舰船甲板油漆红外指纹图谱的无损鉴别  

Nondestructive identification of ship deck paint with infrared fingerprint spectrum

作　　者：何欣龙 王继芬 韩育林 梁震龙 张玉龙 HE Xinlong;WANG Jifen;HAN Yulin;LIANG Zhenlong;ZHANG Yulong(Institute of Forensic Science and Technology,People′s Public Security University of China,Beijing 102600,China;Institute of Safety and Order,People′s Public Security University of China,Beijing 102600,China)

机构地区：[1]中国人民公安大学刑事科学技术学院,北京102600 [2]中国人民公安大学治安学院,北京102600

出　　处：《兵器材料科学与工程》2019年第6期97-101,共5页Ordnance Material Science and Engineering

基　　金：2019年拔尖人才培养专项资助硕士研究生科研创新项目（2019ssky003）

摘　　要：为实现舰船甲板油漆的快速无损检测和精确识别与分类,利用光谱检测和数据分析,采集并获取酚醛漆型等5类共计50个不同品牌、不同生产厂家的舰船甲板油漆样本的红外指纹原始图谱、一阶导数图谱和二阶导数图谱,选择自动基线校正、峰面积归一化、多元散射校正和Savitzky-Golay 7点平滑图谱,建立基于Bayes判别、K近邻、径向基函数神经网络和多层感知器4种判别模型。结果表明:用二阶导数红外指纹图谱结合Bayes-KNN-RBF-MLP法能为舰船甲板油漆品牌的准确检测提供新的快速无损分析手段,且模型检测精度高,方法具有普适性和一定的借鉴意义。It is of great significance to realize the rapid nondestructive detection and accurate identification of different types of ship deck paints via spectral detection and data analysis.The infrared fingerprint original spectra,first derivative spectra and second derivative spectra of a total of 50 ship deck paints from 5 brands were collected and Automated Baseline Correction,Peak Area Normalization,Multiple Scatter Correction and Savitzky-Golay 7-point smoothing were selected to pretreat infrared spectra.Finally,four discriminant models,including Bayes discriminant,K-nearest neighbor,radial basis function neural network and multi-layer perceptron were established.The results show that second derivative infrared fingerprint combined with Bayes-KNN-RBF-MLP can provide a new analysis method for more accurate detection of ship deck paints and the accuracy of model detection is higher.The method has universality and certain reference significance.

关 键 词：舰船甲板油漆 二阶导数红外指纹图谱 模式识别 鉴别 

分 类 号：U668.3[交通运输工程—船舶及航道工程] O657.3[交通运输工程—船舶与海洋工程]