基于最近邻比较和中值滤波的拉曼光谱spike自动检测及剔除方法研究

大气与环境光学学报 ›› 2017, Vol. 12 ›› Issue (2): 128-135.

基于最近邻比较和中值滤波的拉曼光谱spike自动检测及剔除方法研究

闫静文, 李颖*, 陈靓, 杨德旺, 郑荣儿

(中国海洋大学光学光电子实验室，山东青岛 266100)

收稿日期:2015-08-12 修回日期:2015-09-01 出版日期:2017-03-28 发布日期:2017-03-23
通讯作者: 李颖，（1971-），女，山东滨州人，副教授，主要从事激光光谱技术及其应用方面的研究。 E-mail:liying@ouc.edu.cn
基金资助:
Supported by National High Technology Research and Development Program of China(国家863计划，2012AA09A405）

Spike Detection and Removal of Raman Spectra by the Method Combined Nearest Neighbor Comparison with Median Filter

YAN Jingwen, LI Ying*, CHEN Jing, YANG Dewang, ZHENG Ronger

(Optics and Optoelectronics Laboratory, Ocean University of China, Qingdao 266100, China)

Received:2015-08-12 Revised:2015-09-01 Published:2017-03-28 Online:2017-03-23

摘要/Abstract

摘要：

激光拉曼光谱技术具有原位、实时、多物质同时探测的优点，在实验室内已经得到了广泛研究，并逐渐应用于深海海底等极端环境下的原位探测。由于CCD等光谱探测器可能受到宇宙射线的干扰，会在拉曼谱图上形成位置随机的spike，对拉曼光谱分析造成干扰。将最近邻比较和中值滤波相结合的方法应用于海水实测拉曼光谱spike的检测剔除，首先将原始拉曼光谱分为拉曼峰区域和非拉曼峰区域，对拉曼峰区域使用最近邻比较法处理，对非拉曼峰区域使用中值滤波方法，结果显示该方法能将非拉曼峰区域和拉曼峰区域的spike完全剔除，没有出现误检和漏检，且不会对拉曼特征峰强度产生影响，有助于提高拉曼光谱定性定量分析的精确性和稳定性。

关键词: 拉曼光谱, spike, 最近邻比较法, 中值滤波

Abstract:

Laser Raman spectroscopy is widely applied in marine chemical environmental detection and analysis of substances, as an in-situ, real-time, multi- substances detection sensor technology. Charge-coupled devices (CCD) as the detector of Raman spectrometer are easily interfered by cosmic ray so that some spikes that have random location ,intensity and peak width occur on Raman spectra. The presence of such spikes can damage spectral presentation, complicate interpretation, and impede further analyses. The method combined nearest neighbor comparison method（NNC）with median filter was used to detect and remove spike from the measured spectral data of seawater in laboratory. The Raman spectra were divided into two areas firstly: Raman peaks area and the area that has no Raman peaks. The NNC processed the area that has Raman peaks and the median filter processed the area that has no Raman peaks. The result shows that the combined method can remove spike of the full spectra completely on the premise that the intensity of Raman peaks remain unchanged. The method contributes to improving the accuracy and stability of qualitative and quantitative analysis for Raman spectra.

Key words: Raman spectra, spike, nearest neighbor comparison method, median filter

中图分类号:

O433

闫静文李颖陈靓杨德旺郑荣儿. 基于最近邻比较和中值滤波的拉曼光谱spike自动检测及剔除方法研究[J]. 大气与环境光学学报, 2017, 12(2): 128-135.

YAN Jing-Wen, LI Ying, CHEN Jing, YANG De-Wang, ZHENG Rong-Er. Spike Detection and Removal of Raman Spectra by the Method Combined Nearest Neighbor Comparison with Median Filter[J]. Journal of Atmospheric and Environmental Optics, 2017, 12(2): 128-135.

参考文献

[1] Brewer P G, Malby G, Pasteris J D, et al. Development of a laser Raman spectrometer for deep-ocean science [J]. Deep Sea Res. PT., 2004, 51(5): 739-753.
[2] White S N, Dunk R M, Peltzer E T, et al. In situ Raman analyses of deep‐sea hydrothermal and cold seep systems (Gorda Ridge and Hydrate Ridge) [J]. Geochem., Geophy., Geosy., 2006, 7(5): 5023.
[3] Du Z, Li Y, Chen J, et al. Feasibility investigation on deep ocean compact autonomous Raman spectrometer developed for in-situ detection of acid radical ions [J]. Chin. J. Oceanol. Limnol., 2015, 33(2): 545-550.
[4] Zhang D M, Ben-Amotz D. Removal of cosmic spikes from hyper-spectral images using a hybrid upper-bound spectrum method [J]. Appl. Spectr., 2002, 56(1): 91-98.
[5] Schulze H G, Turner R F B. A two-dimensionally coincident second difference cosmic ray spike removal method for the fully automated processing of Raman spectra [J]. Appl. Spectr., 2014, 68(2): 185-191.
[6] Bussian B M, H?rdle W. Robust smoothing applied to white noise and single outlier contaminated Raman spectra [J]. Appl. Spectr., 1984, 38(3): 309-313.
[7] Katsumoto Y, Ozaki Y. Practical algorithm for reducing convex spike noises on a spectrum [J]. Appl. Spectr., 2003, 57(3): 317-322.
[8] Hill W, Rogalla D. Spike-correction of weak signals from charge-coupled devices and its application to Raman spectroscopy [J]. Anal. Chem., 1992, 64(21): 2575-2579.
[9] Phillips G R, Harris J M. Polynomial filters for data sets with outlying or missing observations: application to charge-coupled-device-detected Raman spectra contaminated by cosmic rays [J]. Anal. Chem., 1990, 62(21): 2351-2357.
[10] Schulze H G, Turner R F B. A fast, automated, polynomial-based cosmic ray spike-removal method for the high-throughput processing of Raman spectra [J]. Appl. Spectr., 2013, 67(4): 457-462.
[11] Takeuchi H, Hashimoto S, Harada I. Simple and efficient method to eliminate spike noise from spectra recorded on charge-coupled device detectors [J]. Appl. Spectr., 1993, 47(1): 129-131.
[12] Zhang D M, Jallad K N, Ben-Amotz D. Stripping of cosmic spike spectral artifacts using a new upper-bound spectrum algorithm [J]. Appl. Spectr., 2001, 55(11): 1523-1531.
[13] Behrend C J, Tarnowski C P, Morris M D. Identification of outliers in hyperspectral Raman image data by nearest neighbor comparison [J]. Appl. Spectr., 2002, 56(11): 1458-1461.
[14] Zhang L, Henson M J. A practical algorithm to remove cosmic spikes in Raman imaging data for pharmaceutical applications [J]. Appl. Spectr., 2007, 61(9): 1015-1020.
[15] Cappel U B, Bell I M, Pickard L K. Removing cosmic ray features from Raman map data by a refined nearest neighbor comparison method as a precursor for chemometric analysis [J]. Appl. Spectr., 2010, 64(2): 195-200.
[16] Li Sheng，Dai Liankui. A simple spike removal algorithm for online Raman monitoring [J]. Chinese Journal of Light Scattering, 2011, 23(3): 188-194(in Chinese).
李晟, 戴连奎. 一种简单的在线拉曼光谱 spike 剔除方法 [J]. 光散射学报, 2011, 23(3): 188-194.
[17] Li S, Dai L K. An improved algorithm to remove cosmic spikes in Raman spectra for online monitoring [J]. Appl. Spectr., 2011, 65(11): 1300-1306.
[18] Mozharov S, Nordon A, Littlejohn D, et al. Automated cosmic spike filter optimized for process Raman spectroscopy [J]. Appl. Spectr., 2012, 66(11): 1326-1333.