基于红外反射法的金属薄板印涂湿膜厚度测量及曲面拟合研究  

Measurement of Wet Film Thickness Based on Infrared Reflection Method and Research on Surface Fitting

作　　者：陆观[1] 曹冠[1] 徐一鸣[1] 袁江[1] 邵建新[1] 邱自学[1] 

机构地区：[1]南通大学机械工程学院,江苏南通226019

出　　处：《仪表技术与传感器》2017年第6期183-188,共6页Instrument Technique and Sensor

基　　金：国家自然科学基金资助项目(61273024;51307089);交通运输部应用基础研究项目(2014319813180);江苏省产学研前瞻性联合研究项目(BY2015047-11);江苏省教育厅面上项目(14KJB510030);南通大学自然科学研究专项项目(13ZJ003)

摘　　要：为了实现金属薄板印涂厚度的评估,设计了一种基于红外反射法的金属薄板印涂湿膜厚度在线检测系统,此系统主要包括机械结构和测控系统2部分。通过测量生产线上关键点厚度值拟合单板单线和多板全平面的印涂湿膜厚度,运用Matlab和上下四分位截断法对测得的厚度数据进行剔除粗大误差处理,最后利用BP神经网络法对处理后的数据进行曲面拟合和仿真。拟合的结果表明:设计的金属薄板印涂湿膜厚度测量系统可以对生产线上的一批金属薄板进行整体的厚度评估;BP神经网络曲面拟合方法精度高,曲面较平滑,为金属薄板印涂厚度的评估提供了指导作用。In order to realize the evaluation of sheet metal printing and coating thickness, an online detection system based on infrared reflection method of sheet metal priming and coating wet film thickness was designed. The system mainly includes two parts, mechanical structure and control system. By measuring the key point on the production line to fitting single in single-line and multi-board plane printed wet film thickness, Matlab and upper and lower four bit truncation method was used to remove the gross error processing according to measured thickness data. Finally, the BP neural network method was used to fit and simulate the processed data. The fitting results show that the design of the metal sheet printing and coating wet film thickness measurement system can gauge the overall thickness of a number of metal sheet on the production line and the BP neural network curve fitting method has high precision and smooth surface, which provides a strong guidance for the assessment of the thickness of the metal sheet.

关 键 词：金属薄板 测厚 反向传播算法 曲面拟合 

分 类 号：TH741[机械工程—光学工程]