Abstract
Pulp feedstock components (including extractives, lignin and holocellulose) are major products of the pulp and paper industry, and their proportion in production feedstock (e.g., wood) greatly impact pulp yield and pulp production costs. Near-infrared (NIR) spectroscopy offers a potential solution for rapid characterizing physical and chemical properties of woody biomass. However, NIR models are highly specific to the physical form and species of the samples used in the production process. Traditional calibration transfer (CT) methods usually fail to adapt NIR models to significant differences between spectra. Adversarial transfer learning (ATL) strategies, an emerging algorithm in computer imaging, align source and target distributions by introducing adversarial mechanisms. For the first time, deep ATL architecture, coupled with NIR spectroscopy (collected from ASD LabSpecPro spectrometer), was applied to quantitatively predict the pulpwood feedstock component content for adapting models of different physical forms (wood blocks, wood meals) and tree species to each other without the need for constructing a new model. It was discovered that ATL methods, including domain adversarial neural networks (DANN), domain separation networks (DSN), and dynamic adversarial adaptation networks (DAAN) not only remove the differences between wood block and wood meal spectral data sources (i.e., domains), but also make desired predictions for cross-species domain adaptation. The robustness and stability of the transferred models exceed those of traditional CT and transfer learning methods. Our results suggest that ATL models could be effectively adapted to polymorphic, multi-species pulp feedstock data, and can be extended to the detection and inversion of other feedstock properties.
Graphical abstract
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Availability of data and materials
Not applicable.
References
Alencastre-Miranda M, Johnson RM, Krebs HI (2021) Convolutional neural networks and transfer learning for quality inspection of different sugarcane varieties. IEEE Trans Industr Inf 17(2):787–794. https://doi.org/10.1109/tii.2020.2992229
Bagwari S, Choudhary K, Raikwar S, Nijhawan R, Kumar S et al (2022) An edge filter based approach of neural style transfer to the image stylization. IEEE Access 10:104612–104621. https://doi.org/10.1109/access.2022.3204040
Belt T, Awais M, Makela M (2022) Chemical characterization and visualization of progressive brown rot decay of wood by near infrared imaging and multivariate analysis. Front Plant Sci. https://doi.org/10.3389/fpls.2022.940745
Bousmalis K, Trigeorgis G, Silberman N, Krishnan D, Erhan D (2016) Domain separation networks. Paper presented at the Proceedings of the 30th international conference on neural information processing systems, Barcelona, Spain
Chen YY, Wang ZB (2019) Cross components calibration transfer of NIR spectroscopy model through PCA and weighted ELM-based TrAdaBoost algorithm. Chemometr Intell Lab Syst. https://doi.org/10.1016/j.chemolab.2019.103824
Chen JH, Jiang DP, Zhang YZ, Wang KQ (2022) Research on data migration modeling method for bending strength of solid wood based on SWCSS-GFK-SVM. Spectrosc Spectral Anal 42(5):1471–1477. https://doi.org/10.3964/j.issn.1000-0593(2022)05-1471-07
Diaz VF, Roger JM, Saeys W (2022) Unsupervised dynamic orthogonal projection. An efficient approach to calibration transfer without standard samples. Anal Chimica Acta. https://doi.org/10.1016/j.aca.2022.340154
Espinoza JA, Hodge GR, Dvorak WS (2012) The potential use of near infrared spectroscopy to discriminate between different pine species and their hybrids. J Near Infrared Spectrosc 20(4):437–447. https://doi.org/10.1255/jnirs.1006
Fackler K, Schwanninger M (2010) Polysaccharide degradation and lignin modification during brown rot of spruce wood: a polarised Fourier transform near infrared study. J Near Infrared Spectrosc 18(6):403–416. https://doi.org/10.1255/jnirs.901
Ferreira GWD, Roque JV, Soares EMB, Silva IR, Silva EF et al (2018) Temporal decomposition sampling and chemical characterization of eucalyptus harvest residues using NIR spectroscopy and chemometric methods. Talanta 188:168–177. https://doi.org/10.1016/j.talanta.2018.05.073
Ganin Y, Ustinova E, Ajakan H, Germain P, Larochelle H et al (2016) Domain-adversarial training of neural networks. J Mach Learn Res 17(1):2096–2030
He X, Chen YS, Ghamisi P (2020) Heterogeneous transfer learning for hyperspectral image classification based on convolutional neural network. IEEE Trans Geosci Remote Sens 58(5):3246–3263. https://doi.org/10.1109/tgrs.2019.2951445
Hein PRG, Clair B, Brancheriau L, Chaix G (2010) Predicting microfibril angle in Eucalyptus wood from different wood faces and surface qualities using near infrared spectra. J Near Infrared Spectrosc 18(6):455–464. https://doi.org/10.1255/jnirs.905
Horikawa Y (2017) Assessment of cellulose structural variety from different origins using near infrared spectroscopy. Cellulose 24(12):5313–5325. https://doi.org/10.1007/s10570-017-1518-0
Huang AM, Jiang ZH, Li GY (2007) Determination of holocellulose and lignin content in Chinese fir by near infrared spectroscopy. Spectrosc Spectral Anal 27(7):1328–1331
Ji W, Rossel RAV, Shi Z (2015) Accounting for the effects of water and the environment on proximally sensed vis-NIR soil spectra and their calibrations. Eur J Soil Sci 66(3):555–565. https://doi.org/10.1111/ejss.12239
Jiang ZH, Huang AM (2006) Water in wood and its near infrared spectroscopic analysis. Spectrosc Spectral Anal 26(8):1464–1468
Jiang ZH, Yang Z, So CL, Hse CY (2007) Rapid prediction of wood crystallinity in Pinus elliotii plantation wood by near-infrared spectroscopy. J Wood Sci 53(5):449–453. https://doi.org/10.1007/s10086-007-0883-y
Jiang W, Han G, Via BK, Tu M, Liu W et al (2014) Rapid assessment of coniferous biomass lignin–carbohydrates with near-infrared spectroscopy. Wood Sci Technol 48(1):109–122. https://doi.org/10.1007/s00226-013-0590-3
Jin XL, Chen XL, Shi CH, Li M, Guan YJ et al (2017) Determination of hemicellulose, cellulose and lignin content using visible and near infrared spectroscopy in Miscanthus sinensis. Biores Technol 241:603–609. https://doi.org/10.1016/j.biortech.2017.05.047
Kelley SS, Rials TG, Snell R, Groom LH, Sluiter A (2004) Use of near infrared spectroscopy to measure the chemical and mechanical properties of solid wood. Wood Sci Technol 38(4):257–276. https://doi.org/10.1007/s00226-003-0213-5
Kentsch S, Caceres MLL, Serrano D, Roure F, Diez Y (2020) Computer vision and deep learning techniques for the analysis of drone-acquired forest images, a transfer learning study. Remote Sensing. https://doi.org/10.3390/rs12081287
Khiari R, Mhenni MF, Belgacem MN, Mauret E (2010) Chemical composition and pulping of date palm rachis and Posidonia oceanica - A comparison with other wood and non-wood fibre sources. Biores Technol 101(2):775–780. https://doi.org/10.1016/j.biortech.2009.08.079
Kothiyal V, Raturi A, Jaideep, & Dubey, Y. M. (2014) Enhancing the applicability of near infrared spectroscopy for estimating specific gravity of green timber from Eucalyptus tereticornis by developing composite calibration using both radial and tangential face of wood. Eur J Wood Wood Prod 72(1):11–20. https://doi.org/10.1007/s00107-013-0742-z
Kothiyal V, Jaideep B, S., Ginwal, H. S., & Gupta, S. (2015) Multi-species NIR calibration for estimating holocellulose in plantation timber. Wood Sci Technol 49(4):769–793. https://doi.org/10.1007/s00226-015-0720-1
Lahtinen P, Liukkonen S, Pere J, Sneck A, Kangas H (2014) A comparative study of fibrillated fibers from different mechanical and chemical pulps. BioResources 9(2):2115–2127
Lavoie FB, Robert G, Langlet A, Gosselin R (2023) Calibration transfer by likelihood maximization: A standard-free approach capable of handling non-overlapping wavelength ranges. Chemometr Intell Lab Syst. https://doi.org/10.1016/j.chemolab.2023.104766
Li Y, Sharma M, Altaner C, Cookson LJ (2020a) An approach to quantify natural durability of Eucalyptus bosistoana by near infrared spectroscopy for genetic selection. Ind Crops Prod 154:112676. https://doi.org/10.1016/j.indcrop.2020.112676
Li YJ, Sharma M, Altaner C, Cookson LJ (2020) An approach to quantify natural durability of Eucalyptus bosistoana by near infrared spectroscopy for genetic selection. Ind Crops Prod. https://doi.org/10.1016/j.indcrop.2020.112676
Li CT, Chen HZ, Zhang YY, Hong SY, Ai W et al (2022) Improvement of NIR prediction ability by dual model optimization in fusion of NSIA and SA methods. Spectrochimica Acta Part A Mol Biomol Spectrosc. https://doi.org/10.1016/j.saa.2022.121247
Liang L, Wei L, Fang G, Xu F, Deng Y et al (2020) Prediction of holocellulose and lignin content of pulp wood feedstock using near infrared spectroscopy and variable selection. Spectrochim Acta Part A Mol Biomol Spectrosc 225:117515. https://doi.org/10.1016/j.saa.2019.117515
Liu J, Li XY, Jin R, Xu SM, Ku J (2015) Extending hyperspectral detecting model of pH in fresh pork to new breeds. Spectrosc Spectral Anal 35(7):1973–1979. https://doi.org/10.3964/j.issn.1000-0593(2015)07-1973-07
Liu YC, Li YY, Peng YK, Yan S, Zhao XL et al (2021) Non-destructive and rapid detection of the internal chemical composition of granules samples by spectral transfer. Chemometr Intell Lab Syst. https://doi.org/10.1016/j.chemolab.2020.104174
Lopez F, Alfaro A, Garcia MM, Diaz MJ, Calero AM et al (2004) Pulp and paper from tagasaste (Chamaecytisus proliferus LF ssp palmensis). Chem Eng Res Des 82(A8):1029–1036. https://doi.org/10.1205/0263876041580730
Loureiro BA, Arriel TG, Ramalho FMG, Hein PRG, Trugilho PF (2022) NIR-based models for estimating selected physical and chemical wood properties from fast-growing plantations. Iforest Biogeosci For 15:372–380
Malachowska E, Lipkiewicz A, Niemczyk M, Dubowik M, Boruszewski P et al (2021) Influences of Fiber and Pulp properties on papermaking ability of cellulosic pulps produced from alternative fibrous raw materials. J Natl Fibers 18(11):1751–1761. https://doi.org/10.1080/15440478.2019.1697994
Mishra P, Nikzad-Langerodi R (2021) A brief note on application of domain-invariant PLS for adapting near-infrared spectroscopy calibrations between different physical forms of samples. Talanta 232:122461. https://doi.org/10.1016/j.talanta.2021.122461
Munnaf MA, Mouazen AM (2022) Removal of external influences from on-line vis-NIR spectra for predicting soil organic carbon using machine learning. Catena. https://doi.org/10.1016/j.catena.2022.106015
Neto RD, Ramalho FMG, Costa LR, Hein PRG (2021) Estimating hardness and density of wood and charcoal by near-infrared spectroscopy. Wood Sci Technol 55(1):215–230. https://doi.org/10.1007/s00226-020-01232-y
Pan SJ, Tsang IW, Kwok JT, Yang QA (2011) Domain adaptation via transfer component analysis. IEEE Trans Neural Netw 22(2):199–210. https://doi.org/10.1109/tnn.2010.2091281
Pereira LSA, Carneiro MF, Botelho BG, Sena MM (2016) Calibration transfer from powder mixtures to intact tablets: A new use in pharmaceutical analysis for a known tool. Talanta 147:351–357. https://doi.org/10.1016/j.talanta.2015.10.006
Porep JU, Kammerer DR, Carle R (2015) On-line application of near infrared (NIR) spectroscopy in food production. Trends Food Sci Technol 46(2):211–230. https://doi.org/10.1016/j.tifs.2015.10.002
Preys S, Roger JM, BoUlet JC (2008) Robust calibration using orthogonal projection and experimental design Application to the correction of the light scattering effect on turbid NIR spectra. Chemometric Intell Lab Syst 91(1):28–33. https://doi.org/10.1016/j.chemolab.2007.10.007
Samat A, Gamba P, Abuduwaili J, Liu SC, Miao ZL (2016) Geodesic flow kernel support vector machine for hyperspectral image classification by unsupervised subspace feature transfer. Remote Sensing. https://doi.org/10.3390/rs8030234
Sandak A, Sandak J, Waliszewska B, Zborowska M, Mleczek M (2017) Selection of optimal conversion path for willow biomass assisted by near infrared spectroscopy. Iforest-Biogeosci For 10:506–514. https://doi.org/10.3832/ifor1987-010
Saray SN, Tahmoresnezhad J (2022) Iterative joint classifier and domain adaptation for visual transfer learning. Int J Mach Learn Cybern 13(4):947–961. https://doi.org/10.1007/s13042-021-01428-z
Schimleck LR, Kube PD, Raymond CA, Michell AJ, French J (2006) Extending near infrared reflectance(NIR) pulp yield calibrations to new sites and species. J Wood Chem Technol 26(4):299–311. https://doi.org/10.1080/02773810601076683
Schimleck L, Antony F, Dahlen J, Moore J (2018) Wood and fiber quality of plantation-grown conifers: a summary of research with an emphasis on loblolly and radiata pine. Forests 9(6):298. https://doi.org/10.3390/f9060298
Schwanninger M, Rodrigues JC, Fackler K (2011) A review of band assignments in near infrared spectra of wood and wood components. J Near Infrared Spectrosc 19(5):287–308. https://doi.org/10.1255/jnirs.955
Sheng Q, Santos-Rivera M, Ouyang X, Kouba AJ, Vance CK (2022) Near-infrared spectroscopy and mode cloning (NIR-MC) for in-situ analysis of crude protein in bamboo. Remote Sens 14(6):1302
Skotare T, Nilsson D, Xiong SJ, Geladi P, Trygg J (2019) Joint and unique multiblock analysis for integration and calibration transfer of NIR instruments. Anal Chem 91(5):3516–3524. https://doi.org/10.1021/acs.analchem.8b05188
TAPPI (2006) Acid insoluble lignin in wood and pulp. In Technical association of the pulp and paper industry (Vol. T222 om-98).
TAPPI (2007) Solvent extractives of wood and pulp. In Technical association of the pulp and paper industry (Vol. T204 cm-97).
Tsuchikawa S, Kobori H (2015) A review of recent application of near infrared spectroscopy to wood science and technology. J Wood Sci 61(3):213–220. https://doi.org/10.1007/s10086-015-1467-x
Wang JD, Chen YQ, Feng WJ, Yu H, Huang MY et al (2020) Transfer learning with dynamic distribution adaptation. ACM Trans Intell Syst Technol. https://doi.org/10.1145/3360309
Wang HP, Chen P, Dai JW, Liu D, Li JY et al (2022) Recent advances of chemometric calibration methods in modern spectroscopy: algorithms, strategy, and related issues. Trac-Trends Anal Chem. https://doi.org/10.1016/j.trac.2022.116648
Wang YZ, Xiang JY, Tang Y, Chen WJ, Xu YJ (2022b) A review of the application of near-infrared spectroscopy (NIRS) in forestry. Appl Spectrosc Rev 57(4):300–317. https://doi.org/10.1080/05704928.2021.1875481
Wu LW, Liu PS, Yuan YH, Liu SY, Zhang YN (2023) Context-aware style learning and content recovery networks for neural style transfer. Inf Process Manag. https://doi.org/10.1016/j.ipm.2023.103265
Xu EC, Zhou YJ (2007) Synergistic effects between chemical mechanical pulps and chemical pulps from hardwoods. Tappi J 6(11):4–9
Yu C, Wang J, Chen Y, Huang M (2019) Transfer learning with dynamic adversarial adaptation network. Paper presented at the 2019 IEEE International Conference on Data Mining (ICDM)
Zhang Y, Hui J, Qin Q, Sun Y, Zhang T et al (2021) Transfer-learning-based approach for leaf chlorophyll content estimation of winter wheat from hyperspectral data. Remote Sens Environ 267:112724. https://doi.org/10.1016/j.rse.2021.112724
Zhang XX, Chen XY, Xiong ZX, Siesler HW, Liang L (2022a) Transfer of a calibration model for the prediction of lignin in pulpwood among four portable near infrared spectrometers. J Near Infrared Spectrosc 30(4):208–218. https://doi.org/10.1177/09670335221110013
Zhang Z, Li Y, Li Y (2022b) Prediction approach of larch wood density from visible-near-infrared spectroscopy based on parameter calibrating and transfer learning. Front Plant Sci 13:1006292. https://doi.org/10.3389/fpls.2022.1006292
Zhang ZY, Li YX, Li CX, Wang ZC, Chen Y (2022) Algorithm of stability-analysis-based feature selection for NIR calibration transfer. Sensors. https://doi.org/10.3390/s22041659
Zhao X, Liu D, Ma L, Liu Q, Chen K et al (2022) Deep CNN model based on serial-parallel structure optimization for four-class motor imagery EEG classification. Biomed Signal Process Control 72:103338. https://doi.org/10.1016/j.bspc.2021.103338
Zhu L, Chen YS, Ghamisi P, Benediktsson JA (2018) Generative adversarial networks for hyperspectral image classification. IEEE Trans Geosci Remote Sens 56(9):5046–5063. https://doi.org/10.1109/tgrs.2018.2805286
Acknowledgments
The authors received supports from the Innovation Foundation for Doctoral Program of Forestry Engineering of Northeast Forestry University (LYGC202114), the Fundamental Research Funds for the Central Universities (Grant Number 2572022AW45) and the Applied Technology Research and Development Plan of Heilongjiang Province (Grant Numbers GA19C006, GA21C030).
Funding
This work was supported by the Innovation Foundation for Doctoral Program of Forestry Engineering of Northeast Forestry University (LYGC202114), the Fundamental Research Funds for the Central Universities (Grant Number 2572022AW45) and the Applied Technology Research and Development Plan of Heilongjiang Province (Grant Numbers GA19C006, GA21C030).
Author information
Authors and Affiliations
Contributions
All authors contributed to the study conception and design. ZZ: Conceptualization, methodology, software, formal analysis, visualization, funding acquisition, writing-original draft. HZ: conceptualization, methodology, software, formal analysis, visualization, funding acquisition, writing-original draft. YL: Funding acquisition, writing-original draft, writing- review & editing. RAW: Writing- review & editing. RP: investigation, data curation. YC: formal analysis, investigation. XL: investigation.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Consent for publication
The authors declare no sensitive or personal information is reported within this work.
Ethical approval and consent to participate
The authors declare no sensitive or personal information is reported within this work.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Zhang, Z., Zhong, H., Li, Y. et al. Predicting components of pulpwood feedstock for different physical forms and tree species using NIR spectroscopy and transfer learning. Cellulose 31, 551–566 (2024). https://doi.org/10.1007/s10570-023-05619-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10570-023-05619-5