基于事件语义的滑坡监测传感器数据集成研究

项目名称： 基于事件语义的滑坡监测传感器数据集成研究

项目编号： No.41201380

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 地理学

项目作者： 陈雯

作者单位： 同济大学

项目金额： 25万元

中文摘要： 地学传感器网络是当今滑坡监测的重要手段之一，研究分布式异构综合监测环境下的数据集成方法对滑坡预警及应急决策具有重大意义。现有传感器数据集成或采用制定数据标准来进行句法统一、或从生数据（Raw Data）的层次进行语义集成，尚未达到真正"无缝"、且缺乏与地学现象的关联机制，导致各传感器的协同工作能力较弱，且观测网对环境的感知能力不强。针对以上问题，本课题提出从事件语义的角度进行传感器数据集成的思路。首先，以传感器本体和时空本体为基础，建立滑坡监测数据集成事件模型；其次，以复杂事件处理为关键技术，结合定量滑坡预报模型和定性判定规则，探讨监测数据与滑坡地学复杂事件的自动关联方法；最后，综合上述模型和关联方法，设计并开发面向滑坡监测的地学传感器语义中间件。本课题的研究成果可以有效的提高多个异构地学传感器乃至多个异构传感器网络的协同工作能力，同时大幅度增强传感器网络的感知能力。

中文关键词： 滑坡；传感器网；语义集成；地学事件；复杂事件处理

英文摘要： Geo-sensor network is one of the important means for landslide monitoring. Developing a data collection method for comprehensive monitoring of a distributed and heteorgeneously-structured environment is of great significance to landslide early warning and emergency decision making. Currently available methods for senseor data integration either rely on syntactic unification through data standardization or semantic synthesis at raw data level. These methods are far from reaching a "seamless" integration and lacking necessary association with the targested geophysical objects, which may degrade the inter-operability of the sensors and lead to misinterpretations of sensor observations. Aiming to seek solutions to the aforementioned problems, this study proposes a method for sensor data integration from an "event semantics layer" concept. In this framework, an event model for integrating landslide monitoring data is first constructed on the basis of sensor ontology and spatiotemporal ontology, then an auto-linking method (the key technique for complex event processing) for associating simple observations with complex landslide events is developed through both quantitative prediction modeling and qualitative rules, Finally, the geo-sensor semantic middle-ware for landslide monitoring is designed and developed by inte

英文关键词： Landslide；Sensor Web；Semantic Integration；Geographical Event；Complex Event Processing