小波包分析-插值-RBF法同时测定铝合金中铁、锰、铜、锌

摘要
参考文献
相关文章 (15)

全文: PDF (475 KB)
输出: BibTeX | EndNote (RIS)

摘要：提出了一种在小波包分析对吸收光谱数据进行降噪处理的基础上，采用线性插值增加校正样样本数的新思路，为解决多组分分光光度同时测定中经常遇到的样本少变量多的问题，对提高预测结果的准确性提供了一种新方法。应用小波包分析-一维线性插值-RBF网络处理铝合金样品中铁、锰、铜、锌的同时测定。由于小波包分析-线性插值处理既能发挥良好的滤噪作用，又能使训练集样本对待辩识空间形成较好的覆盖，从而使RBF网络能提取到更多的特征信息，改善网络性能，研究结果表明，该方法可以显著降低测定样的相对误差，获得的测定结果令人满意。

关键词：小波包分析;线性插值;RBF网络;铝合金;同时测定

Abstract：A new approach was proposed so that noises in visible spectra are eliminated by wavelet packets, and the specimen numbers of calibration samples are increased by linear interpolation. It can used to solve problems of few specimen numbers and many variables in simultaneous spectrophotometric determination. Thus forecasting resultant veracity is improved. The contents of Fe, Mn, Cu and Zn in aluminous-alloy was determined simultaneously by using wavelet packets analysis-linear interpolation-RBF (radial basis function) neural networks (WPA-Interp-RBF). Wavelet packets analysis can eliminate noises to a desired level. Linear interpolation makes specimen of training samples cover recognizable room better. So RBF can abstract more characteristic information and improve its performance. Relative error(RE) of sample is reduced remarkably with desired results.

Key words：Wavelet packets analysis;Linear interpolation;Radial basis function neural networks;Aluminous-alloy;Simultaneous spectrophotometric determination

收稿日期: 2004-07-20 修订日期: 2004-12-05

中图分类号:

O657.3

通讯作者: 张运陶

引用本文:

程正军，张运陶^*. 小波包分析-插值-RBF法同时测定铝合金中铁、锰、铜、锌[J]. 光谱学与光谱分析, 2005, 25(10): 1658-1661.
CHENG Zheng-jun, ZHANG Yun-tao^*. Application of Wavelet Packets Analysis-Interpolation-RBF to the Simultaneous Determination of Fe, Mn, Cu and Zn in Aluminous-Alloy. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2005, 25(10): 1658-1661.

链接本文:

https://www.gpxygpfx.com/CN/Y2005/V25/I10/1658

[1] MA Ji-ping, WU Hai-ping, WANG Xing-yu(马继平, 吴海平, 王兴宇). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2000, 20(1): 122.
[2] HUANGFU Li-xia, CHEN Wen-jie, SI Sheng-zhu(皇甫立霞, 陈文杰, 司圣柱). Journal of Hefei University of Technology(合肥工业大学学报), 1999, 22(5): 122.
[3] ZHU Jin-lin, CHEN Yi-wei, CAO Yong-sheng, et al(朱金林, 陈奕卫, 曹永生, 等). Chinese Journal of Analysis Laboratory(分析试验室), 2003, 22(1): 39.
[4] LU Li-qiang, JIN Ji-hong(鲁立强, 金继红). Chinese Journal of Analytical Chemistry(分析化学), 1997, 25(7): 818.
[5] BAI Ling, NI Yong-nian(白玲, 倪永年). Chinese Journal of Analysis Laboratory(分析试验室), 2002, 21(1): 39.
[6] CHEN Li-li, JIN Ji-hong, QIN Sun-wei(陈莉莉, 金继红, 秦孙巍). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2003, 23(3):597.
[7] Jawerth B, Swedens W. SIAM Review, 1994, 36: 377.
[8] LI He-sheng, MAO Jian-qin, ZHANG Fu-tang, et al. Acta Automatica Sinica, 2003, 29(2): 242.
[9] CHENG Zheng-xing(程正兴). Algorithm and Application of Wavelet Analysis(小波分析算法与应用). Xi'an: Xi'an JiaoTong University Press(西安: 西安交通大学出版社), 1998. 156.
[10] ZHANG Yun-hua, XI Mei-cheng, CHEN Chang-song(张韵华, 奚梅成, 陈长松). The ways to Data Calculation and Algorithm(数值计算方法和算法). Beijing: Science Press(北京: 科学出版社), 2000. 1, 100.
[11] ZHU Ming-xing, ZHANG De-long(朱明星, 张德龙). Journal of Anhui University, Natural Science Edition(安徽大学学报), 2000, 24(1): 72.
[12] ZHOU Jun-wu, SUN Chuan-yao, WANG Fu-li(周俊武, 孙传尧, 王福利). Mining and Metallurgy(矿冶). 2001, 10(4): 71.
[13] Tin-Yan Kwok, Dit-Yan Yeung. IEEE Transaction on Neural Networks, 1997, 8(3): 630.
[14] CHEN Gui-ming, ZHANG Ming-zhao, QI Hong-yu(陈桂明, 张明照, 戚红雨). Digital Signal and Image Processing Based on MATLAB(应用MATLAB语言处理数字信号与数字图像). Beijing: Science Press(北京: 科学出版社), 2001. 273.
[15] LI Hai-tao, DENG Ying(李海涛, 邓樱). MATLAB How to Program(MATLAB程序设计教程). Beijing: Higher Education Press(北京: 高等教育出版社), 2002. 28, 236.
[16] FANG Kai-tai, MA Chang-xing(方开泰, 马长兴). Orthogonal and Uniform Design(正试验设计). Beijing: Science Press(北京: 科学出版社), 2000. 237, 247.