zhpsЖурнал прикладной спектроскопииZhurnal Prikladnoii Spektroskopii0514-7506B. I. Stepanov Institute of Physics of the National Academy of Scienceszhps-1444Research ArticleАННОТАЦИИ АНГЛОЯЗЫЧНЫХ СТАТЕЙABSTRACTS ENGLISH-LANGUAGE ARTICLESМетод неразрушающего определения содержания солей на фресках с использованием регистрации гиперспектрального изображения и многопараметрических методов его обработкиEstimation of Soluble Salt Concentration in Murals Based on Spectral Transformation and Feature Extraction ModellingGuoZ. Q.GuoZ. Q.

Пекин

Beijing

2108570021067@stu.bucea.edu.cnLyuS. Q.LyuS. Q.

Пекин

Beijing

HouМ. L.HouM. L.

Пекин

Beijing

Школа геоматики и городской пространственной информатики Пекинского университета строительства и архитектуры; Главная лаборатория тонкой реконструкции и мониторинга состояния архитектурного наследияSchool of Geomatics and Urban Spatial Informatics, Beijing University of Civil Engineering and Architecture; Beijing Key Laboratory for Architectural Heritage Fine Reconstruction & Health Monitoring202325112023905808808Copyright © Guo Z.Q., Lyu S.Q., Hou М.L., 20232023Guo Z.Q., Lyu S.Q., Hou М.L.Guo Z.Q., Lyu S.Q., Hou M.L.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://zhps.ejournal.by/jour/article/view/1444

Предложен метод быстрого и неразрушающего определения содержания солей в росписях с использованием гиперспектральных изображений. Спектральные данные образцов настенной росписи собраны с помощью спектрорадиометра и предварительно обработаны путем удаления точек излома с помощью сглаживания Савицкого–Голея. Необработанные спектры подвергались удалению континуума, логарифмированию обратной (LR) обработки, коррекции множественного рассеяния и преобразованию стандартной нормальной переменной в сочетании с дифференцированием первого и второго порядка для получения 15 преобразованных спектров различных форм. Спектры образцов при различных уровнях концентрации классифицированы, характеристические длины волн извлечены с использованием разделения набора образцов на основе совместного расстояния X–Y и алгоритма последовательного проецирования. Для сравнения использованы коэффициент корреляции Пирсона и переменная в проекции. Модели оценки концентрации соли разработаны с использованием частичной регрессии наименьших квадратов (PLSR), регрессии опорных векторов (SVR) и модели случайного леса (RF). Наклоны подгонки рассчитаны и сопоставлены. Показано, что интенсивность спектров отражения уменьшалась, а затем увеличивалась с ростом концентрации соли. Точность RF и SVR лучше, чем PLSR, для модели RF-LR-FD Rс2 = 0.9703, RMSE = 0.0466 и RPD = 16.8350.

Affected by temperature and humidity in the environment, salt crystals expand and accumulate on the surface of murals, causing the pigment layer to peel off, which damages the mural. A method for the rapid and nondestructive detection of salt content in murals using hyperspectral techniques is proposed. The spectral data from mural samples were collected with a spectroradiometer and preprocessed by removing breakpoints with Savitzky‒Golay smoothing. The raw spectra were subjected to continuum removal, logarithm of the reciprocal (LR) processing, multiple scattering correction, and standard normal variate transformation combined with first-order differentiation (FD) and second-order differentiation processing to obtain 15 transformed spectra of different forms. The spectra of samples at different concentration levels were classified, and the characteristic wavelengths were extracted using sample set partitioning based on the joint X–Y distance and successive projection algorithm. The pearson correlation coefficient and variable importance in the projection were used for comparison. Salt concentration estimation models were developed using partial least squares regression (PLSR), support vector regression (SVR), and a random forest (RF) model. The slopes of fit were calculated and compared. The results showed that the reflectance spectra decreased and then increased with increasing salt concentration. The accuracy of RF and SVR was better than that of PLSR, and the Rc2, RMSEc, and RPDc values of the RF-LR-FD model were 0.9703, 0.0466, and 16.8350, respectively. Spectral analysis combined with machine learning models has potential for the nondestructive detection of salt in murals.

разрушение фрескисодержание солейфункция спектрального преобразованиямашинное обучениеmural disruptionsalt contentspectral transformation functionmachine learningThis work was supported by the National Natural Science Foundation of China (Nos. 42171356 and 42171444) and the BUCEA Post Graduate Innovation Project (PG2022111)ReferencesJ. Cao, Z. Zhang, A. Zhao, et al., Herit Sci., 8, 7–20 (2020), doi: 10.1186/s40494-020-0355-x, https://heritagesciencejournal.springeropen.com/articles/10.1186/s40494-020-0355-x.J. Cao, Z. Zhang, A. Zhao, et al., Herit Sci., 8, 7–20 (2020), doi: 10.1186/s40494-020-0355-x, https://heritagesciencejournal.springeropen.com/articles/10.1186/s40494-020-0355-x.Z. Zheng, J. Nat. Disasters., 29, 158–163 (2020), doi: 10.13577/j.jnd.2020.0616, http://zrzh.paperonce.org/oa/darticle.aspx?type=view&id=20200616.Z. Zheng, J. Nat. Disasters., 29, 158–163 (2020), doi: 10.13577/j.jnd.2020.0616, http://zrzh.paperonce.org/oa/darticle.aspx?type=view&id=20200616.C. Liu, W. Sun, J. Huang, et al., Sci. Conserv. Archaeol., 32, 70–81 (2020), http://www.wwbhkgkx.net/wwbh/article/abstract/20200310?st=search.C. Liu, W. Sun, J. Huang, et al., Sci. Conserv. Archaeol., 32, 70–81 (2020), http://www.wwbhkgkx.net/wwbh/article/abstract/20200310?st=search.Y. Zhang, Z. Yu, L. Wang, et al., Dunhuang Res., 48, 148–155 (2021), doi: 10.13584/j.cnki.issn10004106.2021.01.020, http://www.dhyj.net.cn/index.php?m=content&c=index&a=show&catid=17&id=963.Y. Zhang, Z. Yu, L. Wang, et al., Dunhuang Res., 48, 148–155 (2021), doi: 10.13584/j.cnki.issn10004106.2021.01.020, http://www.dhyj.net.cn/index.php?m=content&c=index&a=show&catid=17&id=963.F. Wang, J. Gao, Y. Zha, ISPRS J. Photogramm. Rem. Sens., 136, 73–84 (2018), doi: 10.1016/j.isprsjprs.2017.12.003, https://www.sciencedirect.com/science/article/abs/pii/S0924271617303702?via%3Dihub.F. Wang, J. Gao, Y. Zha, ISPRS J. Photogramm. Rem. Sens., 136, 73–84 (2018), doi: 10.1016/j.isprsjprs.2017.12.003, https://www.sciencedirect.com/science/article/abs/pii/S0924271617303702?via%3Dihub.J. F. García-Martín, A. T. Badaró, D. F. Barbin, et al., Processes, 8, 823–832 (2020), doi: 10.3390/pr8070823, https://www.mdpi.com/2227-9717/8/7/823.J. F. García-Martín, A. T. Badaró, D. F. Barbin, et al., Processes, 8, 823–832 (2020), doi: 10.3390/pr8070823, https://www.mdpi.com/2227-9717/8/7/823.M. Amer, A. Tyler, T. Fouda, et al., Sci. Pap. Ser. Manag. Ec., 17, 47–56 (2017), https://managementjournal.usamv.ro/index.php/scientific-papers/1451-spectral-characteristics-forestimation-heavy-metals-accumulation-in-wheat-plants-and-grain-1451#spucontentCitation5.M. Amer, A. Tyler, T. Fouda, et al., Sci. Pap. Ser. Manag. Ec., 17, 47–56 (2017), https://managementjournal.usamv.ro/index.php/scientific-papers/1451-spectral-characteristics-forestimation-heavy-metals-accumulation-in-wheat-plants-and-grain-1451#spucontentCitation5.Z. E. Mashimbye, M. A. Cho, J. P. Nell, et al., Pedosphere, 22, 640–649 (2012), doi: 10.1016/S10020160(12)60049-6, https://www.sciencedirect.com/science/article/abs/pii/S1002016012600496?via%3Dihub.Z. E. Mashimbye, M. A. Cho, J. P. Nell, et al., Pedosphere, 22, 640–649 (2012), doi: 10.1016/S10020160(12)60049-6, https://www.sciencedirect.com/science/article/abs/pii/S1002016012600496?via%3Dihub.K. Tan, W. Ma, F. Wu, et al., Environ. Mon. Assess., 191, 446–459 (2019), doi: 10.1007/s10661-0197510-4, https://link.springer.com/article/10.1007/s10661-019-7510-4.K. Tan, W. Ma, F. Wu, et al., Environ. Mon. Assess., 191, 446–459 (2019), doi: 10.1007/s10661-0197510-4, https://link.springer.com/article/10.1007/s10661-019-7510-4.O. Mutanga, A. K. Skidmore., Int. J. Appl. Earth Obs. Geoinf., 5, 87–96 (2004), doi: 10.1016/j.jag.2004.01.001, https://www.sciencedirect.com/science/article/abs/pii/S0303243404000029?via%3Dihub.O. Mutanga, A. K. Skidmore., Int. J. Appl. Earth Obs. Geoinf., 5, 87–96 (2004), doi: 10.1016/j.jag.2004.01.001, https://www.sciencedirect.com/science/article/abs/pii/S0303243404000029?via%3Dihub.R. K. Galvão, M. C. Arauj, G. E. José et al., Talanta, 67, 736–740 (2005), doi: 10.1016/j.talanta.2005.03.025, https://www.sciencedirect.com/science/article/abs/pii/S003991400500192X?via%3Dihub.R. K. Galvão, M. C. Arauj, G. E. José et al., Talanta, 67, 736–740 (2005), doi: 10.1016/j.talanta.2005.03.025, https://www.sciencedirect.com/science/article/abs/pii/S003991400500192X?via%3Dihub.S. Wang, P. Han, G. Cui, et al., Spectrosc. Spectr. Anal., 39, 738–742 (2019), doi: 10.3964/j.issn.1000059303-0738-05, http://www.gpxygpfx.com/CN/10.3964/j.issn.1000-059303-0738-05.S. Wang, P. Han, G. Cui, et al., Spectrosc. Spectr. Anal., 39, 738–742 (2019), doi: 10.3964/j.issn.1000059303-0738-05, http://www.gpxygpfx.com/CN/10.3964/j.issn.1000-059303-0738-05.M. C. Araújo, T. C. Saldanha, R. K. Galvão, et al., Chem. Intell. Lab. Syst., 57, 65–73 (2001), doi: 10.1016/S0169-7439(01)00119-8, https://www.sciencedirect.com/science/article/abs/pii/S0169743901001198?via%3Dihub.M. C. Araújo, T. C. Saldanha, R. K. Galvão, et al., Chem. Intell. Lab. Syst., 57, 65–73 (2001), doi: 10.1016/S0169-7439(01)00119-8, https://www.sciencedirect.com/science/article/abs/pii/S0169743901001198?via%3Dihub.K. Liang, Q. Liu, J. Xu, et al., J. Appl. Spectrosc., 85, 953–961 (2018), doi: 10.1007/s10812-018-0745-y, https://link.springer.com/article/10.1007/s10812-018-0745-y.K. Liang, Q. Liu, J. Xu, et al., J. Appl. Spectrosc., 85, 953–961 (2018), doi: 10.1007/s10812-018-0745-y, https://link.springer.com/article/10.1007/s10812-018-0745-y.S. Wold, M. Sjostrom, L. Eriksson, Chem. Intell. Lab. Syst., 58, 109–130 (2001), doi: 10.1016/S01697439(01)00155-1, https://www.sciencedirect.com/science/article/abs/pii/S0169743901001551.S. Wold, M. Sjostrom, L. Eriksson, Chem. Intell. Lab. Syst., 58, 109–130 (2001), doi: 10.1016/S01697439(01)00155-1, https://www.sciencedirect.com/science/article/abs/pii/S0169743901001551.Z. Li, F. Deng, J. He, et al., Spectrosc. Spectr. Anal., 41, 2872–2878 (2021), doi: 10.3964/j.issn.1000059309-2872-07, http://www.gpxygpfx.com/CN/10.3964/j.issn.1000-059309-2872-07.Z. Li, F. Deng, J. He, et al., Spectrosc. Spectr. Anal., 41, 2872–2878 (2021), doi: 10.3964/j.issn.1000059309-2872-07, http://www.gpxygpfx.com/CN/10.3964/j.issn.1000-059309-2872-07.F. Melgani, L. Bruzzone, IEEE Trans. Geosci. Rem. Sens., 42, 1778–1790 (2004), doi: 10.1109/TGRS.2004.831865, https://ieeexplore.ieee.org/document/1323134.F. Melgani, L. Bruzzone, IEEE Trans. Geosci. Rem. Sens., 42, 1778–1790 (2004), doi: 10.1109/TGRS.2004.831865, https://ieeexplore.ieee.org/document/1323134.T. Qiao, J. Ren, C. Craigie, et al., J. Appl. Spectrosc., 82, 137–144 (2015), doi: 10.1007/s10812-0150076-1, https://link.springer.com/article/10.1007/s10812-015-0076-1.T. Qiao, J. Ren, C. Craigie, et al., J. Appl. Spectrosc., 82, 137–144 (2015), doi: 10.1007/s10812-0150076-1, https://link.springer.com/article/10.1007/s10812-015-0076-1.J. Ham, Y. Chen, M. Crawford, et al., IEEE Trans. Geosci. Rem. Sens., 43, 492–501 (2005), doi: 10.1109/TGRS.2004.842481, https://ieeexplore.ieee.org/document/1396322.J. Ham, Y. Chen, M. Crawford, et al., IEEE Trans. Geosci. Rem. Sens., 43, 492–501 (2005), doi: 10.1109/TGRS.2004.842481, https://ieeexplore.ieee.org/document/1396322.W. Zhou, H. Yang, L. Xie, et al., CATENA, 202, No. 12 (2021), doi: 10.1016/j.catena.2021.105222, https://www.sciencedirect.com/science/article/abs/pii/S0341816221000813?via%3Dihub.W. Zhou, H. Yang, L. Xie, et al., CATENA, 202, No. 12 (2021), doi: 10.1016/j.catena.2021.105222, https://www.sciencedirect.com/science/article/abs/pii/S0341816221000813?via%3Dihub.W. Xu, Y. Han, Y. Qin, et al., Spectrosc. Spectr. Anal., 35, 2856–2861 (2015), doi: 10.3964/j.issn.1000059310-2856-06, http://www.gpxygpfx.com/CN/10.3964/j.issn.1000-059310-2856-06.W. Xu, Y. Han, Y. Qin, et al., Spectrosc. Spectr. Anal., 35, 2856–2861 (2015), doi: 10.3964/j.issn.1000059310-2856-06, http://www.gpxygpfx.com/CN/10.3964/j.issn.1000-059310-2856-06.

The authors declare that there are no conflicts of interest present.