基于改進小波閾值的量子點熒光光譜降噪研究
發(fā)布時間：2018-11-13 17:11
【摘要】：使用光譜儀采集到的信號難免受到不同噪聲源的影響。為了提高光譜信號解析的精準度,通過分析小波應用于信號降噪的原理以及經(jīng)典的軟、硬閾值降噪法存在的缺陷,提出了一種改進的閾值降噪法。該方法既克服了硬閾值法產(chǎn)生間斷點,軟閾值法產(chǎn)生恒定偏差的缺陷,又盡量地保留了有用信號。實驗選用的小波基函數(shù)為SymletsA,分解層數(shù)為4,結(jié)合Birge-Massart策略模型確定的分層閾值對硒化鎘量子點熒光光譜信號進行降噪處理。結(jié)果表明,與經(jīng)典的軟、硬閾值降噪法相比,通過改進閾值降噪法得到重構(gòu)信號的信噪比(SNR=47.550 2)、能量占比(PER=0.973 3)和均方誤差(MSE=149.421 3)均有提高和改善。
[Abstract]:The signals collected by spectrometer are inevitably affected by different noise sources. In order to improve the accuracy of spectral signal resolution, an improved threshold denoising method is proposed by analyzing the principle of wavelet application in signal denoising and the defects of classical soft and hard threshold de-noising method. This method not only overcomes the defect that hard threshold method produces discontinuity point, soft threshold method produces constant deviation, but also preserves useful signal as far as possible. The wavelet basis function of the experiment is SymletsA, decomposition layer number is 4, combined with the delamination threshold determined by Birge-Massart strategy model, the fluorescence spectrum signal of cadmium selenide quantum dot is de-noised. The results show that the SNR (SNR=47.550 _ 2), PER=0.973 _ 3 and mean square error (MSE=149.421 _ 3) of reconstructed signal are improved by the improved threshold denoising method compared with the classical soft-hard threshold denoising method.
【作者單位】： 貴州大學大數(shù)據(jù)與信息工程學院;
【基金】：國家國際科技合作專項項目(2014DFA00670)
【分類號】：TH744.1
本文編號：2329798
[Abstract]:The signals collected by spectrometer are inevitably affected by different noise sources. In order to improve the accuracy of spectral signal resolution, an improved threshold denoising method is proposed by analyzing the principle of wavelet application in signal denoising and the defects of classical soft and hard threshold de-noising method. This method not only overcomes the defect that hard threshold method produces discontinuity point, soft threshold method produces constant deviation, but also preserves useful signal as far as possible. The wavelet basis function of the experiment is SymletsA, decomposition layer number is 4, combined with the delamination threshold determined by Birge-Massart strategy model, the fluorescence spectrum signal of cadmium selenide quantum dot is de-noised. The results show that the SNR (SNR=47.550 _ 2), PER=0.973 _ 3 and mean square error (MSE=149.421 _ 3) of reconstructed signal are improved by the improved threshold denoising method compared with the classical soft-hard threshold denoising method.
【作者單位】： 貴州大學大數(shù)據(jù)與信息工程學院;
【基金】：國家國際科技合作專項項目(2014DFA00670)
【分類號】：TH744.1
【相似文獻】
相關期刊論文 前6條
1 程曉鋒,李澤仁,江孝國;一種高能X光轉(zhuǎn)換屏的熒光光譜實驗研究[J];光電工程;1996年S1期
2 凌明勝;錢志余;梁超英;;血糖濃度熒光光譜檢測研究[J];量子電子學報;2007年05期
3 黃珍;;ARL9800XP熒光光譜議故障分析及維修[J];農(nóng)村經(jīng)濟與科技;2009年10期
4 呂飛;石春;劉曉光;秦琳琳;吳剛;趙春江;;基于FPGA和DSP的時間分辨熒光光譜測硼儀[J];電子測量技術;2012年06期
5 陳小強;賴萬昌;張震;陳亨貴;于新華;;小波分析在X熒光光譜去噪聲中的應用[J];核電子學與探測技術;2009年04期
6 ;[J];;年期
相關碩士學位論文 前1條
1 李艷春;基于CCD的農(nóng)藥光纖熒光光譜測量系統(tǒng)的研究[D];燕山大學;2006年
，本文編號：2329798
本文鏈接：http://sikaile.net/kejilunwen/yiqiyibiao/2329798.html
