An algorithm for constructing a topological skeleton for semi-structured spatial data based on persistent homology

Abstract

The construction and use of a topological skeleton for processing semi-structured information based on persistent homology methods is considered in the article. In the work, the main topological feature for analysis of object is a hole. The application of the developed algorithm to solve the actual problem of geoinformatics in the matching of spatial objects at different scales of map is shown. Comparison of topological skeletons at different tree depths is demonstrated.