摘要
将基于统计特征的三维荧光谱参量化方法与神经网络模式识别技术相结合,实现污染油的油种鉴别。选取石油三维荧光谱的特征参数并组成特征向量,然后用神经网络技术实现油种的特征参数模式识别。通过基于实验数据的应用仿真,验证了这一方法的实用性和可靠性,从而可以取代人工观察对比的粗大方法。作为基于物质荧光的“指纹”新技术,也可应用于其它物质分析鉴别技术领域。
With parameterization of oil's 3D fluorescence spectrum based on its statistic features and ANN combined, oil identification is studied. With feature parameters of oil's 3D fluorescence spectrum as input vector of ANN, oil identification using pattern recognition techniques is presented,as so called “feature parameterization identification”. This method has turn out to be reliable and usable in practice in stead of artificial watching. This new fingerprint technique can also be used in other aspects such as public security.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2005年第z1期727-728,734,共3页
Chinese Journal of Scientific Instrument
基金
河北省自然科学基金(D2004000195)资助项目
关键词
三维荧光谱
参量化
统计特征
油种鉴别
Three-dimensional fluorescence spectrum Parameterization Statistic feature Oil identification
