View in EDS HTML Full Text PDF Full Text

Soil Suborder Discrimination Using Machine Learning Is Improved by SWIR Imaging Compared with Full VIS–NIR–SWIR Spectra.

Saved in:
Bibliographic Details
Title: Soil Suborder Discrimination Using Machine Learning Is Improved by SWIR Imaging Compared with Full VIS–NIR–SWIR Spectra.
Authors: Haubert, Daiane de Fatima da Silva1 (AUTHOR), Vedana, Nicole Ghinzelli1,2 (AUTHOR), Mendonça, Weslei Augusto1 (AUTHOR), de Oliveira, Karym Mayara1,2 (AUTHOR), de Oliveira, Caio Almeida1 (AUTHOR), Gonçalves, João Vitor Ferreira1 (AUTHOR), Demattê, José Alexandre M.2 (AUTHOR), de Oliveira, Roney Berti1 (AUTHOR), Reis, Amanda Silveira1 (AUTHOR), Falcioni, Renan1 (AUTHOR) rfalcioni2@uem.br, Nanni, Marcos Rafael1 (AUTHOR)
Source: Remote Sensing. Mar2026, Vol. 18 Issue 6, p898. 26p.
Subjects: Hyperspectral imaging systems, Soil classification, Spectrum analysis, Machine learning
Geographic Terms: Paraná (Brazil : State), Brazil
Abstract: Highlights: What are the main findings? Suborder discrimination was strongest for subsurface horizons and was consistently improved when the models used SWIR-only inputs relative to full VIS–NIR–SWIR spectra. Horizon classification within profiles achieved high accuracy, with misclassifications concentrated between vertically adjacent horizons, indicating gradual spectral transitions along the profile. What are the implications of the main findings? The results highlight the SWIR bands as the most informative spectral region for taxonomic discrimination, supporting the use of SWIR-focused hyperspectral remote-sensing workflows and feature engineering for soil class mapping under bare-soil conditions. Because optical remote sensing samples are mainly surface, surface-only classification is feasible but more uncertain; robust mapping should integrate multitemporal bare-soil compositing, targeted field calibration, and uncertainty-aware outputs to flag ambiguous or transitional areas. Rapid, standardised discrimination of soil taxonomic units remains challenging when relying solely on conventional field descriptions and laboratory analyses, particularly at high sampling densities. This study evaluated whether proximal spectroscopy and hyperspectral imaging can support the classification of Brazilian Soil Classification System (SiBCS) suborders and pedogenetic horizons when surface and subsurface spectra are treated separately. Six intact soil monoliths (0.12 × 1.60 m) were collected in Paraná State, southern Brazil, representing one Organossolo (Ooy), three Latossolos (LVd, LVd1, and LVd2) and two Argissolos (PVAd and PVd). For each monolith, 800 spectra were acquired per sensor with a non-imaging VIS–NIR–SWIR spectroradiometer (350–2500 nm), and 800 spectra per sensor per monolith were extracted from the SWIR hyperspectral images (1200–2450 nm). Principal component analysis (PCA) was used to summarise spectral variability, and supervised classification was performed via k-nearest neighbours, random forest, decision tree and gradient boosting for suborders (10-fold cross-validation), and a neural network was used for within-profile horizon classification. PCA indicated that most of the spectral variance was captured by a dominant axis, with clearer separation among suborders in the SWIR space than in the full VIS–NIR–SWIR range. With respect to suborder classification, subsurface spectra outperformed surface spectra, and SWIR outperformed VIS–NIR–SWIR: the best accuracies were 0.96 for subsurface SWIR (gradient boosting; AUC = 0.99; MCC = 0.95) and 0.89 for surface SWIR (k-nearest neighbours; AUC = 0.98; MCC = 0.87). Within-profile horizon classification via VIS–NIR–SWIR achieved accuracies of 0.84–0.97 with the Neural Network, with most misclassifications occurring between adjacent horizons. Overall, subsurface SWIR information provided the most reliable basis for taxonomic discrimination, whereas horizon classification was feasible but reflected gradual spectral transitions along the profile. [ABSTRACT FROM AUTHOR]
Copyright of Remote Sensing is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
Text:
  Availability: 1
Header DbId: egs
DbLabel: Engineering Source
An: 192591388
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Soil Suborder Discrimination Using Machine Learning Is Improved by SWIR Imaging Compared with Full VIS–NIR–SWIR Spectra.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Haubert%2C+Daiane+de+Fatima+da+Silva%22">Haubert, Daiane de Fatima da Silva</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Vedana%2C+Nicole+Ghinzelli%22">Vedana, Nicole Ghinzelli</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Mendonça%2C+Weslei+Augusto%22">Mendonça, Weslei Augusto</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22de+Oliveira%2C+Karym+Mayara%22">de Oliveira, Karym Mayara</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22de+Oliveira%2C+Caio+Almeida%22">de Oliveira, Caio Almeida</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Gonçalves%2C+João+Vitor+Ferreira%22">Gonçalves, João Vitor Ferreira</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Demattê%2C+José+Alexandre+M%2E%22">Demattê, José Alexandre M.</searchLink><relatesTo>2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22de+Oliveira%2C+Roney+Berti%22">de Oliveira, Roney Berti</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Reis%2C+Amanda+Silveira%22">Reis, Amanda Silveira</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Falcioni%2C+Renan%22">Falcioni, Renan</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> rfalcioni2@uem.br</i><br /><searchLink fieldCode="AR" term="%22Nanni%2C+Marcos+Rafael%22">Nanni, Marcos Rafael</searchLink><relatesTo>1</relatesTo> (AUTHOR)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Remote+Sensing%22">Remote Sensing</searchLink>. Mar2026, Vol. 18 Issue 6, p898. 26p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Hyperspectral+imaging+systems%22">Hyperspectral imaging systems</searchLink><br /><searchLink fieldCode="DE" term="%22Soil+classification%22">Soil classification</searchLink><br /><searchLink fieldCode="DE" term="%22Spectrum+analysis%22">Spectrum analysis</searchLink><br /><searchLink fieldCode="DE" term="%22Machine+learning%22">Machine learning</searchLink>
– Name: SubjectGeographic
  Label: Geographic Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Paraná+%28Brazil+%3A+State%29%22">Paraná (Brazil : State)</searchLink><br /><searchLink fieldCode="DE" term="%22Brazil%22">Brazil</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Highlights: What are the main findings? Suborder discrimination was strongest for subsurface horizons and was consistently improved when the models used SWIR-only inputs relative to full VIS–NIR–SWIR spectra. Horizon classification within profiles achieved high accuracy, with misclassifications concentrated between vertically adjacent horizons, indicating gradual spectral transitions along the profile. What are the implications of the main findings? The results highlight the SWIR bands as the most informative spectral region for taxonomic discrimination, supporting the use of SWIR-focused hyperspectral remote-sensing workflows and feature engineering for soil class mapping under bare-soil conditions. Because optical remote sensing samples are mainly surface, surface-only classification is feasible but more uncertain; robust mapping should integrate multitemporal bare-soil compositing, targeted field calibration, and uncertainty-aware outputs to flag ambiguous or transitional areas. Rapid, standardised discrimination of soil taxonomic units remains challenging when relying solely on conventional field descriptions and laboratory analyses, particularly at high sampling densities. This study evaluated whether proximal spectroscopy and hyperspectral imaging can support the classification of Brazilian Soil Classification System (SiBCS) suborders and pedogenetic horizons when surface and subsurface spectra are treated separately. Six intact soil monoliths (0.12 × 1.60 m) were collected in Paraná State, southern Brazil, representing one Organossolo (Ooy), three Latossolos (LVd, LVd1, and LVd2) and two Argissolos (PVAd and PVd). For each monolith, 800 spectra were acquired per sensor with a non-imaging VIS–NIR–SWIR spectroradiometer (350–2500 nm), and 800 spectra per sensor per monolith were extracted from the SWIR hyperspectral images (1200–2450 nm). Principal component analysis (PCA) was used to summarise spectral variability, and supervised classification was performed via k-nearest neighbours, random forest, decision tree and gradient boosting for suborders (10-fold cross-validation), and a neural network was used for within-profile horizon classification. PCA indicated that most of the spectral variance was captured by a dominant axis, with clearer separation among suborders in the SWIR space than in the full VIS–NIR–SWIR range. With respect to suborder classification, subsurface spectra outperformed surface spectra, and SWIR outperformed VIS–NIR–SWIR: the best accuracies were 0.96 for subsurface SWIR (gradient boosting; AUC = 0.99; MCC = 0.95) and 0.89 for surface SWIR (k-nearest neighbours; AUC = 0.98; MCC = 0.87). Within-profile horizon classification via VIS–NIR–SWIR achieved accuracies of 0.84–0.97 with the Neural Network, with most misclassifications occurring between adjacent horizons. Overall, subsurface SWIR information provided the most reliable basis for taxonomic discrimination, whereas horizon classification was feasible but reflected gradual spectral transitions along the profile. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Remote Sensing is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=192591388
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.3390/rs18060898
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 26
        StartPage: 898
    Subjects:
      – SubjectFull: Hyperspectral imaging systems
        Type: general
      – SubjectFull: Soil classification
        Type: general
      – SubjectFull: Spectrum analysis
        Type: general
      – SubjectFull: Machine learning
        Type: general
      – SubjectFull: Paraná (Brazil : State)
        Type: general
      – SubjectFull: Brazil
        Type: general
    Titles:
      – TitleFull: Soil Suborder Discrimination Using Machine Learning Is Improved by SWIR Imaging Compared with Full VIS–NIR–SWIR Spectra.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Haubert, Daiane de Fatima da Silva
      – PersonEntity:
          Name:
            NameFull: Vedana, Nicole Ghinzelli
      – PersonEntity:
          Name:
            NameFull: Mendonça, Weslei Augusto
      – PersonEntity:
          Name:
            NameFull: de Oliveira, Karym Mayara
      – PersonEntity:
          Name:
            NameFull: de Oliveira, Caio Almeida
      – PersonEntity:
          Name:
            NameFull: Gonçalves, João Vitor Ferreira
      – PersonEntity:
          Name:
            NameFull: Demattê, José Alexandre M.
      – PersonEntity:
          Name:
            NameFull: de Oliveira, Roney Berti
      – PersonEntity:
          Name:
            NameFull: Reis, Amanda Silveira
      – PersonEntity:
          Name:
            NameFull: Falcioni, Renan
      – PersonEntity:
          Name:
            NameFull: Nanni, Marcos Rafael
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 15
              M: 03
              Text: Mar2026
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-print
              Value: 20724292
          Numbering:
            – Type: volume
              Value: 18
            – Type: issue
              Value: 6
          Titles:
            – TitleFull: Remote Sensing
              Type: main
ResultId 1