农业大数据时代到来，使我们面临不同领域、不同来源的数据集的多样性，如何挖掘与发现存在潜在联系的数据集中的知识是目前面临解决的重要科学问题。本项目在已构建农业领域本体和知识库基础上，研究从多数据源中抽取、农业知识转换和融合问题，有效地提升农业知识服务能力。首先结合农业本体，研究农业知识体抽取、标注、聚类等算法研究，并利用语义关系验证知识关联，在此基础上研究基于农业本体的语义规则构建方法，研究基于属性、实例和概念三个粒度层面的不同知识粒度的知识一致性处理方法。最后研究用户需求的农业知识融合求解结果验证方法，判断知识融合方法是否有效，能否达到实际应用，并实现在农业大数据环境下，针对农业领域开发农业知识融合系统原型。本项目的研究方法、技术方案、原型系统和实验数据不仅对农业知识服务中的知识获取具有重要意义，而且为农业领域的智能决策、远程教学和知识管理等相关研究提供理论指导和工程方法。

With the agricultural big data era coming, we are facing the diversity of data sets from different areas and different sources, and the knowledge mining and discovery in the data set with potential relationships has become the important research problem. On the basis of small agricultural ontology and knowledge basis, the research focuses on the problem of agricultural knowledge extraction, knowledge conversion and knowledge fusion under the environment of agricultural big data, and aims to improve the agricultural knowledge service quality. The project will do research into key technologies of multi-granularity knowledge fusion based on agricultural ontology and fusion rules from the three perspectives: (1) Research on agricultural on agricultural ontology-based knowledge extraction, clearing and annotation methods, and semantic relations are adopted for knowledge connection validating; (2) Research on fusion & semantic rules construction based ontology, rule choice and result evaluation, and knowledge fusion will process the knowledge consistency method based on three different knowledge granularity levels of property, instance, and concept. (3) Research on knowledge fusion algorithm for user demands, including fusion result ' transformation, optimizing and evaluation. At last, we will develop a prototype system to prove and analyze the method’ feasibility. The research method, technical architecture, prototype system and experimental data have far-reaching significance in knowledge acquisition for the users, and provide theory guide and engineering techniques for intelligent decision, data mining, distance learning, and knowledge management in agricultural domain.