数据融合技术在环境监测网络中的应用与思考

按照环境监测网络的信息处理结构,将网络划分为采集层、网络层和应用层,归纳了数据级融合、特征级融合、决策级融合3层融合级别,介绍了加权平均、卡尔曼滤波、人工神经网络、支持向量机、遗传算法、贝叶斯系统等在环境监测数据分析中的应用,指出当前面临的数据处理技术不成熟、评价体系不完善和数据质量监管能力不足等问题,提出优化网络结构、数据多元应用和强化决策支持等研究建议。

Applications and Reflections of Data Fusion in Environmental Monitoring Network

The structure of environmental monitoring data processing is divided into three layers:data collection, network transmission, application. And then, data fusion, feature fusion, decision fusion are introduced as three levels of environmental data fusion. In terms of monitoring data analysis, several fusion algorithms are illustrated and compared, such as weighted mean, Calman filter, artificial neural network, support vector machine, genetic algorithm and Bayesian system. Finally, existing problems including incomplete data processing technology, immature evaluation system and lacking of data quality supervision are pointed out. Recommendations such as monitoring network optimization, multivariate data application and enhancement of decision support are proposed accordingly.