面向地质灾害监测的无线视频传感器网络协同感知与灾害情景视觉分析研究

项目名称： 面向地质灾害监测的无线视频传感器网络协同感知与灾害情景视觉分析研究

项目编号： No.41202232

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 天文学、地球科学

项目作者： 王勇

作者单位： 中国地质大学（武汉）

项目金额： 24万元

中文摘要： 基于传感器网络技术开展地质灾害监测预警是当前的重要发展方向。传统无线传感网络主要通过灾害预测模型分析传感器节点实时采集的多种监测数据，来确定灾害发生的可能性，存在监测数据量大、各参量相互关系不明确以及缺少公认有效的预测模型等问题。通过将视觉图像信息引入到地质灾害监测中来，能有效提高预警的适时性与准确性。本项目在已有工作的基础上，研究无线视频传感器网络在地质灾害监测中的应用，重点研究能有效兼顾灾害预警准确性和网络能耗的传感器节点协同感知与灾情视觉分析方法。结合传感器节点信息收益与网络能耗构建效能函数，采用遗传算法选定参与协同感知的传感器节点。通过选定节点协同获取场景动态监测数据与图像信息，完成监控区域内人员、车辆等目标的跟踪定位，根据多视角场景图像的空间相关性实现三维场景重建，为指挥决策提供实时的动态灾情信息。项目研究成果将提高地质灾害监测的可视化程度，为防灾减灾提供更好的信息技术支撑。

中文关键词： 无线传感器网络；协同感知；视觉分析；监测预警；

英文摘要： The monitoring and early warning of geological disasters based on wireless sensor networks (WSN) is an important development trend. The traditional WSN forecasts the disasters through analyzing various monitoring data with a prediction model. The method has disadvantages, such as large amount of data, the ambiguity relationship between the parameters and the lack of effective disaster prediction model. The visual image information can be introduced into the geological disasters monitoring and early warning, which will effectively improve the timing and accuracy. On the basis of previous work, the project emphasizes on the application of the wireless video sensor networks in geological disasters monitoring. The key problem is to build the mechanism of the sensor nodes collaborative perception and the method of disaster visual analysis which adequately give consideration to the accuracy of early warning and the network energy consumption. We construct an efficiency function by combining the information utility with the energy consumption of sensor nodes, and apply genetic algorithms to determine the sensor nodes which participate in the cooperative sensing. The selected nodes collaboratively acquire the dynamic monitoring data and the scene image information, and accomplish the tracking and localization of personn

英文关键词： wireless sensor networks；collaborative perception；visual analysis；monitoring and early warning；