pith. sign in

arxiv: 2510.18935 · v3 · submitted 2025-10-21 · 💻 cs.CV

Feature Extraction in the Remote Sensing Data Value Chain: A Systematic Review of Methods and Applications

Nathan Mankovich , Kai-Hendrik Cohrs , Homer Durand , Vasileios Sitokonstantinou , Tristan Williams , Gustau Camps-Valls This is my paper

Pith reviewed 2026-05-18 04:41 UTC · model grok-4.3

classification 💻 cs.CV
keywords remote sensingfeature extractiondata value chainfoundation modelsrepresentation learningEarth observationmachine learninginterpretability
0
0 comments X

The pith

A framework for feature extraction traces its evolution in remote sensing from task-specific tools to unified representations that support multiple applications.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a structured framework to organize the diverse methods of feature extraction used in remote sensing data processing. It follows how these methods have changed as data moves through the full value chain from collection and preprocessing to analysis and decision support. The review identifies a clear progression away from models built for one narrow task toward shared representations that serve many goals at once. It points to two key needs in the current era of large foundation models: making extracted features more robust against noise and variation, and keeping them interpretable so users can understand why decisions are made. The authors argue that connecting older, hand-designed extraction techniques with newer learned representations offers a practical path forward for handling the growing volume of Earth observation data.

Core claim

The paper establishes a framework that classifies feature extraction techniques according to their position and role in the remote sensing data value chain. Using this structure, it documents the historical shift from single-task, often hand-crafted methods to unified, multi-purpose representations. In the foundation-model setting, the work concludes that future progress requires both stronger robustness and interpretability in extracted features and deliberate efforts to combine classical extraction approaches with modern representation learning.

What carries the argument

A proposed framework that organizes feature extraction methods by their stage and function within the remote sensing data value chain, enabling comparison across classical and modern techniques.

If this is right

  • Unified representations reduce feature redundancy and lower the computational cost of analyzing high-dimensional satellite and aerial data.
  • Robust and interpretable feature extraction becomes a required property for foundation models applied to environmental monitoring and disaster response.
  • Bridging classical extraction methods with learned representations can preserve domain knowledge while scaling to larger datasets.
  • The data-value-chain view helps practitioners choose extraction techniques that match the specific stage and goal of their remote-sensing pipeline.
  • Future work should test whether hybrid classical-modern features improve performance on downstream tasks such as land-cover classification and change detection.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The framework could be extended to quantify how much interpretability is gained or lost when moving from classical to deep-learned features.
  • Applying the same value-chain lens to non-remote-sensing domains such as medical imaging might reveal similar shifts toward unified representations.
  • Developers of new foundation models for Earth observation could use the framework to decide where to insert explicit feature-extraction modules rather than relying solely on end-to-end learning.
  • Empirical tests could measure whether models built on the proposed unified representations actually reduce the need for task-specific retraining across different remote-sensing applications.

Load-bearing premise

The review assumes that the collection of papers surveyed is broad enough and representative enough to reveal the true trends in how feature extraction has developed.

What would settle it

Discovery of a large, coherent body of recent remote-sensing literature on feature extraction that falls outside the proposed framework or that shows no movement toward unified representations.

Figures

Figures reproduced from arXiv: 2510.18935 by Gustau Camps-Valls, Homer Durand, Kai-Hendrik Cohrs, Nathan Mankovich, Tristan Williams, Vasileios Sitokonstantinou.

Figure 1
Figure 1. Figure 1: A timeline of common DR methods for RS. DR for feature extraction began with linear multivariate analysis methods like Principal Component Analysis (PCA) [10] in the 1930s, and became popular in the RS community in the 1970s [11]. The manifold learning boom began in the late 1900s, including nonlinear DR like kernel PCA [12], and was quickly adopted by the remote sensing community within 10 years [13]. Inc… view at source ↗
Figure 2
Figure 2. Figure 2: DR improves performance across the RS data value chain. Data from RS sensors come in various data types, often high-dimensional in spatial, spectral, and temporal dimensions. DR reduces these dimensions to address challenges at each phase of the value chain of RS applications, from pre-processing (e.g., data compression Sec. III-A1, cleaning Sec. III-A2, and fusion Sec. III-A3) to analysis (e.g., visualiza… view at source ↗
Figure 3
Figure 3. Figure 3: A taxonomy of DR. DR characteristics are separated into the mapping, dataset, and constraints/property preservation. The DR mapping can be either explicit or implicit. Explicit DR mappings are either linear or nonlinear parameterized functions. Different DR methods are used for different tasks based on the input dataset. Unsupervised DR methods just input a dataset D ⊂ X, whereas supervised methods take a … view at source ↗
Figure 4
Figure 4. Figure 4: The utility of DR for RS data from various widely used Earth observation sensors. The dimensions of these data are spectral, spatial, and temporal. A higher value for a data type indicates a higher need for DR to reduce features in that dimension. Hyperspectral Imaging (HSI) sensors exhibit high spectral redundancy, requiring dimensionality reduction in the spectral domain. Synthetic Aperture Radar (SAR) a… view at source ↗
Figure 5
Figure 5. Figure 5: ). Spatial Spectral Temporal Time Spectrum Space All Input data (Ambient space) Reduced space Reconstructed data [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: DR for data cleaning. DR for data cleaning generally use an explicit mapping ϕ with approximate inverse ψ. For image restoration, enhancement, and denoising, the data are mapped to a reduced space, then back to the ambient space, and this filtering removes the noisy (often high-frequency) information. In contrast, gap-filling first identifies the dimensions with the gaps (e.g., spatial or temporal). Then, … view at source ↗
Figure 7
Figure 7. Figure 7: DR for data fusion. Data fusion can be applied at different levels: pixel-level fusion is applied to satellite images, combining them pixel-wise to enhance spatial and spectral details. Feature-level fusion integrates extracted features from potentially diverse modalities, and decision-level fusion com￾bines independent predictions. Dimensionality reduction techniques aid pixel and feature-level fusion by … view at source ↗
Figure 8
Figure 8. Figure 8: DR for data visualization. Different visualization techniques preserve specific aspects of remote sensing data. Spatial visualization reduces spectral and/or temporal dimensions to generate maps, preserving spatial structures and patterns. Temporal visualization compresses spectral and spatial information into time series, highlighting different temporal patterns. Abstract space projection discards domain-… view at source ↗
Figure 9
Figure 9. Figure 9: Two-dimensional embeddings of spectral data generated using various dimensionality reduction techniques. The data originates from the HyperLabelme dataset [177], specifically from the FlightLineC1 site and the M7scanner sensor. The dimensionality reduction algorithms were trained on 498 samples, each with a spectral dimensionality of 128. Spatial Variables Temporal Input data (Ambient space) Spatial Spectr… view at source ↗
Figure 10
Figure 10. Figure 10: DR for anomaly detection. Two main approaches to anomaly detection with DR. First, DR to separate anomalies from the background, followed by a statistical or ML-based anomaly detection algorithm. This concept is illustrated in the upper figure, which displays a spatio-temporal multivariate data cube [178] for detecting spatio-temporal anomalies (credit: ©ESA). In the second approach, anomalies are detecte… view at source ↗
Figure 11
Figure 11. Figure 11: DR for improving prediction tasks in RS. DR methods enhance classification and regression by increasing the discriminatory power of ex￾tracted features, reducing overfitting, and eliminating noise. Four key perspec￾tives guide this enhancement. Dataset augmentation using DR improves model generalization. Reducing spectral-spatial redundancy in HS data improves feature extraction and predictive performance… view at source ↗
Figure 12
Figure 12. Figure 12: Article counts for each DR task and metric. These counts only include articles cited in this paper. DWT ICA Isomap LDA LLE NMF PCA kPCA t-SNE Method 10 3 10 2 10 1 10 0 Computation Time (s, log scale) FlightLineC1 N=469 P=12 Barrax99 N=498 P=128 Botswana N=504 P=145 KSC N=494 P=176 Indian Pines N=453 P=220 [PITH_FULL_IMAGE:figures/full_fig_p017_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Computation times for common DR algorithms for different HyperLabelMe datasets [177]. We evaluate unsupervised methods and one supervised method for classification datasets (LDA) to reduce from P to K = 2 dimensions. All methods are run on a 2020 MacBook Pro with M1 chip and 16GB of memory. A. Anomaly detection and predictions DR for compression, denoising, and fusion is evaluated on the capacity of the r… view at source ↗
Figure 14
Figure 14. Figure 14: DR methods are characterized by dataset (first, inner circle), mapping (second circle), and optimization problem/ property preservation (third circle). Although methods similar to DFT and DWT can be seen as matrix factorization methods, we choose to separate them as signal processing transforms. A. Compression Standard DR methods for RS treat all dimensions of RS data uniformly, overlooking the sequential… view at source ↗
Figure 15
Figure 15. Figure 15: Using the feature space of ResNet as a pre-processing step for DR for visualizing SAR data. There is an improvement in class separation when DR is run on features extraced by ResNet [PITH_FULL_IMAGE:figures/full_fig_p027_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Two-dimensional embeddings of spectral data were produced using a range of dimensionality reduction techniques. The data is sourced from the HyperLabelme dataset [177], comprising the FlightLineC1, Barrax, Botswana, KSC, and Indian Pines sites, ordered by increasing spectral dimensionality. For each site and sensor, the number of samples and spectral dimensionality are reported as (N, P) [PITH_FULL_IMAGE… view at source ↗
read the original abstract

Earth observation involves collecting, analyzing, and processing an ever-growing mass of data. This planetary data is crucial for addressing relevant societal, economic, and environmental challenges, ranging from environmental monitoring to urban planning and disaster management. However, its high dimensionality entails significant feature redundancy and computational overhead, limiting the effectiveness of machine learning models. Feature extraction (FE) techniques address these challenges by preserving essential data properties while reducing redundancy and enhancing tasks in Remote Sensing (RS). The landscape of FE for RS is diverse, disorganized, and rapidly evolving. We offer a practical guide for this landscape by introducing a framework of FE. Using this framework, we trace the evolution of FE across the data value chain in RS. Finally, we synthesize these trends and offer perspectives for the future of FE in RS by first characterizing this shift from single-task models to unified representations, then identifying two perspectives in the foundation model era: the need for robust and interpretable FE and the potential of bridging classical FE with modern representation learning.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper presents a systematic review of feature extraction (FE) techniques in remote sensing (RS). It introduces a framework for organizing FE methods, traces their evolution across the RS data value chain, characterizes a shift from single-task models to unified representations, and synthesizes two perspectives for the foundation-model era: the need for robust and interpretable FE and the potential to bridge classical FE with modern representation learning.

Significance. If the literature base is representative, the work offers a practical organizing framework and forward-looking synthesis for a rapidly evolving subfield, potentially helping researchers connect classical dimensionality-reduction techniques with foundation-model representations in Earth-observation applications.

major comments (2)
  1. [Methods] Methods section: the literature-selection protocol (search strings, databases, time window, inclusion/exclusion criteria, and handling of 2023–2024 foundation-model papers) is not described with sufficient specificity or accompanied by a PRISMA-style flow diagram. Because the central claims about evolutionary trends and the two foundation-model perspectives rest on the representativeness of the surveyed corpus, this omission is load-bearing for the synthesis.
  2. [Framework definition] Framework introduction (early sections): the proposed FE framework is presented as a practical guide, yet the manuscript does not explicitly compare it against prior taxonomies in RS (e.g., those based on spectral, spatial, or deep-feature categories) or demonstrate how the new framework adds non-redundant structure. This weakens the justification for using it to trace the claimed single-task-to-unified shift.
minor comments (2)
  1. [Abstract and Introduction] Several long sentences in the abstract and introduction could be split to improve readability.
  2. [Figures] Figure captions should explicitly state the data sources or example images used so readers can assess representativeness.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We sincerely thank the referee for the constructive and detailed feedback on our systematic review. We address each major comment below and describe the revisions we will implement to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Methods] Methods section: the literature-selection protocol (search strings, databases, time window, inclusion/exclusion criteria, and handling of 2023–2024 foundation-model papers) is not described with sufficient specificity or accompanied by a PRISMA-style flow diagram. Because the central claims about evolutionary trends and the two foundation-model perspectives rest on the representativeness of the surveyed corpus, this omission is load-bearing for the synthesis.

    Authors: We agree that greater specificity is required for a systematic review. The current manuscript outlines the overall search approach at a high level but does not provide the granular protocol details requested. In the revised manuscript we will expand the Methods section with: (i) the precise search strings used (combinations of terms such as “feature extraction” AND (“remote sensing” OR “Earth observation”) together with keywords for classical and deep-learning methods); (ii) the databases queried (Scopus, Web of Science, IEEE Xplore, arXiv, and Google Scholar); (iii) the time window (2010–2024, with explicit handling of the 2023–2024 foundation-model surge via targeted supplementary searches); (iv) clear inclusion/exclusion criteria; and (v) a PRISMA-style flow diagram documenting the screening and selection process. These additions will directly substantiate the representativeness of the corpus and thereby support the evolutionary-trend and foundation-model-perspective claims. revision: yes

  2. Referee: [Framework definition] Framework introduction (early sections): the proposed FE framework is presented as a practical guide, yet the manuscript does not explicitly compare it against prior taxonomies in RS (e.g., those based on spectral, spatial, or deep-feature categories) or demonstrate how the new framework adds non-redundant structure. This weakens the justification for using it to trace the claimed single-task-to-unified shift.

    Authors: We acknowledge that an explicit side-by-side comparison with established taxonomies would strengthen the justification. While the framework’s primary contribution lies in situating feature extraction within the RS data value chain and in highlighting the transition toward unified representations, the manuscript does not currently include a dedicated comparison subsection. In the revision we will add a concise comparison (text plus table) in the framework-introduction section that contrasts our value-chain-oriented organization against prior spectral/spatial, handcrafted/learned, and deep-feature taxonomies. This addition will clarify the non-redundant elements—particularly the explicit linkage to downstream tasks and the synthesis toward foundation models—thereby reinforcing the rationale for employing the framework to trace the single-task-to-unified shift. revision: yes

Circularity Check

0 steps flagged

No significant circularity in this literature review synthesis

full rationale

This paper is a systematic review that synthesizes existing literature on feature extraction methods in remote sensing. It introduces a conceptual framework and traces evolutionary trends across the data value chain solely by surveying and organizing published work, without any mathematical derivations, equations, fitted parameters, or predictions that could reduce to the paper's own inputs by construction. No self-citation chains, ansatzes, or uniqueness theorems are invoked as load-bearing elements for new results. The central claims rest on the external literature base rather than internal self-reference, rendering the synthesis self-contained with no circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

As a systematic review paper, no new mathematical derivations, fitted parameters, or invented entities are introduced; the contribution rests on synthesis of prior literature.

pith-pipeline@v0.9.0 · 5724 in / 1081 out tokens · 35183 ms · 2026-05-18T04:41:11.260045+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

234 extracted references · 234 canonical work pages · 4 internal anchors

  1. [1]

    Deep learning and process understand- ing for data-driven Earth system science,

    M. Reichstein, G. Camps-Valls, B. Stevens, M. Jung, J. Denzler, N. Carvalhais, and F. Prabhat, “Deep learning and process understand- ing for data-driven Earth system science,”Nature, vol. 566, no. 7743, pp. 195–204, 2019

    work page 2019

  2. [2]

    A review of remote sensing for environmental monitoring in China,

    J. Li, Y . Pei, S. Zhao, R. Xiao, X. Sang, and C. Zhang, “A review of remote sensing for environmental monitoring in China,”Remote Sensing, vol. 12, no. 7, p. 1130, 2020

    work page 2020

  3. [3]

    Application of remote sensing and GIS in agriculture and natural resource management under changing climatic conditions

    P. K. Kingra, D. Majumder, and S. P. Singh, “Application of remote sensing and GIS in agriculture and natural resource management under changing climatic conditions.”Agricultural Research Journal, vol. 53, no. 3, 2016

    work page 2016

  4. [4]

    Remote sensing in urban planning: Contributions towards ecologically sound policies?

    T. Wellmann, A. Lausch, E. Andersson, S. Knapp, C. Cortinovis, J. Jache, S. Scheuer, P. Kremer, A. Mascarenhas, R. Kraemeret al., “Remote sensing in urban planning: Contributions towards ecologically sound policies?”Landscape and urban planning, vol. 204, p. 103921, 2020

    work page 2020

  5. [5]

    Remote sensing for agricultural applications: A meta-review,

    M. Weiss, F. Jacob, and G. Duveiller, “Remote sensing for agricultural applications: A meta-review,”Remote sensing of environment, vol. 236, p. 111402, 2020

    work page 2020

  6. [6]

    Remote sensing for natural disaster management,

    C. Van Westen, “Remote sensing for natural disaster management,” International archives of photogrammetry and remote sensing, vol. 33, no. B7/4; PART 7, pp. 1609–1617, 2000

    work page 2000

  7. [7]

    Sentinels optical communications payload (ocp) operations: From test to in-flight experience,

    I. Shurmer, F. Marchese, J.-M. Morales-Santiago, and P. P. Emanuelli, “Sentinels optical communications payload (ocp) operations: From test to in-flight experience,” in2018 SpaceOps Conference, 2018, p. 2654

    work page 2018

  8. [8]

    GOES-R series introduction,

    S. J. Goodman, “GOES-R series introduction,” inThe GOES-R Series. Elsevier, 2020, pp. 1–3

    work page 2020

  9. [9]

    The curse(s) of dimensionality,

    N. Altman and M. Krzywinski, “The curse(s) of dimensionality,” Nature Methods, vol. 15, no. 6, pp. 399–400, 2018

    work page 2018

  10. [10]

    Analysis of a complex of statistical variables into principal components

    H. Hotelling, “Analysis of a complex of statistical variables into principal components.”Journal of educational psychology, vol. 24, no. 6, p. 417, 1933

    work page 1933

  11. [11]

    Passive remote sensing techniques for mapping water depth and bottom features,

    D. R. Lyzenga, “Passive remote sensing techniques for mapping water depth and bottom features,”Applied optics, vol. 17, no. 3, pp. 379–383, 1978

    work page 1978

  12. [12]

    Kernel principal com- ponent analysis,

    B. Schölkopf, A. Smola, and K.-R. Müller, “Kernel principal com- ponent analysis,” inInternational conference on artificial neural net- works. Springer, 1997, pp. 583–588

    work page 1997

  13. [13]

    Camps-Valls and L

    G. Camps-Valls and L. Bruzzone,Kernel methods for remote sensing data analysis. John Wiley & Sons, 2009

    work page 2009

  14. [14]

    Auto-Encoding Variational Bayes

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,”arXiv preprint arXiv:1312.6114, 2013

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  15. [15]

    Deep learning in remote sensing: A comprehensive review and list of resources,

    X. X. Zhu, D. Tuia, L. Mou, G.-S. Xia, L. Zhang, F. Xu, and F. Fraundorfer, “Deep learning in remote sensing: A comprehensive review and list of resources,”IEEE Geoscience and Remote Sensing Magazine, vol. 5, no. 4, pp. 8–36, 2017

    work page 2017

  16. [16]

    Comprehensive review of hyperspectral image compression algorithms,

    Y . Dua, V . Kumar, and R. S. Singh, “Comprehensive review of hyperspectral image compression algorithms,”Optical Engineering, vol. 59, no. 9, pp. 090 902–090 902, 2020

    work page 2020

  17. [17]

    Image restoration for remote sensing: Overview and toolbox,

    B. Rasti, Y . Chang, E. Dalsasso, L. Denis, and P. Ghamisi, “Image restoration for remote sensing: Overview and toolbox,”IEEE Geo- science and Remote Sensing Magazine, vol. 10, no. 2, pp. 201–230, 2021

    work page 2021

  18. [18]

    Deep learning in multimodal remote sensing data fusion: A compre- hensive review,

    J. Li, D. Hong, L. Gao, J. Yao, K. Zheng, B. Zhang, and J. Chanussot, “Deep learning in multimodal remote sensing data fusion: A compre- hensive review,”International Journal of Applied Earth Observation and Geoinformation, vol. 112, p. 102926, 2022

    work page 2022

  19. [19]

    Big data for remote sensing: Visualization, analysis and interpretation,

    N. Dey, C. Bhatt, and A. S. Ashour, “Big data for remote sensing: Visualization, analysis and interpretation,”Cham: Springer, vol. 104, 2018

    work page 2018

  20. [20]

    Hyperspectral anomaly detection using deep learning: A review,

    X. Hu, C. Xie, Z. Fan, Q. Duan, D. Zhang, L. Jiang, X. Wei, D. Hong, G. Li, X. Zenget al., “Hyperspectral anomaly detection using deep learning: A review,”Remote Sensing, vol. 14, no. 9, p. 1973, 2022

    work page 1973

  21. [21]

    Implementation of machine-learning classification in remote sensing: An applied review,

    A. E. Maxwell, T. A. Warner, and F. Fang, “Implementation of machine-learning classification in remote sensing: An applied review,” International journal of remote sensing, vol. 39, no. 9, pp. 2784–2817, 2018

    work page 2018

  22. [22]

    Feature extraction for hyperspectral imagery: The evolution from shallow to deep: Overview and toolbox,

    B. Rasti, D. Hong, R. Hang, P. Ghamisi, X. Kang, J. Chanussot, and J. A. Benediktsson, “Feature extraction for hyperspectral imagery: The evolution from shallow to deep: Overview and toolbox,”IEEE Geoscience and Remote Sensing Magazine, vol. 8, no. 4, pp. 60–88, 2020

    work page 2020

  23. [23]

    Low- rank and sparse representation for hyperspectral image processing: A review,

    J. Peng, W. Sun, H.-C. Li, W. Li, X. Meng, C. Ge, and Q. Du, “Low- rank and sparse representation for hyperspectral image processing: A review,”IEEE Geoscience and Remote Sensing Magazine, vol. 10, no. 1, pp. 10–43, 2022

    work page 2022

  24. [24]

    Tensor decompositions for hyperspectral data processing in remote sensing: A comprehensive review,

    M. Wang, D. Hong, Z. Han, J. Li, J. Yao, L. Gao, B. Zhang, and J. Chanussot, “Tensor decompositions for hyperspectral data processing in remote sensing: A comprehensive review,”IEEE Geoscience and Remote Sensing Magazine, vol. 11, no. 1, pp. 26–72, 2023

    work page 2023

  25. [25]

    Advanced feature extraction for Earth observation data processing,

    E. Izquierdo-Verdiguier, V . Laparra, J. M. Marí, L. G. Chova, and G. Camps-Valls, “Advanced feature extraction for Earth observation data processing,” inComprehensive remote sensing, volume 2: data processing and analysis methodology. Elsevier, 2017, pp. 108–133

    work page 2017

  26. [26]

    Dimensionality reduction: A comparative review,

    L. Van Der Maaten, E. O. Postma, H. J. Van Den Heriket al., “Dimensionality reduction: A comparative review,”Journal of machine learning research, vol. 10, no. 66-71, p. 13, 2009

    work page 2009

  27. [27]

    J. A. Lee and M. Verleysen,Nonlinear dimensionality reduction. Springer Science & Business Media, 2007

    work page 2007

  28. [28]

    Review of dimension reduction methods,

    S. Nanga, A. T. Bawah, B. A. Acquaye, M.-I. Billa, F. D. Baeta, N. A. Odai, S. K. Obeng, and A. D. Nsiah, “Review of dimension reduction methods,”Journal of Data Analysis and Information Processing, vol. 9, no. 3, pp. 189–231, 2021

    work page 2021

  29. [29]

    A short survey of hyperspectral remote sensing applications in agri- culture,

    M. Teke, H. S. Deveci, O. Halilo ˘glu, S. Z. Gürbüz, and U. Sakarya, “A short survey of hyperspectral remote sensing applications in agri- culture,” in2013 6th international conference on recent advances in space technologies (RAST). IEEE, 2013, pp. 171–176

    work page 2013

  30. [30]

    Autoencoders,

    D. Bank, N. Koenigstein, and R. Giryes, “Autoencoders,”Machine Learning for Data Science Handbook: Data Mining and Knowledge Discovery Handbook, pp. 353–374, 2023

    work page 2023

  31. [31]

    A survey on canonical correlation analysis,

    X. Yang, W. Liu, W. Liu, and D. Tao, “A survey on canonical correlation analysis,”IEEE Transactions on Knowledge and Data Engineering, vol. 33, no. 6, pp. 2349–2368, 2019

    work page 2019

  32. [32]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning. PMLR, 2021, pp. 8748–8763

    work page 2021

  33. [33]

    E. J. Candes, D. L. Donohoet al.,Curvelets: A surprisingly effective nonadaptive representation for objects with edges. Department of Statistics, Stanford University Stanford, CA, USA, 1999

    work page 1999

  34. [34]

    Fast Fourier transforms: A tutorial review and a state of the art,

    P. Duhamel and M. Vetterli, “Fast Fourier transforms: A tutorial review and a state of the art,”Signal processing, vol. 19, no. 4, pp. 259–299, 1990

    work page 1990

  35. [35]

    Dictionary learning algorithms for sparse representa- tion,

    K. Kreutz-Delgado, J. F. Murray, B. D. Rao, K. Engan, T.-W. Lee, and T. J. Sejnowski, “Dictionary learning algorithms for sparse representa- tion,”Neural computation, vol. 15, no. 2, pp. 349–396, 2003

    work page 2003

  36. [36]

    S. A. Broughton and K. Bryan,Discrete Fourier analysis and wavelets: applications to signal and image processing. John Wiley & Sons, 2018

    work page 2018

  37. [37]

    Empirical Orthogo- nal Functions and related techniques in atmospheric science: A review,

    A. Hannachi, I. T. Jolliffe, and D. B. Stephenson, “Empirical Orthogo- nal Functions and related techniques in atmospheric science: A review,” International Journal of Climatology, vol. 27, no. 9, pp. 1119–1152, 2007

    work page 2007

  38. [38]

    Generalized discriminant analysis using a kernel approach,

    G. Baudat and F. Anouar, “Generalized discriminant analysis using a kernel approach,”Neural computation, vol. 12, no. 10, pp. 2385–2404, 2000

    work page 2000

  39. [39]

    Blind separation of sources, part i: An adap- tive algorithm based on neuromimetic architecture,

    C. Jutten and J. Herault, “Blind separation of sources, part i: An adap- tive algorithm based on neuromimetic architecture,”Signal processing, vol. 24, no. 1, pp. 1–10, 1991

    work page 1991

  40. [40]

    The isomap algorithm and topological stability,

    M. Balasubramanian and E. L. Schwartz, “The isomap algorithm and topological stability,”Science, vol. 295, no. 5552, pp. 7–7, 2002

    work page 2002

  41. [41]

    A kernel method for canonical correlation analysis

    S. Akaho, “A kernel method for canonical correlation analysis,”arXiv preprint cs/0609071, 2006

    work page internal anchor Pith review Pith/arXiv arXiv 2006

  42. [42]

    Kernel maximum autocorrelation factor and minimum noise fraction transformations,

    A. A. Nielsen, “Kernel maximum autocorrelation factor and minimum noise fraction transformations,”IEEE Transactions on Image Process- ing, vol. 20, no. 3, pp. 612–624, 2010

    work page 2010

  43. [43]

    Kernel partial least squares regression in reproducing kernel Hilbert space,

    R. Rosipal and L. J. Trejo, “Kernel partial least squares regression in reproducing kernel Hilbert space,”Journal of Machine Learning Research, vol. 2, no. Dec, pp. 97–123, 2001

    work page 2001

  44. [44]

    Classification of hyperspectral images with regularized linear discriminant analysis,

    T. V . Bandos, L. Bruzzone, and G. Camps-Valls, “Classification of hyperspectral images with regularized linear discriminant analysis,” IEEE Transactions on Geoscience and Remote Sensing, vol. 47, no. 3, pp. 862–873, 2009

    work page 2009

  45. [45]

    An introduction to locally linear embedding,

    L. K. Saul and S. T. Roweis, “An introduction to locally linear embedding,”unpublished. Available at: http://www. cs. toronto. edu/˜ roweis/lle/publications. html, 2000

    work page 2000

  46. [46]

    Semisupervised alignment of manifolds,

    J. Ham, D. Lee, and L. Saul, “Semisupervised alignment of manifolds,” inInternational Workshop on Artificial Intelligence and Statistics. PMLR, 2005, pp. 120–127. 21

    work page 2005

  47. [47]

    Prithvi-eo-2.0: A versatile multi-temporal foundation model for Earth observation applications,

    D. Szwarcman, S. Roy, P. Fraccaro, Þ. E. Gíslason, B. Blumenstiel, R. Ghosal, P. H. de Oliveira, J. L. de Sousa Almeida, R. Sedona, Y . Kang, S. Chakraborty, S. Wang, C. Gomes, A. Kumar, M. Truong, D. Godwin, H. Lee, C.-Y . Hsu, A. A. Asanjan, B. Mujeci, D. Shidham, T. Keenan, P. Arevalo, W. Li, H. Alemohammad, P. Olofsson, C. Hain, R. Kennedy, B. Zadrozn...

    work page

  48. [48]

    Romain Thoreau, Laurent Risser, Véronique Achard, Béatrice Berthelot, and Xavier Briottet

    [Online]. Available: https://arxiv.org/abs/2412.02732

    work page arXiv

  49. [49]

    A survey on multidimensional scaling,

    N. Saeed, H. Nam, M. I. U. Haq, and D. B. Muhammad Saqib, “A survey on multidimensional scaling,”ACM Computing Surveys (CSUR), vol. 51, no. 3, pp. 1–25, 2018

    work page 2018

  50. [50]

    Graph embedding and extensions: A general framework for dimensionality reduction,

    S. Yan, D. Xu, B. Zhang, H.-J. Zhang, Q. Yang, and S. Lin, “Graph embedding and extensions: A general framework for dimensionality reduction,”IEEE Transactions on Pattern Analysis and Machine Intel- ligence, vol. 29, no. 1, pp. 40–51, 2006

    work page 2006

  51. [51]

    A transfor- mation for ordering multispectral data in terms of image quality with implications for noise removal,

    A. A. Green, M. Berman, P. Switzer, and M. D. Craig, “A transfor- mation for ordering multispectral data in terms of image quality with implications for noise removal,”IEEE Transactions on Geoscience and Remote Sensing, vol. 26, no. 1, pp. 65–74, 1988

    work page 1988

  52. [52]

    Deep unsupervised embedding for remotely sensed images based on spatially augmented momentum contrast,

    J. Kang, R. Fernandez-Beltran, P. Duan, S. Liu, and A. J. Plaza, “Deep unsupervised embedding for remotely sensed images based on spatially augmented momentum contrast,”IEEE Transactions on Geoscience and Remote Sensing, vol. 59, no. 3, pp. 2598–2610, 2021

    work page 2021

  53. [53]

    Nonnegative matrix factorization: A comprehensive review,

    Y .-X. Wang and Y .-J. Zhang, “Nonnegative matrix factorization: A comprehensive review,”IEEE Transactions on knowledge and data engineering, vol. 25, no. 6, pp. 1336–1353, 2012

    work page 2012

  54. [55]

    A beginner’s guide to partial least squares analysis,

    M. Haenlein and A. M. Kaplan, “A beginner’s guide to partial least squares analysis,”Understanding statistics, vol. 3, no. 4, pp. 283–297, 2004

    work page 2004

  55. [56]

    An introduction to the proper orthogonal decomposi- tion,

    A. Chatterjee, “An introduction to the proper orthogonal decomposi- tion,”Current science, pp. 808–817, 2000

    work page 2000

  56. [57]

    The self-organizing map,

    T. Kohonen, “The self-organizing map,”Proceedings of the IEEE, vol. 78, no. 9, pp. 1464–1480, 1990

    work page 1990

  57. [58]

    Singular spectrum analysis in nonlinear dynamics, with applications to paleoclimatic time series,

    R. Vautard and M. Ghil, “Singular spectrum analysis in nonlinear dynamics, with applications to paleoclimatic time series,”Physica D: Nonlinear Phenomena, vol. 35, no. 3, pp. 395–424, 1989

    work page 1989

  58. [59]

    Visualizing data using t-sne

    L. Van der Maaten and G. Hinton, “Visualizing data using t-sne.” Journal of machine learning research, vol. 9, no. 11, 2008

    work page 2008

  59. [60]

    Probabilistic principal component analysis,

    M. E. Tipping and C. M. Bishop, “Probabilistic principal component analysis,”Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 61, no. 3, pp. 611–622, 1999

    work page 1999

  60. [61]

    The varimax criterion for analytic rotation in factor analysis,

    H. F. Kaiser, “The varimax criterion for analytic rotation in factor analysis,”Psychometrika, vol. 23, no. 3, p. 187–200, 1958

    work page 1958

  61. [62]

    Dimensionality reduction of multimodal labeled data by local fisher discriminant analysis

    M. Sugiyama, “Dimensionality reduction of multimodal labeled data by local fisher discriminant analysis.”Journal of Machine Learning Research, vol. 8, no. 5, 2007

    work page 2007

  62. [63]

    Fast discrete curvelet transforms,

    E. Candes, L. Demanet, D. Donoho, and L. Ying, “Fast discrete curvelet transforms,”multiscale modeling & simulation, vol. 5, no. 3, pp. 861– 899, 2006

    work page 2006

  63. [64]

    Camps-Valls and L

    G. Camps-Valls and L. Bruzzone, Eds.,Kernel methods for remote sensing data analysis. UK: Wiley & Sons, Dec 2009

    work page 2009

  64. [65]

    Universal kernels

    C. A. Micchelli, Y . Xu, and H. Zhang, “Universal kernels.”Journal of Machine Learning Research, vol. 7, no. 12, 2006

    work page 2006

  65. [66]

    Signal- to-noise ratio in reproducing kernel Hilbert spaces,

    L. Gómez-Chova, R. Santos-Rodríguez, and G. Camps-Valls, “Signal- to-noise ratio in reproducing kernel Hilbert spaces,”Pattern Recogni- tion Letters, vol. 112, pp. 75–82, 2018

    work page 2018

  66. [67]

    Learning repre- sentations by back-propagating errors,

    D. E. Rumelhart, G. E. Hinton, and R. J. Williams, “Learning repre- sentations by back-propagating errors,”Nature, vol. 323, no. 6088, pp. 533–536, 1986

    work page 1986

  67. [68]

    Deep representation learning: Fundamentals, technolo- gies, applications, and open challenges,

    A. Payandeh, K. T. Baghaei, P. Fayyazsanavi, S. B. Ramezani, Z. Chen, and S. Rahimi, “Deep representation learning: Fundamentals, technolo- gies, applications, and open challenges,”IEEE Access, vol. 11, pp. 137 621–137 659, 2023

    work page 2023

  68. [69]

    Self-supervised learning in remote sensing: A review,

    Y . Wang, C. M. Albrecht, N. A. A. Braham, L. Mou, and X. X. Zhu, “Self-supervised learning in remote sensing: A review,”IEEE Geoscience and Remote Sensing Magazine, vol. 10, no. 4, pp. 213– 247, 2022

    work page 2022

  69. [70]

    A simple frame- work for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple frame- work for contrastive learning of visual representations,” inProceed- ings of the 37th International Conference on Machine Learning, ser. ICML’20. JMLR.org, 2020

    work page 2020

  70. [71]

    SatCLIP: Global, general-purpose location embeddings with satellite imagery,

    K. Klemmer, E. Rolf, C. Robinson, L. Mackey, and M. Rußwurm, “SatCLIP: Global, general-purpose location embeddings with satellite imagery,”arXiv preprint arXiv:2311.17179, 2023

    work page arXiv 2023

  71. [72]

    A Survey of Large Language Models

    W. X. Zhao, K. Zhou, J. Li, T. Tang, X. Wang, Y . Hou, Y . Min, B. Zhang, J. Zhang, Z. Dong, Y . Du, C. Yang, Y . Chen, Z. Chen, J. Jiang, R. Ren, Y . Li, X. Tang, Z. Liu, P. Liu, J.-Y . Nie, and J.-R. Wen, “A survey of large language models,” 2025. [Online]. Available: https://arxiv.org/abs/2303.18223

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  72. [73]

    Foundation models defining a new era in vision: A survey and outlook,

    M. Awais, M. Naseer, S. Khan, R. M. Anwer, H. Cholakkal, M. Shah, M.-H. Yang, and F. S. Khan, “Foundation models defining a new era in vision: A survey and outlook,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 47, no. 4, pp. 2245–2264, 2025

    work page 2025

  73. [74]

    CROMA: Remote sensing representations with contrastive radar-optical masked autoencoders,

    A. Fuller, K. Millard, and J. R. Green, “CROMA: Remote sensing representations with contrastive radar-optical masked autoencoders,” inThirty-seventh Conference on Neural Information Processing Systems, 2023. [Online]. Available: https://openreview.net/forum?id= ezqI5WgGvY

    work page 2023

  74. [75]

    Panopticon: Advancing any-sensor foundation models for Earth observation,

    L. Waldmann, A. Shah, Y . Wang, N. Lehmann, A. J. Stewart, Z. Xiong, X. X. Zhu, S. Bauer, and J. Chuang, “Panopticon: Advancing any-sensor foundation models for Earth observation,”

    work page

  75. [76]

    J., X IONG , Z., Z HU, X

    [Online]. Available: https://arxiv.org/abs/2503.10845

    work page arXiv

  76. [77]

    Contrastive and non-contrastive self-supervised learning recover global and local spectral embedding methods,

    R. Balestriero and Y . LeCun, “Contrastive and non-contrastive self-supervised learning recover global and local spectral embedding methods,” inAdvances in Neural Information Processing Systems, A. H. Oh, A. Agarwal, D. Belgrave, and K. Cho, Eds., 2022. [Online]. Available: https://openreview.net/forum?id=jQgsZDspz5h

    work page 2022

  77. [78]

    Compression of multispectral remote sensing images using clustering and spectral reduction,

    A. Kaarna, P. Zemcik, H. Kalviainen, and J. Parkkinen, “Compression of multispectral remote sensing images using clustering and spectral reduction,”IEEE transactions on Geoscience and Remote Sensing, vol. 38, no. 2, pp. 1073–1082, 2000

    work page 2000

  78. [79]

    Remote sensing image compression based on high-frequency and low-frequency components,

    S. Xiang and Q. Liang, “Remote sensing image compression based on high-frequency and low-frequency components,”IEEE Transactions on Geoscience and Remote Sensing, 2024

    work page 2024

  79. [80]

    The JPEG 2000 still image compression standard,

    A. Skodras, C. Christopoulos, and T. Ebrahimi, “The JPEG 2000 still image compression standard,”IEEE Signal processing magazine, vol. 18, no. 5, pp. 36–58, 2001

    work page 2000

  80. [81]

    Progressive 3-D coding of hyperspectral images based on JPEG2000,

    B. Penna, T. Tillo, E. Magli, and G. Olmo, “Progressive 3-D coding of hyperspectral images based on JPEG2000,”IEEE Geoscience and Remote Sensing Letters, vol. 3, no. 1, pp. 125–129, 2006

    work page 2006

Showing first 80 references.