These results validate the effectiveness of GLSAPP in suppressing spectral variability and confusion through spatial–spectral collaborative optimization, providing a more accurate feature representation method for hyperspectral image classification.
期刊介绍:
The Journal covers the entire field of infrared physics and technology: theory, experiment, application, devices and instrumentation. Infrared'' is defined as covering the near, mid and far infrared (terahertz) regions from 0.75um (750nm) to 1mm (300GHz.) Submissions in the 300GHz to 100GHz region may be accepted at the editors discretion if their content is relevant to shorter wavelengths. Submissions must be primarily concerned with and directly relevant to this spectral region.
Its core topics can be summarized as the generation, propagation and detection, of infrared radiation; the associated optics, materials and devices; and its use in all fields of science, industry, engineering and medicine.
Infrared techniques occur in many different fields, notably spectroscopy and interferometry; material characterization and processing; atmospheric physics, astronomy and space research. Scientific aspects include lasers, quantum optics, quantum electronics, image processing and semiconductor physics. Some important applications are medical diagnostics and treatment, industrial inspection and environmental monitoring.