摘要
光纤电容液滴分析仪利用液滴分析技术,综合在液体形成液滴的监测过程中获得的物理、化学特性参数,进行液体识别.基于液滴分析仪的功能原理,结合其现有的识别算法,提出了一个基于BP神经网络技术的液体识别方案,设计了3层BP神经网络分类器,完成了嵌入式系统功能算法的设计.对典型样品进行了测试实验,结果表明,对大部分液体可以进行完全识别,部分液体的识别率达96%以上,识别精度可达95%.
The fiber capacitance drop analyzer,using drop analysis technology,can analyze the liquid synthetically by the affections of their physical and chemical parameters obtained by monitoring the drop formation process.Based on the function and principle of fiber capacitance drop analyzer and with its present recognition algorithm considered,this paper proposes an overall implementing project by means of BP neural network technology.Accord-ing to the principle of BP neural network,a three-layer BP neural network recognition method is designed,and em-bedded system and its analysis program are also developed.The experiments of typical samples are carried out and the results show that the majority of the samples is recognized correctly,the recognition ratio of some samples is above 96% and precision achieves 95%.
出处
《天津大学学报》
EI
CAS
CSCD
北大核心
2011年第5期445-449,共5页
Journal of Tianjin University(Science and Technology)
关键词
液滴分析
液滴指纹图
BP神经网络
嵌入式系统
识别算法
liquid drop analysis
liquid drop fingerprint
BP neural network
embedded system
recognition algorithm
