Integration of heterogeneous agriculture information system based on interoperation of domain ontology

Existing agriculture information systems can meet general storage and general management requirements, however, on intensive management mode, users need to visit and gather resources and data in multiple heterogeneous information systems that distribute in different location of the web. These heterogeneous information systems not only have difference in platform construction, geography location, business logic storage mode, but also in area professional terminology. In this article, 3 levels of agriculture information systems integrated processing of systems heterogeneity, model heterogeneity and semantic heterogeneity were analyzed separately. Systems heterogeneity refers to different data sources of application system, database system, operation system and hardware platform; model heterogeneity means data sources differentiate in storage mode, semantic heterogeneity indicates information sources have differentiation in Semantics. Agriculture Information Bus (AI-Bus) model was proposed which including 4 levels that are service layer, protocol layer, data layer and routing layer, and achieved rule oriented flexible mechanism, provided configurable process definition, heterogeneous differences of information system were diminished from systematic level and mode level. Since domain ontology has good concept layer structures and rich in semantic relationships, it is important in information resources gathering and knowledge expressing. As to semantic heterogeneity problem appearing in information system integration, agriculture ontology was introduced to represent domain knowledge sharing and reusing, and colligated concept similarity computing method and description similarity computing method. For concept similarity, it includes 3 quantitative calculation indexes that are semantic coincidence ratio, semantic distance and hierarchy depth; for description similarity, it includes 2 quantitative calculation indexes that are relationship similarity and property similarity. Based on concept similarity and description similarity computation results, an ontology mapping approach was presented which solves the interoperation of multi-source heterogeneous information at the semantic level. Through hierarchical model and domain ontology interoperability applied to agricultural information system integration, currently widely existing “information isolated island” can be eliminated, and thus improve business application systems data sharing and service efficiency.