Spatio-temporal knowledge is essential in understanding the dynamic aspects of complex scenes. However, existing knowledge graphs have limitations, such as inadequate time description, inflexible expression of semantic relationships, and difficulties in accessing GIS platforms. The article proposes the spatio-temporal object knowledge graph (STOKG), consisting of the object concept layer, spatio-temporal object layer, and dynamic version layer. To demonstrate the practical usefulness of the STOKG model, the Henan epidemic knowledge graph is created using epidemiological data from early 2020, which shows the dynamic evolution of the spatio-temporal objects of cases from the geography and semantic perspectives. Finally, the STOKG model is compared with the existing models in terms of accuracy, completeness and repetitiveness. The experimental results show that the STOKG model provides a more flexible and comprehensive approach to representing spatio-temporal knowledge, which is useful for applications in fields such as geography, epidemiology, and environmental science.

中文翻译：

---

用于动态场景表示的分层时空对象知识图模型

时空知识对于理解复杂场景的动态方面至关重要。然而，现有的知识图谱存在时间描述不充分、语义关系表达不灵活、接入GIS平台困难等局限性。文章提出时空对象知识图谱（STOKG），由对象概念层、时空对象层和动态版本层组成。为了证明STOKG模型的实用性，利用2020年初的流行病学数据创建了河南疫情知识图谱，从地理和语义角度展示了病例时空对象的动态演变。最后，将STOKG模型与现有模型在准确性、完整性和重复性方面进行了比较。实验结果表明，STOKG模型提供了一种更灵活、更全面的时空知识表示方法，对于地理学、流行病学和环境科学等领域的应用很有用。