盧 濤,白薇薇
(河南工業(yè)大學(xué)電氣工程學(xué)院, 河南 鄭州 450001)
基于多普勒頻移和頻帶展寬的自相關(guān)光聲流速矢量測(cè)量
盧濤*,白薇薇
(河南工業(yè)大學(xué)電氣工程學(xué)院, 河南 鄭州 450001)
摘要:為了測(cè)量碳顆粒懸混溶液的流速矢量,將自相關(guān)方法引入了光聲多普勒流速測(cè)量。縱向和橫向速度分量分別由多普勒頻移和多普勒頻帶展寬得到。光聲信號(hào)由波長532 nm,重復(fù)頻率20 Hz的脈沖激光激勵(lì),由中心頻率10 MHz的點(diǎn)聚焦壓電超聲換能器采集。由微量注射泵驅(qū)動(dòng)的碳顆粒懸混液模擬血液的流動(dòng)。時(shí)域光聲信號(hào)由希爾伯特變換為復(fù)信號(hào)后進(jìn)行自相關(guān)計(jì)算。多普勒頻移和多普勒頻寬的標(biāo)準(zhǔn)偏差由若干獨(dú)立A掃的自相關(guān)平均得到。對(duì)比之前采用序列A掃的互相關(guān)方法,自相關(guān)中的信號(hào)時(shí)移大小可自定義,避免了對(duì)高重復(fù)頻率脈沖激光的要求,有利于探測(cè)深度的提高。該方法的可行性通過對(duì)流速為16-32 mm/s、多普勒角度為50°的碳顆粒懸混液流速矢量的測(cè)量得到了初步驗(yàn)證。
關(guān)鍵詞:光聲多普勒;流速矢量;自相關(guān);多普勒頻移;多普勒頻譜展寬
0引言
光聲成像(photoacoustic tomography,PAT)結(jié)合了純光學(xué)成像和超聲成像的優(yōu)點(diǎn),可對(duì)生物組織內(nèi)部進(jìn)行高空間分辨率和高圖像對(duì)比度的成像。基于光聲多普勒效應(yīng)(photoacoustic Doppler,PAD)的流速測(cè)量也是近年研究的熱點(diǎn)。在相關(guān)的研究工作中,基于光聲多普勒效應(yīng)的速度縱向分量、橫向分量和速度矢量的測(cè)量大多采用了短時(shí)傅里葉變換或互相關(guān)算法[1-11]。在互相關(guān)中,通過將速度的縱向分量和橫向分量相結(jié)合,進(jìn)行了速度矢量的測(cè)量[2-9]。為了避免序列A掃光聲信號(hào)對(duì)之間的信號(hào)混疊或不相關(guān),互相關(guān)方法需要采用重復(fù)頻率為數(shù)千赫茲的脈沖激光[2-5]。但是,一般在光聲成像中采用的高能脈沖激光很難達(dá)到數(shù)千赫茲的脈沖重復(fù)頻率,且速度測(cè)量范圍也取決于對(duì)光聲信號(hào)序列A掃的時(shí)間間隔,此外,高重復(fù)頻率的脈沖激光由于單脈沖能量較小,也不利于探測(cè)深度的提高。眾所周知,在純超聲多普勒和光學(xué)多普勒成像(optical Doppler tomography,ODT)中,自相關(guān)方法已經(jīng)得到較為成熟的應(yīng)用[12, 13]。自相關(guān)基于對(duì)單幀信號(hào)的A掃,信號(hào)的時(shí)移為自定義參數(shù),最小時(shí)移取決于信號(hào)的采樣間隔[13]。據(jù)此,本文將自相關(guān)方法應(yīng)用于光聲多普勒流速矢量的測(cè)量,軸向和橫向速度分量分別由多普勒頻移和多普勒頻帶展寬得到,進(jìn)而由軸向和橫向速度分量得到速度矢量和多普勒角度。實(shí)驗(yàn)中采用了波長為532 nm的脈沖激光為光聲信號(hào)激勵(lì)源,光聲信號(hào)由中心頻率為10 MHz的點(diǎn)聚焦壓電超聲換能器采集,碳顆粒懸混液作為模擬血流樣品由微量注射泵驅(qū)動(dòng)。該方法的可行性通過對(duì)流速為18-32 mm/s的碳顆粒懸混液流速矢量的測(cè)量得到了初步驗(yàn)證。
1自相關(guān)速度矢量測(cè)量方法
1.1縱向速度分量測(cè)量
縱向速度分量平行于超聲探頭的信號(hào)接收軸線。在相關(guān)研究中,縱向速度分量采用了基于序列A掃的互相關(guān)算法[2-9]。借鑒在純超聲流速測(cè)量和ODT中采用的自相關(guān)流速測(cè)量方法[12,13],縱向速度分量表示為[3]:
(1)
其中:△φ是信號(hào)的相移, c是超聲在水中的傳播速度,f0是超聲探頭的中心頻率, k是由實(shí)驗(yàn)確定的校準(zhǔn)系數(shù),△T為時(shí)移(time lag),△φ可由自相關(guān)表示為[12,14]:

(2)
其中:Ry和Rx分別為復(fù)自相關(guān)函數(shù)R的虛部和實(shí)部,R表示為[14]:
(3)

1.2橫向速度分量測(cè)量
橫向速度分量垂直于超聲探頭的信號(hào)接收軸線。當(dāng)流速較低時(shí),懸混液中顆粒的布朗運(yùn)動(dòng)是頻譜展寬的主要原因,當(dāng)流速較高時(shí),懸混液中運(yùn)動(dòng)顆粒通過超聲探頭聚焦區(qū)域的渡越時(shí)間是頻譜展寬的主要因素[15],橫向速度分量表示為[3]:
(4)
其中:Bd為頻譜的3dB帶寬, W是超聲探頭的陣元直徑,F是超聲探頭的焦距, b是由布朗運(yùn)動(dòng)、速度梯度和其他原因造成的頻譜背景展寬。c、f0和k的定義與公式(1)中相同。Bd的大小可由頻譜的標(biāo)準(zhǔn)偏差σ確定如下[13,14]:
(5)

2實(shí)驗(yàn)系統(tǒng)
如圖1所示, 采用波長為532nm、重復(fù)頻率20Hz的Nd:YAG脈沖激光(Quanta-RayINDI,SpectrumPhysics)為光聲信號(hào)激勵(lì)源,點(diǎn)聚焦壓電超聲換能器(OlympusIR-1008-S-SU)中心頻率10MHz、焦距19mm、焦點(diǎn)直徑0.23mm。低噪聲前置放大器 (Olympus5676) 帶寬50kHz-20MHz。單幀光聲信號(hào)由示波器(Agilent90404A)經(jīng)10次采集平均得到。 模擬血流樣品為容積比為5%的碳顆粒懸混液,懸混液通過將適量的聚鎢酸鈉 (71913,Sigma-Aldrich) 溶解于蒸餾水中制作,碳顆粒的平均直徑小于75μm(C3345,Sigma-Aldrich)。 模擬血流樣品由微量注射泵(LSP01-1A,Longerpump,China) 和10mL注射器驅(qū)動(dòng),采用內(nèi)徑為 0.95mm的透明塑料軟管作為模擬血管(Saint-GobainPerformancePlastics)。 實(shí)驗(yàn)中預(yù)設(shè)流速范圍16至 32mm/s,速度增量1.0mm/s,多普勒角度50°。

圖1 實(shí)驗(yàn)系統(tǒng)原理Fig.1 Experimental scheme
3結(jié)果與討論

圖2 時(shí)域光聲信號(hào).(a)碳顆粒懸混液的原始光聲信號(hào)(b)低通濾波后的光聲信號(hào)Fig.2 Time domain signal. (a)The original detected photoacoustic signal of carbon particles suspension (b) the low-pass filtered photoacoustic signal

圖3 速度矢量大小測(cè)量結(jié)果與預(yù)設(shè)值對(duì)比Fig.3 The measured flow vector values versus preset flow velocity
在圖2中,為了提高信號(hào)的信噪比,原始光聲信號(hào)(圖2a)在自相關(guān)計(jì)算前進(jìn)行了低通濾波(圖2b),濾波器上限截止頻率12.5 MHz。速度橫向、縱向分量及速度矢量大小的測(cè)量結(jié)果如圖3所示,速度矢量大小根據(jù)橫向和縱向速度分量的擬合直線得到。從圖3中可看出,速度縱向分量的擬合直線的線性度優(yōu)于橫向分量,其原因是碳顆粒的布朗運(yùn)動(dòng),懸混液在流動(dòng)過程中的速度梯度、湍流等效應(yīng)等所造成背景噪聲對(duì)頻譜展寬的影響較大,且其與橫向速度分量的大小無關(guān)。實(shí)驗(yàn)系統(tǒng)的空間分辨率是另一個(gè)影響測(cè)量精度的主要因素,主要取決于激光和聚焦超聲探頭焦點(diǎn)的大小。實(shí)驗(yàn)中,激光焦斑直徑為5 mm,10 MHz點(diǎn)聚焦超聲探頭的焦點(diǎn)直徑為0.23 mm,如采用更高中心頻率的超聲探頭提高實(shí)驗(yàn)系統(tǒng)的空間分辨率,可進(jìn)一步抑制頻譜背景噪聲對(duì)測(cè)量線性度的影響。此外,在互相關(guān)測(cè)量中[2-9],為了保證信號(hào)間的相關(guān)性并獲得較高時(shí)間分辨率的序列光聲信號(hào)對(duì),需采用重復(fù)頻率為數(shù)千赫茲脈沖激光,但光聲成像中常采用的高能脈沖激光的重復(fù)頻率一般為幾十赫茲。本文所采用的自相關(guān)方法基于對(duì)光聲信號(hào)的獨(dú)立A掃,自相關(guān)中的時(shí)移為自定義參數(shù),避免了對(duì)高重復(fù)頻率脈沖激光的要求,入射激光單脈沖能量較高,采集光聲信號(hào)中心頻率較低,有利于探測(cè)深度的提高。由橫向和縱向速度分量的擬合直線得到的多普勒角度為51.7°,根據(jù)速度矢量擬合直線計(jì)算的平均測(cè)量誤差大小為0.06 mm/s。結(jié)果表明,速度矢量大小與多普勒角度的測(cè)量值與預(yù)設(shè)值符合較好,該方法的可行性得到了初步驗(yàn)證。
4結(jié)論
對(duì)采用自相關(guān)方法的光聲多普勒速度矢量測(cè)量進(jìn)行了研究,光聲信號(hào)由脈沖激光激勵(lì)產(chǎn)生并由點(diǎn)聚焦超聲探頭采集,軸向和橫向速度分量分別通過光聲多普勒頻移和頻帶展寬得到。實(shí)驗(yàn)測(cè)量了預(yù)設(shè)流速為16 至 32 mm/s,多普勒角度為50°的碳顆粒懸混液的速度矢量大小和多普勒角度。該方法避免了對(duì)高重復(fù)頻率脈沖激光的要求,有利于探測(cè)深度的提高,其可行性通過實(shí)驗(yàn)得到了初步驗(yàn)證。
參考文獻(xiàn)
[1]ZHANG H F, MASLOV K,WANG L H. Photoacoustic doppler effect from flowing small light-absorbing particles [J]. Phys Rev Lett, 2007, 99(18):184501.
[2]YAO J , WANG L H. Transverse flow imaging based on photoacoustic Doppler bandwidth broadening [J]. J Biomed Opt, 2010, 15(2):021304.
[3]YAO J, MASLOV K, WANG L H.Invivophotoacoustic tomography of total blood flow and Doppler angle [J]. Proc of SPIE, 2012, 8223:82230U1.
[4]ZHANG R, YAO J, Maslov K I,etal. Structured-illumination photoacoustic Doppler flowmetry of axial flow in homogeneous scattering media [J]. Appl Phys Lett ,2013,103(9):94101.
[5]ZHANG R, YAO J, Maslov K Ietal.Photoacoustic Doppler axial flow measurement of homogenous media using structured illumination [J].Proc of SPIE, 2014, 8943:89431U.
[6]BRUNKER J, BEARD P. Acoustic resolution photoacoustic Doppler velocity measurements influids using time-domain cross-correlation [J]. Proc of SPIE, 2013, 8581:85811U.
[7]BRUNKER J, BEARD P. Pulsed photoacoustic Doppler flowmetry using time-domaincross-correlation:Accuracy, resolution and scalability [J]. J Acoust Soc Am, 2012,132(3):1780-1791.
[8]BRUNKER J, BEARD P. Acoustic resolution photoacoustic Doppler flowmetry:practical considerations for obtaining accurate measurements of blood flow [J]. Proc of SPIE, 2014, 8943:89431K.
[9]BRUNKER J, BEARD P. Pulsed photoacoustic Doppler flowmetry using a cross correlation method [J]. Proc SPIE, 2010, 7564:756426.
[10]SHEINFELD A, EYAL A. Laser-diode based 10MHz photoacoustic Doppler flowmetry at 830 nm [J]. Proc of SPIE, 2012, 8223:82233G.
[11]SHEINFELD A, GILEAD S, EYAL A . Coded Photoacoustic Doppler excitation with near-optimal utilization of the time and frequency domains [J]. Proc of SPIE, 2011, 7899:789928.
[12]KASAI C. Real-time two-dimensional blood-flow imaging using an autocorrelation technique [J]. IEEE Trans Ultrason Ferroelectr Freq Control, 1986, 33(1):458-464.
[13]PIAO D , ZHU Q. Quantifying Doppler angle and mapping flow velocity by a combination of Doppler-shift and Doppler-bandwidth measurements in optical Doppler tomography [J]. Appl Opt, 2003, 42(25):5158-5166.
[14]ZHAO Y, CHEN Z, SAXER C,etal. Doppler standard deviation imaging for clinical monitoring ofinvivohuman skin blood flow [J]. Opt Lett, 2000, 25(18):1358-1360.
[15]REN H, BRECKE K M, DING Z ,etal. Imaging and quantifying transverse flow velocity with the Doppler bandwidth in a phase-resolved functional optical coherence tomography [J]. Opt Lett, 2002, 27(6):409-411.
Photoacoutic Flow Velocity Vector Measurements Based on Doppler-shift and Dopper Bandwidth Broadening Using an Autocorrelation Method
LUTao*,BAIWeiwei
(College of Electrical Engineering, Henan University of Technology, Zhengzhou 450001, Henan, China)
Abstract:In order to measure the flow velocity vector of the carbon particles suspension, the autocorrelation method was introduced into in the photoacoustic Doppler flowmetry. The axial and transverse flow velocity component was measured from the photoacoustic Doppler frequency shift and Doppler bandwidth broadening, respectively. A 532 nm pulsed laser with the repetition rate of 20 Hz was used as a pumping source to generate the photoacoustic signal. The photoacoustic signal was then detected using a spherical focused PZT ultrasonic transducer with the central frequency of 10 MHz. The carbon particles suspension was driven by a syringe pump as the blood mimicking phantom. The complex photoacoustic signal was calculated by Hilbert transformation from the time domain signal before autocorrelation. Both the Doppler frequency shift and standard deviation of the Doppler spectrum bandwidth broadening was calculated by the autocorrelation of a series of individual A scans. Compared with the previously reported photoacoustic flow vector measurement using a cross-correlation method, the time delay in autocorrelation is user defined,the requirement of the high repetition rate pulsed laser is avoided and is also beneficial to the improvement of detection depth . The feasibility of the proposed method was preliminarily demonstrated by measuring the flow vector of the carbon particles suspension flow with the preset velocity from 16 to 32 mm/s and the Doppler angle of 50°.
Key words:photoacoustic Doppler; flow velocity vector; autocorrelation; doppler shift; doppler bandwidth broadening
文章編號(hào):1007-7146(2015)06-0519-04
文獻(xiàn)標(biāo)志碼:A
中圖分類號(hào):Q631
作者簡介:盧濤(1974-),男,河南鄭州人,副教授,工學(xué)博士,主要從事光聲成像與檢測(cè)研究。(電話)0371-67758835; (電子郵箱)hautlutao@hotmail.com
基金項(xiàng)目:NSFC-河南人才培養(yǎng)聯(lián)合基金項(xiàng)目(No. U1204612)
收稿日期:2015-08-24;修回日期:2015-10-13
doi:10.3969/j.issn.1007-7146.2015.06.005