Quality Assessment of Papers Using Hyperspectral Imaging and Non-Imaging Techniques: A Review

Abstract

The preservation of both ancient and modern paper documents involves analytical procedures that can determine their status without physical destruction. This review assesses the use of Hyperspectral Non-Imaging Spectroscopy (HNIS), alongside Hyperspectral Imaging (HSI) and other spectroscopic methods, in the identification of paper composition, quality and deterioration. This work offers insights into the detection of chemical and structural variations like acidity, oxidation, fungal activity, and additive presence in non-contact analysis based on studies conducted between 2001 and 2025. HNIS offers a high spectral resolution and sensitivity to small chemical differences, and is suitable to delicate artefacts, whereas HSI can map large areas complementary to HNIS. The implementation of machine learning to predict degradation, classify, and estimate chronologically is also discussed in the article. Some of the most notable issues determined include the lack of common spectral reference collections, inadequate combination of real-time environmental information, and inadequate variance of training data. The offered HNIS-ML integration framework provides a scalable, non-invasive method of early damage detection and informed preservation planning in museums, libraries and archives.

Introduction

The text reviews the development of papermaking and the use of Hyperspectral Non-Imaging Spectroscopy (HNIS) and Hyperspectral Imaging (HSI) for non-destructive paper analysis and conservation.

Paper has been used for communication, documentation, education, and creative expression for centuries. Its composition has evolved from natural fibres such as linen, cotton, hemp, flax, and mulberry bark to modern wood pulp, recycled fibres, bamboo, and other alternative materials. Changes in fibres, sizing agents, fillers, and bleaching methods have significantly affected paper quality, durability, acidity, and ageing behaviour.

Evolution of papermaking

• 1800s: Paper was mainly made from linen and cotton rags, with gelatin sizing and limited fillers. These papers were generally durable and stable.
• 1900s: Wood pulp became the dominant raw material. Rosin-alum sizing, chlorine bleaching, and various fillers became common. However, acidic materials and chemical residues made many papers more prone to yellowing and deterioration.
• 2000s: Wood pulp and recycled fibres became common, with synthetic sizing agents such as AKD and ASA, calcium carbonate and titanium dioxide fillers, and ECF/TCF bleaching processes. Papers generally became brighter, stronger, and more alkaline.
• 2020s: Papermaking increasingly incorporates bamboo, agricultural waste, recycled materials, nanocellulose, and bio-based materials. Modern processes aim for improved sustainability, although contaminants such as PFAS and microplastics in recycled pulp remain concerns.

Need for non-destructive analysis

Traditional paper evaluation methods, including pH testing, tensile-strength testing, and fibre microscopy, can provide useful information but may be time-consuming, labour-intensive, or unsuitable for large collections. This creates a need for rapid, non-contact techniques, particularly for valuable historical documents, books, manuscripts, and museum collections.

HSI and HNIS

The text focuses particularly on Hyperspectral Imaging (HSI) and Hyperspectral Non-Imaging Spectroscopy (HNIS).

• HSI collects both spatial and spectral information, producing a hyperspectral data cube. It is useful for mapping degradation, locating damaged regions, recovering faded text, and examining large document surfaces.
• HNIS collects detailed spectral information from a specific point or small area. It provides high-resolution chemical information and is particularly useful for identifying paper components and localized degradation.

Both techniques are non-contact and non-destructive and can help identify:

• Fibre types such as cotton, flax, and wood pulp.
• Additives and sizing materials such as gelatin, rosin, and calcium carbonate.
• Cellulose, lignin, and other chemical components.
• Oxidation, acidity, fungal growth, and fibre degradation.
• Differences between historical and modern paper.
• Early stages of chemical and environmental deterioration.

Role of machine learning

Combining spectroscopy with machine learning can improve the classification and prediction of paper properties. Machine-learning models can identify different levels of degradation and help predict characteristics such as paper composition, ageing, and even publication date.

Literature findings

The literature shows a gradual progression from conventional spectroscopy to advanced hyperspectral analysis and machine learning. Important research has demonstrated that:

• NIR spectroscopy can monitor artificial ageing and identify paper and ink components.
• HSI can improve the readability of faded or damaged historical documents.
• Spectroscopy can estimate chemical and mechanical properties such as pH and tensile strength without damaging historical paper.
• HSI can identify pigments, underdrawings, degradation regions, and other material characteristics in cultural heritage objects.
• Machine learning can improve hyperspectral classification, feature extraction, and prediction accuracy.
• Recent research has demonstrated that NIR spectroscopy combined with machine learning can estimate the publication year of historical books with high accuracy.

Conclusion

This review has discussed Hyperspectral Non-Imaging Spectroscopy as a non-destructive method of measuring the quality and the state of paper materials such as the rare and historically important documents. The offered methodology, including sample collection and categorisation, spectral data acquisition, preprocessing, feature extraction, machine learning classification, and iterative validation, offers an organised and scalable approach to the assessment of paper degradation under no physical influence. HNIS has shown to be quite appropriate in paper assessment due to ability to identify variations in chemical makeup such as oxidation, acidity and type of fibre by the reaction of paper to light of varying wavelengths. Together with sophisticated methods of data analysis and pattern recognition, these spectral properties allow making an effective classification of paper based on its age and condition, providing conservators with a better foundation to make long-term preservation decisions. The review also indicates the significance of building a detailed spectral reference database, finding early degradation predictors, and using predictive tools to conserve proactively. Combined with the capabilities enabled by new data analysis methods and the combination of HNIS with machine learning, these capabilities form a basis to further advance the preservation of paper heritage and its protection to the future generations.

References

[1] D. Hunter, Papermaking: The History and Technique of an Ancient Craft. New York, NY, USA: Dover Publications, 1978. [2] W. J. Barrow, Deterioration of Book Stock: Causes and Remedies. Richmond, VA, USA: Barrow Research Laboratory, 1974. [3] R. Giorgi et al., \"A new method for paper deacidification based on calcium hydroxide dispersed in nonaqueous media,\" Studies in Conservation, vol. 47, no. sup3, pp. 69-73, 2002. [4] M. Strlic, I. Kralj Cigic, and J. Kolar, \"Evaluation of paper stability using pH and DP measurements,\" International Journal for the Preservation of Library and Archival Material, vol. 26, no. 3, pp. 181-189, 2005. [5] P. Calvini and A. Gorassini, \"FTIR and micro-Raman spectroscopy of cellulose-based materials,\" Cellulose Chemistry and Technology, vol. 40, no. 9-10, pp. 595-605, 2006. [6] M. Strlic et al., \"Non-destructive characterisation of historical paper using spectroscopic methods,\" Analytica Chimica Acta, vol. 623, no. 1, pp. 119-127, 2008. [7] P. Gupta and R. Singh, \"Spectral analysis of aging and deterioration in historical papers,\" Journal of Paper Conservation, vol. 22, no. 3, pp. 124-135, 2021. [8] R. M. Smith, \"The historical use of gelatin in paper sizing,\" Journal of the American Institute for Conservation, vol. 34, no. 2, pp. 125-138, 1995. [9] P. Mishra and S. Mishra, \"Application of hyperspectral sensing for paper material classification,\" Sensors and Materials, vol. 31, no. 6, pp. 1803-1814, 2019. [10] S. H. Zeronian and H. L. Inglesby, \"Bleaching of cellulosic fibres,\" Cellulose Chemistry and Technology, vol. 32, no. 5-6, pp. 365-394, 1998. [11] M. Hubbe, A. P. Singh, and S. Rojas, \"Recycled fiber quality and impacts on paper properties: A review,\" BioResources, vol. 2, no. 4, pp. 739-788, 2007. [12] S. K. Bajpai, Bleaching and Deinking of Recycled Paper, 2nd ed. Amsterdam, Netherlands: Elsevier, 2014. [13] J. K. M. Kalliala and A. Talvenmaa, \"Trends in paper industry: Non-wood fibres and sustainability,\" Nordic Pulp & Paper Research Journal, vol. 30, no. 4, pp. 547-556, 2015. [14] J. C. Prata et al., \"Occurrence of microplastics in paper products and their environmental implications,\" Environmental Pollution, vol. 288, p. 117771, 2021. [15] A. F. H. Goetz, \"Three decades of hyperspectral remote sensing of the Earth: A personal view,\" Remote Sensing of Environment, vol. 113, no. S1, pp. S5-S16, 2009. [16] M. R. Dickinson et al., \"Hyperspectral imaging for manuscript analysis: Recovering lost texts and identifying materials,\" Heritage Science, vol. 6, no. 1, p. 63, 2018. [17] R. P. Baraldi et al., \"Applications of hyperspectral imaging on artworks: From technical documentation to monitoring,\" Applied Physics A, vol. 121, no. 3, pp. 879-886, 2015. [18] J. Zhao, L. Wang, and X. Li, \"Review of hyperspectral analysis techniques in cultural heritage documentation,\" Spectrochimica Acta Part A, vol. 210, pp. 268-281, 2019. [19] A. R. Tobjork, \"Photooxidation and chemiluminescence in cellulosic paper: Mechanistic insights,\" Cellulose, vol. 21, no. 4, pp. 2349-2362, 2014. [20] H. Pretzel and T. Fearn, \"Spectroscopic tools for linking environmental conditions and paper degradation in archives,\" Studies in Conservation, vol. 65, no. sup1, pp. S85-S94, 2020. [21] M. Ali, B. Sablier, M. Vignon, and P. Viallat, \"Near-infrared spectroscopy for monitoring artificial aging of paper,\" in Proc. International Conference on Conservation Science, 2001. [22] C. Chabries, \"Enhancing the legibility of faint or burnt documents,\" Journal of Imaging Science and Technology, vol. 47, no. 5, pp. 452-458, 2003. [23] M. Attas, L. Cloutis, M. Collins, and D. Goltz, \"Hyperspectral imaging for enhanced visualization of historical documents,\" Journal of Cultural Heritage, vol. 5, no. 3, pp. 285-290, 2004. [24] G. Bonifazzi, \"Hyperspectral imaging for artwork inspection,\" in Proc. Optical Methods for Art Inspection, 2005. [25] J. Havermans, \"Near-infrared spectroscopy as a tool for the identification of paper and inks in conservation research,\" Restaurator, vol. 26, no. 3, pp. 189-197, 2005. [26] L. Scholten et al., \"Hyperspectral imaging for plant disease detection: A study on spectral variability,\" Computers and Electronics in Agriculture, vol. 50, no. 2, pp. 209-222, 2006. [27] C. Fischer and I. Kakoulli, \"Multispectral and hyperspectral imaging technologies in conservation: Current research and potential applications,\" Reviews in Conservation, vol. 7, pp. 3-16, 2006. [28] P. Rychly, L. Hradilova, and J. Kucera, \"Chemiluminescence and photoinduced chemiluminescence studies of madder lakes on paper,\" Journal of Photochemistry and Photobiology A: Chemistry, vol. 183, no. 1-2, pp. 1-9, 2006. [29] M. Strlic et al., \"Non-destructive characterization of historical paper using NIR spectroscopy and chemometrics,\" Analytical Chemistry, vol. 80, no. 14, pp. 5252-5258, 2008. [30] L. Barbin, D. Sun, and P. Allen, \"Hyperspectral imaging for quality assessment of food products: A review,\" Critical Reviews in Food Science and Nutrition, vol. 52, no. 10, pp. 827-835, 2012. [31] J. M. Bioucas-Dias et al., \"Hyperspectral unmixing overview: Geometrical, statistical, and sparse regression-based approaches,\" IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 5, no. 2, pp. 354-379, 2012. [32] G. Boldrini, M. Zanon, and L. Ballerini, \"Hyperspectral imaging: Best practices, performance and pitfalls for in-line and on-line applications,\" Journal of Imaging, vol. 8, no. 2, pp. 1-14, 2012. [33] T. Vo-Dinh, \"Multivariate optical computing for spectroscopic analysis,\" Analytical Chemistry, vol. 86, no. 3, pp. 1670-1675, 2014. [34] A. Cucci, J. K. Delaney, and M. Picollo, \"Reflectance hyperspectral imaging for investigation of works of art: Old master paintings and illuminated manuscripts,\" Accounts of Chemical Research, vol. 49, no. 10, pp. 2070-2079, 2016. [35] P. Calvini and A. Gorassini, \"FTIR and Raman spectroscopy techniques for non-invasive analysis of historical documents,\" Restaurator, vol. 38, no. 3, pp. 235-255, 2017. [36] M. Giacometti, F. Martelli, and M. G. Lognoli, \"Enhancement of old documents with hyperspectral imaging,\" Journal of Cultural Heritage, vol. 25, pp. 83-90, 2017. [37] J. Sandak, A. Sandak, and D. Pauliny, \"Hyperspectral imaging of weathered wood samples in transmission mode,\" Wood Science and Technology, vol. 51, pp. 135-151, 2017. [38] R. Gewali, S. T. Monteiro, and E. Saber, \"Machine learning based hyperspectral image analysis: A survey,\" arXiv preprint arXiv:1802.08701, 2018. [39] L. Lehnert, M. Meyer, and U. Schmidtlein, \"The hsdar package: Hyperspectral data analysis in R,\" Journal of Statistical Software, vol. 89, no. 1, pp. 1-23, 2019. [40] A. Smith and D. Taylor, \"Non-imaging spectroscopy for material analysis of historical artifacts,\" Journal of Cultural Heritage, vol. 29, pp. 54-62, 2018. [41] A. De Viguerie, L. Bertolotti, and E. Catherinot, \"Hyperspectral imaging for the analysis of photographic emulsions,\" Applied Physics A, vol. 125, no. 3, p. 187, 2019. [42] G. Gabrieli, M. Fontana, and F. Mazzinghi, \"Diffuse reflectance spectroscopy for in-situ characterization of artists\' materials,\" Spectrochimica Acta Part A, vol. 222, p. 117226, 2019. [43] H. Yang, Y. Wang, and H. Li, \"Rapid determination of holocellulose and lignin in wood by near infrared spectroscopy and kernel extreme learning machine,\" Cellulose, vol. 26, pp. 331-342, 2019. [44] S. Tonazzini, L. Bedini, and E. Salerno, \"Multispectral and hyperspectral document restoration: A review,\" Pattern Recognition Letters, vol. 129, pp. 60-67, 2019. [45] A. De Viguerie, A. Walter, and J. Mazel, \"Principal component analysis of spectral data: A contribution to the knowledge of the materials constituting works of art,\" Spectrochimica Acta Part B, vol. 159, p. 105654, 2019. [46] H. Wang, J. Yang, and Q. Zhao, \"Remote sensing in the preservation of cultural heritage materials,\" Journal of Cultural Heritage, vol. 40, pp. 94-102, 2019. [47] R. Barrett and L. Stullich, \"Overview of hyperspectral techniques in heritage conservation,\" Journal of Imaging, vol. 6, no. 4, p. 28, 2020. [48] A. Jones, F. Thompson, and C. Reed, \"Understanding multispectral imaging of cultural heritage: Best practices for MSI analysis,\" Studies in Conservation, vol. 65, no. 4, pp. 224-238, 2020. [49] L. Orazi, \"Mid-wave infrared reflectography and thermography for the study of ancient books: A review,\" Heritage Science, vol. 8, no. 1, pp. 1-15, 2020. [50] F. Pretzel and T. Fearn, \"Environmental impacts on archival papers: A spectroscopic approach,\" Studies in Conservation, vol. 65, no. 3, pp. 150-162, 2020. [51] Y. Rasti, H. R. Yazdi, and G. Camps-Valls, \"Feature extraction for hyperspectral image classification: A review,\" IEEE Geoscience and Remote Sensing Magazine, vol. 8, no. 4, pp. 63-99, 2020. [52] V. Ramanathan and R. Iyer, \"Advances in hyperspectral non-imaging for material characterization,\" Journal of Cultural Heritage, vol. 53, pp. 75-89, 2022. [53] Y. Wang, Z. Liu, and J. Zhang, \"Tensor decomposition techniques for hyperspectral image restoration,\" IEEE Transactions on Image Processing, vol. 31, pp. 5872-5886, 2022. [54] M. Strlic, T. Trafela, and B. Pihlar, \"Non-destructive dating of paper books using near-infrared spectroscopy and machine learning,\" Heritage Science, vol. 11, no. 1, pp. 1-12, 2023. [55] G. Wan et al., \"Hyperspectral imaging technology for nondestructive identification of quality deterioration in fruits and vegetables: A review,\" Critical Reviews in Food Science and Nutrition, 2025. doi: 10.1080/10408398.2025.2487134. [56] V. Dhangar, D. Barode, P. Dhole, A. Kshirsagar, and B. Gawali, \"Exploring the spectral database of food samples using ASD Field Spec 4 spectroradiometer,\" SSRN 4900729, 2024. [57] V. D. Dhangar et al., \"Assessment of adulteration from food products using ASD Field Spec 4,\" in Proc. 2023 IEEE International Conference on Contemporary Computing and Communications (InC4), vol. 1. IEEE, 2023.

Copyright

Copyright © 2026 Dinesh M. Barode, Rupali S. Awhad, Vijay D. Dhangar, Seema S. Kawathekar. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.