Study of fumigant residues on the surface of paper artefacts based on hyperspectroscopy

Paper cultural relics are important carriers of splendid history and culture, and have important historical research value. As paper is mainly rich in cellulose, starch and protein, paper cultural relics are prone to mould, insects and other microorganisms in the process of long-term preservation, leading to corrosion, deterioration and even destruction of cultural relics. Fumigation method is currently more widely used in a rapid means of control of cultural relics of mould and mildew, fumigant residue detection is the establishment of a set of scientific fumigation method in an indispensable part. In this paper, for the surface of paper cultural relics there are fumigant residues and no residues of spectral characteristics of the variability, based on the characteristics of spectral nondestructive testing, using BP neural network algorithm, SVM algorithm, KNN algorithm, 1D-CNN algorithm were established to establish discriminatory models, according to the different models of the discrimination accuracy of the model performance assessment, select the optimal modelling method.