本论文结合模式识别技术，在基于本体理论的基础上研究并实现强对流天气推理系统。该系统通过模式识别技术对图像数据进行预处理和特征提取，从中获取有用的特征信息，并结合气象专业知识，对特征信息进行分析，抽象出符合一定气象意义的知识，建立领域本体库，在此基础上研究并设计强对流领域推理规则，进而建立起强对流天气推理系统，以提高强对流气象预报领域的计算机化、快速化和实时化。-This paper combines the pattern recognition technology, ontology-based research based on the theory and reasoning to achieve strong convective weather systems. The system uses pattern recognition technology for image data preprocessing and feature extraction, feature information from which to obtain useful, combined with meteorological expertise to analyze information on the characteristics of the abstract meaning that meet certain meteorological knowledge, build domain ontology library, Based on research and design of inference rules convective areas, and setting up a strong convective weather inference system to increase the area of severe convective weather forecast computerized, fast and real-time.