瓦斯传感网络时空数据的预处理

由于煤矿巷道环境恶劣,提高瓦斯传感网络数据可靠性十分重要。从利用传感网络数据的时空关联性出发,研究了基于时空融合的瓦斯传感网络数据的预处理方法。首先,给出了传感网络时间、空间颗粒的定义,提出了瓦斯传感网络时间、空间颗粒的划分方法。接着,论述了异常值检测,以及基于时空融合的瓦斯传感网络数据的平滑预处理方法。仿真结果表明,提出的瓦斯传感网络时空数据预处理方法是有效、合理的。

Pre-processing for spatiotemporal data of methane sensor networks

It is urgent to improve the data reliability of methane sensor networks,because the coal laneway is narrow,and its environment with excessive dirt,dust,moisture is harsh.In view of the data spatiotemporal correlations of methane sensor networks,the pre-processing methods based on spatiotemporal fusion for data of methane sensor networks are presented.First,Temporal and spatial granules are defined and the methods for determining the size of the granules are given.Next,the outlier detection and smooth in pre-processing are described.Finally,simulations show that the pre-processing methods presented for data of methane sensor networks are effective and relational.