激光多普勒测振啁啾噪声分析及其抑制方法
Based on the physical principle of the heterodyne laser Doppler vibration measurement process, we analyze the measurement process of high-frequency and low-speed movement in the presence of low-frequency and high-speed background movement. In the process, the measurement noise generated by the presence of stray light exhibits chirp characteristics, with the effects and patterns of chirp noise explained. In response to the chirp noise, we propose a derivative preprocessing method for demodulation. The theoretical analysis shows the method exerts a significant effect on suppressing chirp noise, which is verified by simulations and experiments. Meanwhile, we set a heterodyne laser Doppler vibration measurement system with stray light and measure the target vibration. The normal method and derivative preprocessing method are adopted respectively for demodulation. The experimental results verify the existence of chirp noise and the effectiveness of the derivative preprocessing method in suppressing chirp noise, which decreases the chirp noise power by about 81.8%. The method can effectively reduce the influence of stray light on vibration measurement.
Heterodyne laser Doppler vibration measurement technology is a widely adopted non-contact and non-destructive movement measurement method, with the advantages of fast response speed and high resolution. It has a strong detection ability for single frequency movement and can quickly identify the characteristic frequency of target movement. However, in the presence of low-frequency high-speed background movement, the measurement of high-frequency low-speed movement is severely affected by chirp noise, which is caused by stray light and closely related to low-frequency high-speed background movement. The chirp noise can seriously affect the measurement of high-frequency low-speed movement, with errors even reaching tens of times larger than those of real movement. There is a lack of research on the principle of chirp noise caused by stray light and suppression methods of chirp noise. We deeply analyze the principle of chirp noise caused by stray light and propose a novel demodulation method called the derivative preprocessing method (DPM). This demodulation method is easy to implement and exhibits a good effect for suppressing chirp noise. This demodulation method is expected to provide a reliable noise suppression method for measuring high-frequency low-speed movement in the presence of low-frequency high-speed background movement. This plays a significant role in analyzing high-frequency vibration modes of precision devices in some special measurement scenarios, such as in the presence of background movement.
Our study consists of theoretical analysis, simulation, and experimental verification. Firstly, the working principle and demodulation method of the heterodyne laser Doppler vibration measurement system, which is in the presence of stray light, are deeply analyzed. According to the analysis, the stray light would generate chirp noise in the process of normal demodulation method (ARCTAN). Based on the generation and characteristics of the chirp noise, a new demodulation method DPM is proposed. Then, the influence of chirp noise on the measurement of target movement velocity and the effect of DPM on chirp noise suppression are simulated. In the simulation, a low-frequency sinusoidal movement is utilized as the background movement, while a high-frequency sinusoidal movement is employed as the target movement. In the simulation, the background movement generates corresponding chirp noise to affect demodulation results severely. The normal demodulation method DPM is leveraged to restore the target movement by demodulating the overall movement and performing high-pass filtering. Finally, a heterodyne laser Doppler vibration measurement experiment is conducted to utilize a piezoelectric ceramic plate fixed on the pendulum device. In the experiment, the piezoelectric ceramic plate vibrates at a single and high frequency, which is regarded as the target movement, and the pendulum's movement is considered as the background movement. According to the experiment results, the existence of chirp noise is verified, and the DPM suppresses the chirp noise too.
In the process of the normal demodulation method (ARCTAN), the cause of chirp noise is the combination of stray light and background movement. The frequency of chirp noise changes in real time with the background movement speed, and specifically, it is proportional to the absolute speed of background movement and inversely proportional to the laser wavelength (Formula 11). When the background movement is low-frequency high-speed movement and the target movement is high-frequency low-speed movement, the chirp noise will seriously affect the measurement of high-frequency part movement [Fig. 2(b)]. Compared with the normal demodulation method, DPM can effectively restore target movement [Fig. 2(c)], but will generate erroneous velocity spikes in the positions that are near zero speed locations. Since adopting the normal demodulation method's results to partly replace the demodulation results of the DPM at the corresponding positions (near zero speed locations), the velocity spikes can be suppressed, and the demodulation results approach target movement more closely. In the experiment, the background movement of the pendulum indeed generates corresponding chirp noise [Figs. 4(a) and (b)], and DPM can effectively suppress the chirp noise [Fig. 4(c)]. DPM has a significant suppression effect on chirp noise, which reduces the power of chirp noise by about 81.8% at the peak of the chirp noise (Fig. 5).
The principle of heterodyne laser Doppler vibration measurement is deeply analyzed. It is pointed out that chirp noise is generated due to stray light and background movement. The frequency of the chirp noise changes in real time with the background movement speed, which is proportional to the absolute background movement speed and inversely proportional to the laser wavelength. Simulations and experiments have confirmed the existence of chirp noise and its frequency variation pattern. A novel demodulation method DPM has been proposed for this type of chirp noise. Simulations and experiments prove that DPM can effectively suppress chirp noise. Above the target movement frequency, the chirp noise power can be reduced by about 81.8%.
1 引言
外差激光多普勒测振技术通过外差干涉方式,利用多普勒效应获取待测物体的运动信息,具有非接触、响应速度快、无损伤、分辨率高等优点[1-3],在成分检测[4]、结构检测[5-6]、医学监测[7-10]等方面有着广阔的应用场景。
针对外差激光多普勒测振系统,研究者们分析了多种噪声来源,如幅度调制[11]、频率漂移[12]、散斑噪声[13-14]、散粒噪声[15]和背景运动[16]等。然而,在一种特定的测量需求——存在低频高速运动的情况下,对高频低速振动进行测量——的情况下(例如,对悬停直升机螺旋桨转轴、流水线上物体的高频振动进行实时测量),对杂散光产生噪声分析和抑制方法尚缺乏针对性的研究,因此本文将针对这种测量需求,对杂散光所产生的噪声的形式和性质等进行分析,并提出了一种针对性的抑制方法。
杂散光的可能来源较多,其中一些离轴方向上的杂散光可以通过吸收或偏折的方式进行抑制[17],但一些光学器件的折、反射产生的轴向杂散光将与测量光和参考光形成多光束干涉,从而带来测量误差。本文针对其中一种典型轴向杂散光——镜头后向反射光的存在,分析了其与测量光、参考光发生三光束干涉的情况,以及这种情况下其产生的测量误差的原理。
在分析了杂散光产生测量误差的基础上,本文针对特定的测量需求进行了理论分析,论证了此时杂散光导致的测量误差将以啁啾噪声的表现形式出现,并通过仿真程序直观展现了这种啁啾噪声的表现形式,并分析了其特点。
在明确了该啁啾噪声的产生原因和特点的基础上,本文提出了一种基于对原始信号进行微分预处理的新解调方法(以下称微分预处理方法),该方法可对该啁啾噪声进行抑制。本文通过仿真程序和实验验证,证明了这种微分预处理方法能够有效地抑制啁啾噪声。该方法有望在一些运动测量场景,例如对运动状态下的精密器件的高频振动模态分析中,起到明显的作用。
2 基本原理
2.1 外差激光多普勒测振信号解调原理
一种外差激光多普勒测振仪的基本结构如
图 1. 一种典型外差激光多普勒测振仪的主要组成部分示意图
Fig. 1. Schematic of the main composition of a typical heterodyne laser Doppler vibrometer
首先,将差频项分别乘以角频率为调制频率
之后通过反正切即可获取多普勒相移量
获取多普勒相移量
2.2 杂散光存在导致解调相位误差原理
在激光多普勒测振仪中,一些特定的杂散光将会以多光束干涉的形式影响测量过程,如
式中:
之后再求取反正切,就得到了解调相位,将其记为
显然,此时解调得到的相位
2.3 解调速度与真实速度之间的误差——啁啾噪声
物体位移和相位之间存在线性关系,因此
即解调速度
根据
而在信号光场强
显然,速度比率
而当物体的运动是由低频高速背景运动和高频低速待测运动组合而成时,根据
2.4 微分预处理方法抑制啁啾噪声原理
根据ARCTAN法的解调原理,啁啾噪声的产生是解调所用的正余弦项中多了和杂散光有关的正余弦项。基于这一点,提出了一种新的解调方式,由于其本质是将正余弦项通过微分方式进行预处理,因此将其称为微分预处理方法(DPM)。此方法的前置部分与常规ARCTAN解调方法相同,区别如下。
获取正余弦项之后,对
记新的解调相位为
此时的新解调相位
实际应用时,需要考虑
3 仿真分析
通过MATLAB程序,对存在杂散光和低频高速背景运动的情况下测量物体高频低速的待测运动时常规解调方法所产生的啁啾噪声的效果,以及微分预处理方法对啁啾噪声的抑制作用进行仿真。仿真中,选取一个振幅为10 nm、频率为10 kHz的正弦运动作为待测物体的待测运动,而附加的背景运动为一个振幅为5 mm、频率为0.5 Hz的正弦运动,信号光与杂散光的场强比值
通过常规解调方法和微分预处理方法,可以得到两个不同的解调结果。对常规方法解调速度、微分预处理方法解调速度的高频部分进行短时傅里叶变换,获取它们的时频图。为了使显示效果更明显,
图 2. 仿真结果。(a)常规方法得到的解调速度的绝对值;(b)常规方法得到的解调速度的时频图;(c)微分预处理方法得到的解调速度时频图
Fig. 2. Simulation results. (a) Absolute value of demodulated velocity obtained by normal method; (b) spectrogram of demodulated velocity obtained by normal method; (c) spectrogram of demodulated velocity obtained by DPM
从时频图中可以发现,在有杂散光的情况下,常规方法得到的解调速度的时频图中出现了额外的谱线,这就是啁啾噪声的直观体现,其与物体的正弦背景运动对应,直观地验证了
4 实验验证
4.1 实验设置
为了验证前文中对啁啾噪声的相关分析的正确性,以及微分预处理方法对抑制啁啾噪声的效果,搭建了一套外差激光多普勒测振系统进行验证实验。实验实物如
图 3. 微分预处理方法验证实验实物图。(a)测量目标;(b)测量系统
Fig. 3. Verification experiment objects of derivative preprocessing method. (a) Measurement target; (b) measurement system
在实验过程中,通过给压电陶瓷片施加频率为10 kHz、电压峰峰值为1 V的单频正弦电激励,使其产生基频为10 kHz的运动,10 kHz基频运动的速度相对摆锤的摆动来说属于高频低速运动。
实验时,先将实验装置布置在测振系统镜头外约5 m的位置,调整测振系统以使得测量光能够在摆锤自然悬垂、静止不动时,聚焦在压电陶瓷片表面中心附近。将摆锤由自然悬垂状态向一旁拨动约1 cm并固定,确认此时测量光仍能照射在压电陶瓷片表面上,之后通过信号发生器向压电陶瓷片施加电激励,并快速解除对摆锤的固定,摆开始运动。
摆开始运动后,通过该外差激光多普勒测振系统对待测目标进行测量,并使用
图 4. 验证实验结果图。(a)常规方法得到的解调速度绝对值;(b)常规方法得到的解调速度时频图;(c)部分替换后的微分预处理方法得到的解调速度时频图
Fig. 4. Verification experiment results. (a) Absolute value of demodulated velocity obtained by normal method; (b) spectrogram of demodulated velocity obtained by normal method; (c) spectrogram of demodulated velocity obtained by DPM after partial replacement
4.2 实验结果及分析
根据常规解调结果,剔除了速度过大或过小的部分数据,对剩余部分的数据分别使用常规解调方法和微分预处理方法进行解调,并对微分预处理解调结果在0速度附近处进行部分替换。
可以明显地看出,
从
图 5. 两种方法所得解调速度的功率谱密度。(a)常规方法得到的解调速度的功率谱密度;(b)部分替换后的微分预处理方法得到的解调速度的功率谱密度
Fig. 5. Power spectral density (PSD) of demodulated velocity of two methods. (a) PSD of demodulated velocity of normal method; (b) PSD of demodulated velocity of DPM after partial replacement
从
5 结论
本文通过深入分析外差式激光多普勒测振仪的工作原理、常用解调方法等,针对有低频高速背景运动情况下,对高频低速振动进行测量的需求,得到了因杂散光而产生啁啾噪声的原理和表现形式,并通过仿真对其进行了验证。理论分析和仿真结果表明,这种情况下因镜头后向反射的杂散光而产生的啁啾噪声,其瞬时频率与目标物体的瞬时运动速度成正比,与激光波长成反比。在有低频高速背景运动的情况下,这种啁啾噪声会对高频低速振动的测量产生严重干扰。
在此基础上,本文提出了一种全新的解调方法——微分预处理方法。仿真和实验结果表明,这种方法对啁啾噪声有明显的抑制效果。开展了激光多普勒测振实验,对摆动中的压电陶瓷片进行了测量,并使用常规解调方法和微分预处理方法分别进行解调。实验结果验证了啁啾噪声的存在,且其表现形式及特点和理论分析结果一致;而微分预处理方法也起到了对啁啾噪声明显的抑制,可使啁啾噪声功率下降约81.8%。
本文的研究内容还存在一定的局限性。首先,本文仅针对杂散光这一因素产生的噪声进行了分析,而未考虑其他因素对测量结果的影响,因此所得结果必然会与实际情况有一定的偏差。其次,本文提出的微分预处理方法的有效性是建立在回光功率较为稳定的情况下的,而在光强剧烈变化的环境下,这种方法的处理效果会下降。因此还需对其他噪声来源和光强变化等因素进行分析,完善相关模型,对微分预处理方法进行调整,以提高这种解调方法的鲁棒性。
[1] 张洪玮, 吴松华, 刘金涛, 等. 激光多普勒测速技术在海洋微尺度湍流测量中的可行性分析[J]. 光学学报, 2023, 43(24): 2401011.
[2] 陈兰剑, 席崇宾, 周健, 等. 无人机机载激光多普勒测速仪研究与飞行验证[J]. 光学学报, 2023, 43(17): 1712002.
[3] 张驰, 王顺, 关向雨, 等. 激光多普勒测振技术应用的新进展[J]. 激光与光电子学进展, 2022, 59(19): 1900006.
[4] 刘晓利, 王紫薇, 傅愉. 基于远距离激光测振仪的痕量危险品探测[J]. 光学学报, 2023, 43(13): 1312005.
[5] Maio L, Ricci F, Memmolo V, et al. Application of laser Doppler vibrometry for ultrasonic velocity assessment in a composite panel with defect[J]. Composite Structures, 2018, 184: 1030-1039.
[6] Balasubramaniam K, Fiborek P, Ziaja D, et al. Global and local area inspection methods in damage detection of carbon fiber composite structures[J]. Measurement, 2022, 187: 110336.
[7] Fredriksson I, Larsson M. On the equivalence and differences between laser Doppler flowmetry and laser speckle contrast analysis[J]. Journal of Biomedical Optics, 2016, 21(12): 126018.
[8] Sorelli M, Francia P, Bocchi L, et al. Assessment of cutaneous microcirculation by laser Doppler flowmetry in type 1 diabetes[J]. Microvascular Research, 2019, 124: 91-96.
[9] Dremin V, Kozlov I, Volkov M, et al. Dynamic evaluation of blood flow microcirculation by combined use of the laser Doppler flowmetry and high-speed videocapillaroscopy methods[J]. Journal of Biophotonics, 2019, 12(6): e201800317.
[10] 许越, 聂立铭. 光学成像技术在中医针刺研究中的应用进展[J]. 中国激光, 2023, 50(3): 0307105.
[11] 鹿彤彤, 陆彦婷, 杜福嘉, 等. 瞳面干涉激光多普勒测振[J]. 光学学报, 2022, 42(15): 1512005.
[12] Yang H X, Yin Z Q, Yang R T, et al. Design for a highly stable laser source based on the error model of high-speed high-resolution heterodyne interferometers[J]. Sensors, 2020, 20(4): 1083.
[13] Lü T, Han X Y, Wu S S, et al. The effect of speckles noise on the Laser Doppler vibrometry for remote speech detection[J]. Optics Communications, 2019, 440: 117-125.
[14] Jin Y, Dollevoet R, Li Z L. Numerical simulation and characterization of speckle noise for laser Doppler vibrometer on moving platforms (LDVom)[J]. Optics and Lasers in Engineering, 2022, 158: 107135.
[15] 蒲梦瑶, 胡以华, 曲芳慧, 等. 变速目标的光子回波外差信号处理方法研究[J]. 中国激光, 2023, 50(10): 1011002.
[16] Jiang L A, Albota M A, Haupt R W, et al. Laser vibrometry from a moving ground vehicle[J]. Applied Optics, 2011, 50(15): 2263-2273.
[17] 邵晶, 李卓, 聂真威, 等. 抑制光学系统中杂散光的非对称微结构的设计与加工[J]. 光学学报, 2023, 43(11): 1122001.
[18] 肖文哲, 程静, 张大伟, 等. 用于光纤干涉传感器的高稳定PGC解调技术[J]. 光电工程, 2022, 49(3): 210368.
Article Outline
王亚豪, 沈杨翊, 孔新新, 张文喜. 激光多普勒测振啁啾噪声分析及其抑制方法[J]. 光学学报, 2024, 44(5): 0507001. Yahao Wang, Yangyi Shen, Xinxin Kong, Wenxi Zhang. Chirp Noise Analysis in Laser Doppler Vibration Measurement and Its Suppression Methods[J]. Acta Optica Sinica, 2024, 44(5): 0507001.