Abstract: 【Purpose/significance】Faced with a large amount of cultural relic information resources on the Internet, how to extract
knowledge from them to build a knowledge graph and conduct application research is the basis for deep utilization of cultural relic
knowledge in the new media era.【Method/process】The article first classifies information resources according to their content and
structural features, then uses keyword extraction methods to obtain keywords that summarize the topics of information resources, per? forms named entity identification of cultural relic information resources in knowledge bases through SPARQL, and finally uses lexical similarity algorithms to rely on ontology for knowledge fusion to build knowledge graph.【Result/conclusion】Knowledge extraction and knowledge graph construction of online cultural relic information resources were carried out in the experiments. The results showed that the article method can make full use of the cultural relic information resources on the Internet for knowledge graph construction and meet the analysis requirements in a variety of application scenarios.【Innovation/limitation】Due to the huge content of cultural relic domain, the relevant research data need to be further expanded to better investigate the applicability of the article method.