Ancient ceramic fragment image automatic splicing method, device, equipment and medium

Abstract

This application relates to a method, apparatus, device, and medium for automatically stitching together images of ancient ceramic fragments. The method includes: constructing a knowledge graph of ancient ceramic archaeological typology; extracting the geometric and apparent feature vectors of each fragment from received images of ancient ceramic fragments; querying corresponding entity nodes and attributes in the knowledge graph; generating a set of candidate matching pairs based on preset archaeological logic rules; calculating the matching degree score for each pair of candidate fragments in the candidate matching pair set; constructing a global stitching graph based on the matching degree score and a preset score threshold, and performing constraint optimization; and generating a three-dimensional virtual reconstruction model of the ancient ceramic based on the optimal connected subgraph obtained from the solution. This method ensures that the reconstruction results not only match precisely geometrically but also maintain a high degree of logical consistency in archaeological dimensions such as vessel shape, kiln date, and decorative style, significantly improving the accuracy, reliability, and overall processing efficiency of automatic stitching together of ancient ceramic fragments.

Images