Research on glass classification and identification based on heritage samples

Abstract

The study of ancient glass has been intensified in archaeological circles today, but its susceptibility to weathering can affect archaeologists' ability to classify it. In order to classify glass artefacts, this paper establishes a 'subclass classification model based on the sum of the absolute values of relative changes' and uses the K-means clustering algorithm to classify the subclasses. In order to identify different classes of glass objects, the chemical composition of the glass is extracted from the centroids of the different subclasses, and the Euclidean distance between the class centres is solved for, and the class represented by the class centre with the lowest Euclidean distance is taken as the class corresponding to the unknown object.