摘要
分布式光纤是一种新型在线监测技术,能极大程度提高电缆健康状态的监测效率;由于光纤处于复杂环境影响中,以及分布式节点实时产生庞大的数据群,这对监测数据的处理技术提出了更高的要求;基于传统的数字式平均法,引入一种改进的k-means聚类算法,实时对各节点产生的数据集处理,能准确的识别因噪声影响而产生的奇异数据,提高了数据反馈的效率和准确性,从而减少了监测系统的漏报和误报现象;现有的实验仿真表明改进的算法较传统算法在数据处理的准确性和快速性上都有明显的提升。
Distributed optical fiber is a new type of online monitoring technology, which can greatly improve the efficiency of monitoring the health status of the cable. The fiber is affected in complex environment, and distributed node will produce huge data base in real time, which proposed higher requirements of data processing technology. In this paper, based on digital average method, we introduce an improved k--means clustering algorithm, which can process data flow generated by each node, to remove all of singular values generated by noise accu- rately and improve the efficiency and accuracy of data feedback. Thanks to the developed algorithm, the phenomenon of misinformation gets ameliorated. The existing simulation results show that the improved algorithm, compared to the traditional algorithm, has significantly im- proved the accuracy and speed of data processing.
出处
《计算机测量与控制》
2015年第12期3969-3971,共3页
Computer Measurement &Control
基金
国家自然科学基金项目(51207088)
上海市自然科学基金(12ZR1412100)
上海市地方能力建设项目(14110500900)
