Differentiation of apple varieties using leaf hyperspectral spectroscopy and supervised modeling

DOI: 10.52945/rac.v39i.2404
Published 2026-10-26

Authors

Abstract

Varietal traceability in apple production is essential for cultivar certification, origin authentication, and the protection of breeding programs, but it still relies on destructive and costly laboratory-based methods. This study aimed to determine the feasibility of leaf hyperspectral reflectance spectroscopy, integrated with supervised machine learning models, as a complementary tool for apple cultivar identification. A total of 540 leaves were collected from a homogeneous orchard containing six commercial apple cultivars (Fuji, Gala, Gala Gui, Galídia, Luiza, and Isadora), generating 1,620 spectral measurements acquired using a FieldSpec® 4 spectroradiometer coupled with a Plant Probe. After removing spectral regions with a low signal-to-noise ratio, spectral preprocessing, principal component analysis (PCA), and Support Vector Classification (SVC, RBF kernel) were applied using stratified cross-validation and an independent hold-out validation set (20%). PERMANOVA revealed a significant effect of cultivar on the overall spectral signature (p = 0.001). The best-performing pipeline (CR + DER + MSC + PCA + SVC) achieved a mean macro-F1 score of 0.74 during cross-validation. Independent validation resulted in an overall accuracy of 0.67. Fuji achieved 100% classification accuracy, whereas Luiza reached 83%, while cultivars belonging to the Gala lineage showed higher levels of confusion (44–56%), reflecting their close genetic relationship. Spectral importance was concentrated in the VIS, red-edge, and early NIR regions, which are associated with leaf pigments and mesophyll structure, whereas SWIR contributed little to cultivar discrimination. The results demonstrate that leaf optical traits can discriminate between apple cultivars, particularly those with distinct genealogies, supporting the proposed approach as a proof of concept. However, the achieved accuracy remains insufficient for commercial certification and traceability, especially for distinguishing closely related clones and somatic mutants.

Author Biographies

  • Dr. Elisângela B. Silva

    PhD in Soil Science (UFRGS – Brazil). Master’s degree in Civil Engineering (UFSC – Brazil). Researcher at the Santa Catarina Agricultural Research and Rural Extension Company (Epagri), at the Center for Environmental Resources and Hydrometeorology Information. E-mail: elisangelasilva@epagri.sc.gov.br

    • Conceptualization
    • Data Curation
    • Formal Analysis
    • Funding Acquisition
    • Investigation
    • Methodology
    • Resources
    • Software
    • Supervision
    • Validation
    • Visualization
    • Writing – Original Draft Preparation
    • Writing – Review & Editing
    Competing Interests

    The authors declare that there are no conflicts of interest in this work.

  • Dr. Gabriel B. Leite

    PhD in Plant Physiology (UBP – France). Master’s degree in Agronomy (UFPel – Brazil). Researcher at the Santa Catarina Agricultural Research and Rural Extension Company (Epagri), at the Center for Environmental Resources and Hydrometeorology Information. E-mail: gabriel@epagri.sc.gov.br

     

    • Conceptualization
    • Investigation
    • Project Administration
    • Writing – Original Draft Preparation
    • Writing – Review & Editing
    • Validation
    Competing Interests

    The authors declare that there are no conflicts of interest in this work.

  • Dr. Marcus V. Kvitschal

    Doctor in Genetics and Breeding (UEM – Brazil). Master’s degree in Agronomy (UEM – Brazil). Researcher at the Santa Catarina Agricultural Research and Rural Extension Company (Epagri), at the Caçador Experimental Station. E-mail: marcusvinicius@epagri.sc.gov.br

    • Conceptualization
    • Investigation
    • Writing – Review & Editing
    Competing Interests

    The authors declare that there are no conflicts of interest in this work.

  • Dr. Valeria Pohlmann

    Doctor in Agronomy (UFPel – Brazil). Master’s degree in Agronomy (UFSM – Brazil). Researcher at the Santa Catarina Agricultural Research and Rural Extension Company (Epagri), at the Center for Environmental Resources and Hydrometeorology Information. E-mail: valeriapohlmann@epagri.sc.gov.br

    • Investigation
    • Writing – Review & Editing
    Competing Interests

    The authors declare that there are no conflicts of interest in this work.

  • Dr. Cristina Pandolfo

    Doctor in Genetic Resources (UFSC – Brazil). Master’s degree in Crop Science (UFRGS – Brazil). Researcher at the Santa Catarina Agricultural Research and Rural Extension Company (Epagri), at the Center for Environmental Resources and Hydrometeorology Information. E-mail: cristina@epagri.sc.gov.br

    • Funding Acquisition
    • Resources
    • Writing – Review & Editing
    Competing Interests

    The authors declare that there are no conflicts of interest in this work.

References

ANALYTICAL SPECTRAL DEVICES INC. ASD Accessories User Manual. Document 600544. Boulder, CO: ASD Inc., 2007.

ARRUDA, D.C.; DUCATI, J.R.; HOFF, Rosemary; BELLOLI, T.F.; THUM, A.B. Proximal hyperspectral analysis in grape leaves for region and variety identification. Ciência Rural, v. 53, n.12, e20220313, 2023. Disponível em: https://doi: https://10.1590/0103-8478cr20220313. Acesso em: 15 maio 2025.

BLACKBURN, G.A. Hyperspectral remote sensing of plant pigments. Journal of Experimental Botany, Oxford, v.58, n. 4, p. 855–867, 2007. Disponível em: http://doi.org/10.1093/jxb/erl123. Acesso em: 14 ago. 2025.

CARTER, G.A.; KNAPP, A.K. Leaf optical properties in higher plants: linking spectral characteristics to stress and chlorophyll concentration. American Journal of Botany, v.88, n.4, p. 677–684, 2001. Disponível em: https://doi.org/10.2307/2657068. Acesso em: 13 jul. 2025.

CRUZ, J.; GERGOV, G.; TARAPOULOUZI, M.; KIRILOVA, E.; KOSTADINOVA, O.; NIKOLOVA, K. Classification of apple varieties by VIS-NIR diffuse reflectance spectroscopy and chemometrics. Bulgarian Chemical Communications, v.56, p.106-115, 2024. Disponível em: https://10.34049/bcc.56.D.S1P38. Acesso em: 10 jun. 2025.

CUSHNAHAN, T.A.; GRAFTON, M.C.E.; PEARSON, D.; RAMILAN, T. Hyperspectral data can differentiate species and cultivars of C3 and C4 Turf despite measurable diurnal variation. Remote Sensing, v.16, p.3142, 2024. Disponível em: https://doi.org/10.3390/rs16173142. Acesso em: 14 jul. 2025.

DENARDI, F.; KVITSCHAL, M.V.; HAWERROTH, M.C.; PETRI, J.L. Cultivares de macieira desenvolvidos pela Epagri. Florianópolis: Epagri, 2023. 97p. (Epagri. Boletim Técnico, 211).

FALCIONI, R.; GONÇALVES, J.V.F.; OLIVEIRA, K.M.D.; ANTUNES, W.C.; NANNI, M. R. VIS-NIR-SWIR hyperspectroscopy combined with data mining and machine learning for classification of predicted chemometrics of green lettuce. Remote Sensing, v. 14, p. 6330, 2022. Disponível em: https://doi.org/10.3390/rs14246330. Acesso em: 24 out. 2024.

FALCIONI, R.; ANTUNES, W.C.; DEMATTÊ, J.A.M.; NANNI, M.R. Reflectance spectroscopy for the classification and prediction of pigments in agronomic crops. Plants, v.12, p.2347, 2023. Disponível em: https://doi.org/10.3390/plants12122347. Acesso em: 24 out. 2024.

FAORO, I.D. (org.). Características agronômicas de cultivares de macieira. Florianópolis, SC: Epagri, 2025. 211p. Disponível em: https:////publicacoes.epagri.sc.gov.br/livros/article/view/2200. Acesso em: 24 out. 2025.

GREWAL, R.; KASANA, S.S.; KASANA, G. Machine learning and deep learning techniques for spectral spatial classification of hyperspectral images: A comprehensive survey. Electronics, v.12, p. 488, 2023. Disponível em: https://doi.org/10.3390/electronics12020488. Acesso em: 28 jan. 2025.

HENNESSY, A.; CLARKE, K.; LEWIS, M. Hyperspectral classification of plants: A review of waveband selection generalisability. Remote Sens., v.12, p.113, 2020. Disponível em: http://doi:10.3390/rs12010113. Acesso em: 12 jan. 2025.

HUANG, Y.; YANG, Y.; SUN, Y.; ZHOU, H.; CHEN, K. Identification of Apple Varieties Using a Multichannel Hyperspectral Imaging System. Sensors, 20, 5120, 2020. Disponível em: https://doi:10.3390/s20185120. Acesso em: 15 maio 2023.

KVITSCHAL, M.V.; DENARDI, F.; HAWERROTH, M.C.; PETRI, J.L. Cultivares de macieira desenvolvidos pela Epagri. Florianópolis, 2023. 97p. (Epagri. Boletim Técnico, 211).

LI, J. Genetic diversity and conservation strategies of apple germplasm resources. Tree Genetics and Molecular Breeding, v.14, 1-7, 2025. Disponível em: https://doi.org/10.5376/tgmb.2024.14.0001. Acesso em: 10 mar. 2026.

LI, Y.; XU, X.; WU, W.; ZHU, Y.; GAO, L.; JIANG, X.; MENG, Y.; Yang, G.J.; XUE, H. Hyperspectral estimation of chlorophyll content in grapevine based on feature selection and GA-BP. Scientific Reports, v.15, p.8029, 2025. Disponível em: https://doi.org/10.1038/s41598-024-84977-x. Acesso em: 01 fev. 2026.

LIU, W.; CHEN, Z.; JIANG, S.; WANG, Y.; FANG, H.; ZHANG, Z.; CHEN, X.; WANG, N.. Research Progress on Genetic Basis of Fruit Quality Traits in Apple (Malus × domestica). Front. Plant Sci. 13:918202, 2022. Disponível em: https://doi.org/10.3390/plants12122347. Acesso em: 15 fev. 2023.

MIRZAEI, M.; MAROFI, S.; ABBASI, M.; SOLGI, E.; KARIMI, R.; VERRELST, J. Scenario-based discrimination of common grapevine varieties using in-field hyperspectral data in the western of Iran. International Journal of Applied Earth Observation and Geoinformation, v.80, p.26-37, 2019. Disponível em: https://doi: https://10.1016/j.jag.2019.04.002 . Acesso em: 15 maio 2023.

PASQUINI, C. Near infrared spectroscopy: fundamentals, practical aspects and analytical applications. Journal of the Brazilian Chemical Society, São Paulo, v. 14, p. 198-219, 2003. Disponível em: https://doi.org/10.1590/S0103-50532003000200006. Acesso em: 25 jun. 2024.

RINNAN, Å.; van den BERG, F.; ENGELSEN, S.B. Review of the most common pre-processing techniques for near-infrared spectra. Trends in Analytical Chemistry, Amsterdam, v. 28, p. 1201, 2009. Disponível em: https://doi.org/10.1016/j.trac.2009.07.007. Acesso em: 19 mar. 2017.

ROCCHINI, D.; SANTOS, M.J.; USTIN, S.L.; FÉRET, J-B.; ASNER, G.P.; BEIERKUHNLEIN, C.; DALPONTE, M.; FEILHAUER, H.; FOODY, G.M.; GELLER, G.N.; GILLESPIE, T.W.; HE, K.S.; KLEIJN, D.; LEITÃO, P.J.; MALAVASI, M.; MOUDRÝ, V.; MÜLLEROVÁ, Ja; NAGENDRA, H.; NORMAND, S.; RICOTTA, C.; SCHAEPMAN, M.E.; SCHMIDTLEIN, S.; SKIDMORE, A.K.; ŠÍMOVÁ, P.; TORRESANI, M.; TOWNSEND, P.A.; TURNER, W.; VIHERVAARA, P.; WEGMANN, M.; LENOIR, J. The spectral species concept in living color. Journal of Geophysical Research, Biogeosciences, 127, e2022JG007026, 2022. Disponível em:https://doi. org/10.1029/2022JG007026. Acesso em: 28 jan. 2025.

SAVITZKY, A.; GOLAY, M.J. E. Smoothing and differentiation of data by simplified least squares procedures. Analytical Chemistry, v. 36, n. 8, p.1627–1639, 1964. Disponível em: https://doi.org/10.1021/ac60214a047. Acesso em: 19 mar. 2017.

SEVINDIK, E.; YANGILAR, F.; ATASAGUN, B.; SOFYALIOĞLU, E.; ŞIMŞEK, S.A.; ÇAYIR, M.E.; VIVODÍK, M. Genetic diversity and structure analysis of 'Ak Sakı' and 'Kara Sakı' apple cultivars growing in Erzincan/Türkiye. Biologica Nyssana, 15, p.61-68, 2024. Disponível em: https://doi.org/10.5281/zenodo.14258237. Acesso em: 29 jan. 2026.

SOUZA, R.P; BRIGHENTI, A.F.; GOULART JUNIOR, R. Análise da evolução da produção de maçã em Santa Catarina no período de 2014 a 2021. Agropecuária Catarinense, Florianópolis, v.37, n.3, p. 40-45, 2024. Disponível em: http://doi: https://doi.org/10.52945/rac.v37i3.1801. Acesso em: 01 jun. 2026.

SU, W-H. Advanced Machine Learning in Point Spectroscopy, RGB- and Hyperspectral-Imaging for Automatic Discriminations of Crops and Weeds: A Review. Smart cities, v.3, p.767-792, 2020. Disponível em: http://doi.org/10.3390/smartcities3030039. Acesso em: 01 abr. 2025.

YAMASHITA, H.; SONOBE, R.; HIRONO, Y..; MORITA, A.; IKKA, T. Dissection of hyperspectral reflectance to estimate nitrogen and chlorophyll contents in tea leaves based on machine learning algorithms. Scientific Reports, v.10, p.17360, 2020. Disponível em: https://doi.org/10.1038/s41598-020-73745-2. Acesso em: 29 jun. 2025.

YAMASHITA, H.; SONOBE, R.; HIRONO, Y..; MORITA, A.; IKKA, T. Potential of spectroscopic analyses for non-destructive estimation of tea quality-related metabolites in fresh new leaves. Scientific Reports, v.11, p.4169, 2021. Disponível em: https://doi.org/10.1038/s41598-021-83847-0. Acesso em: 29 jun. 2025.

ZHANG, J.; WANG, C.; YUAN, L.; LIU, P.; ZHANG, Y.; WU, K.. Construction of a plant spectral library based on an optimised feature selection method. Biosystems Engineering, v.195, p.1–16, 2020. Disponível em: https://doi.org/10.1016/j.biosystemseng.2020.04.008. Acesso em: 13 jul. 2025.

ZHANG, C.; VALENTE, J.; KOOISTRA, L.; GUO, L.; Wang, W. Orchard management with small unmanned aerial vehicles: a survey of sensing and analysis approaches. Precision Agriculture, v.22, p.2007–2052, 2021. Disponível em: https://doi.org/10.1007/s11119-021-09813-y. Acesso em: 03 set. 2025.

ZHANG, Y.; Li, X.; WANG, C.; ZHANG, R.; JIN, L.; HE, Z.; TIAN, S.; WU, K.; WANG, F. PROSPECT-PMP+: Simultaneous Retrievals of Chlorophyll a and b, Carotenoids and Anthocyanins in the Leaf Optical Properties Model. Sensors, v.22, p.3025, 2022. Disponível em: https://doi.org/10.3390/s22083025. Acesso em: 03 set. 2025.

Metrics

50
Views
19
Downloads

Keywords

  • Apple varieties
  • Hyperspectral spectroscopy
  • Leaf reflectance
  • Machine learning
  • Varietal traceability