Identification of ancient ceramics by digital shape characterization

Shapes are important clues for the identification of ancient ceramics. In this study, digital extraction methods of shape characterization were discussed, using samples of flared mouthed porcelain bowls produced in Jingdezhen Hutian Kiln from the Five Dynasties to the Song and Yuan period. By employing the techniques of image enhancement, edge detection and curve fitting of Matlab, the method using digital shape characterization has advanced accuracy and effectiveness of traditional identification which is based on visual and tactual inspections. Together with the increasingly applied compositional data analysis and thermoluminescence dating technique, the method will further improve and upgrade the scientific identification system for ancient ceramics.