基于传感器网络的水环境监测及其时空场重建方法研究

The monitoring and early warning of water environment based on wireless sensor networks (WSNs) is an important development trend. The traditional WSNs collect and analyze various monitoring data and the unitary form is unable to meet the multi-level and all-round demands of water environmental monitoring. Visual perception and mobile sensor nodes can be introduced into the water environmental monitoring, which will effectively improve the diversity, accuracy and timeliness of the monitoring and early warning. On the basis of previous work, the project focuses on mobile sensor networks with visual perception, and researches the water environmental monitoring and its spatiotemporal field reconstruction which adequately give consideration to both the monitoring performance and network energy consumption. The specific contents are as follows. Firstly, visual perception algorithms for water environment and dynamic task offloading strategy are presented to implement the aquatic debris detection and salient regions recognition. Secondly, cooperative scheduling mechanism of mobile sensor nodes is proposed to improve the reliability and fine-granularity of monitoring using the mobility and group sensing of sensor nodes. Finally, the dynamic information about motion processes of aquatic debris and diffusion trends of pollution sources can be obtained in real time according to the spatiotemporal field reconstruction based on multi-sensor information fusion. Research results of the project will improve the automation and intelligence level of water environmental monitoring and provide better information technology support for the prevention and control of water pollution.

基于传感器网络开展水环境监测与预警是当前重要的发展方向。现有无线传感器网络主要通过节点实时采集水质参数，监测形式单一，已无法满足水环境监测多层次、全方位的需求。将视觉感知与移动传感器节点引入到水环境监测中来，能有效提高监测的多样性、准确性和时效性。在已有工作的基础上，本项目以具有视觉感知能力的移动传感器网络为对象，研究能有效兼顾监测性能与节点能耗的水环境监测及其时空场重构方法。具体内容包括: 研究水环境视觉感知算法及其动态卸载策略，实现水面漂浮物检测与显著性区域识别；研究移动传感器节点的协同调度机制，利用节点的机动性和群体感知方式提高水环境监测的细粒度与可靠性；研究基于多传感器节点信息融合的水环境时空场重构方法，实时获取水面漂浮物运动过程与污染源扩散趋势等动态信息。项目研究成果将提高水环境监测的自动化与智能化程度，为水污染防治提供更好的信息技术支撑。

将具有视觉感知能力的移动传感器网络应用于水环境监测与预警是当前重要的发展方向，有助于提升水环境监测的自动化与智能化水平。项目围绕移动传感器网络水环境监测与时空场重构方法开展了研究，提出了水面漂浮物检测与显著性区域识别的视觉感知算法、移动传感器节点运动控制策略与协同感知方法以及基于多传感器节点信息融合的污染源定位与时空场建模方法，并在传感器模块设计、传感网无线通信性能提升以及传感网数据压缩等相关技术方面进行了试验研究，取得了一系列的研究成果，实验测试也取得了良好效果。主要内容及创新点在于，将压缩感知理论与自适应图像分块技术引入背景减除算法实现水面漂浮物的检测；基于图像压缩重构与颜色系数最优线性组合的显著性检测方法来识别水面污染物区域；综合考虑无线网络链路质量和节点能耗设计传感器节点运动控制策略，通过融合节点运动过程中的位置与浓度信息采用支撑向量回归算法实现污染源时空场的重构。本项目的科学意义在于通过传感器节点视觉感知与协同调度能够获取准确、全面的水环境信息，实现了水面污染物的有效检测以及监测水域污染源时空场的动态重建，为水环境监测与应急响应提供新的思路。