Abstract
Marine data is valuable for understanding and exploring the ocean. Given the massive, multi-source, heterogeneous, and time-location correlated characteristics of marine data, it is not easy to manage the marine data efficiently and query them quickly. In this paper, we construct an undirected weighted label graph for marine data, establishing the vertices and the edges based on marine events. Then we compress the marine graph based on the graph pattern matching with bisimulation relationship, and build views on the compressed graph to obtain efficient query results. Finally, we carried out experiments with different types of marine data on the 'typhoon' event, validated the effectiveness of the graph construction and compression on marine data and the query efficiency with the marine data graph compression and views.