瓦斯传感网络时空数据的预处理  

Pre-processing for spatiotemporal data of methane sensor networks

作　　者：胡圣波[1] 舒恒[1] 

机构地区：[1]贵州师范大学智能信息处理研究所,贵州贵阳550001

出　　处：《贵州师范大学学报（自然科学版）》2012年第3期75-78,共4页Journal of Guizhou Normal University：Natural Sciences

基　　金：国家自然科学基金项目(61064614)

摘　　要：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。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 detec- tion 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.

关 键 词：瓦斯 传感网络 时空 预处理 

分 类 号：TD76[矿业工程—矿井通风与安全]