Self-adaptive threshold peak searching method based on distributed FBG sensing system

Abstract

A self-adaptive threshold peak searching method based on a distributed FBG sensing system belongs to the technical field of optical fiber sensing, and comprises the following steps: carrying out noise reduction processing on a collected and normalized spectral signal by using a wavelet threshold denoising method to obtain a noise reduction signal; spectral signal pre-segmentation: constructing an adaptive threshold calculation model to pre-segment a spectrum; finely segmenting a fixed value bandwidth area of the spectrum; carrying out peak searching processing, and carrying out peak value positioning through a Gaussian fitting algorithm; according to the method, the FBG array reflection spectrum is processed by using the wavelet denoising algorithm of the improved threshold function and the twice segmentation algorithm based on the adaptive threshold, spectrum segmentation and peak value demodulation can be rapidly and stably realized, and the effect of real-time demodulation is achieved. According to the method, noise in spectral information can be well reduced, 3dB bandwidth areas of all sub-spectrums can be quickly and stably segmented, and the FBG array reflection spectrum can be quickly and stably demodulated.

Description

Technical Field

[0001] The present application relates to the field of optical fiber sensing technology, and in particular to an adaptive threshold peak finding method based on a distributed FBG sensing system. Background Art

[0002] Fiber Bragg Grating (FBG), as a fiber optic sensor, is widely used in various scenarios due to its reliability, accuracy, sensitivity, and anti-electromagnetic interference in harsh environments, such as structural health and safety monitoring, atrial fibrillation ablation surgery, and high-resolution esophageal manometry. The FBG sensing network indirectly detects changes in the external environment by establishing a functional relationship between the center wavelength drift value of the reflection spectrum and the change of environmental parameters. Therefore, it is very important to obtain the precise center wavelength value of the FBG reflection spectrum signal, which is directly related to the accuracy of the demodulation system.

[0003] At present, the peak-finding algorithms mainly include direct peak-finding algorithm, general polynomial fitting algorithm, centroid method, Gaussian fitting algorithm, Gaussian-polynomial fitting algorithm, genetic algorithm, and Steger-based peak-finding algorithm. Among them, the direct peak-finding method has low computational complexity and short response time, but its scope of application is limited. When the spectral resolution is low and contains a certain amount of noise, the peak-finding accuracy of this method is very low; the general polynomial fitting algorithm is simple to implement, but the peak-finding accuracy is limited; the Gaussian-polynomial fitting algorithm is between the Gaussian fitting algorithm and the general polynomial fitting algorithm, but this method is greatly affected by the observed data. If the peak is not within the observation point, the peak-finding error is large; the genetic algorithm can improve the peak-finding accuracy, but it takes a long time to train the model and is not suitable for real-time measurement; the Steger-based peak-finding algorithm introduces the idea of ​​image processing, and realizes the peak-finding of asymmetric peaks by extracting grayscale curves and combining super-Gaussian models, but the selection of its model parameters has a great influence on the accuracy of peak-finding, so the application scenario of this algorithm is greatly limited. The reflection spectrum of FBG sensor is approximately Gaussian. Usually, the Gaussian fitting algorithm is used to detect the peak value of FBG to obtain high peak detection accuracy. The Gaussian fitting algorithm is simple and has good anti-noise performance. However, the reflection spectrum of distributed FBG sensor network has multi-peak characteristics, and the above algorithms cannot effectively solve the problem of multi-peak detection.

[0004] To solve the above problems, patent CN115717906A discloses a FBG multi-peak spectrum peak finding method based on a symmetry-corrected Gaussian function. In its implementation case, the multi-peak spectrum is firstly directly Hilbert transformed and first-order derived, and then a fixed threshold is set to segment the spectrum. Finally, the precise peak position of the single-peak spectrum is calculated through the symmetry-corrected Gaussian function model.

[0005] Patent CN113358239A discloses a wavelength feature recognition method based on FBG. In its implementation case, the multi-peak spectrum is first smoothed and filtered to reduce noise, then the spectrum after noise reduction is Hilbert transformed and first-order derivative is taken, and the spectrum is segmented by selecting a fixed threshold, and finally the central wavelength is found by the Gaussian-LM algorithm.

[0006] However, patent CN115717906A discloses a FBG multi-peak spectrum peak finding method based on symmetry-corrected Gaussian function. In its implementation case, the multi-peak spectrum is first directly Hilbert transformed and first-order derivative, then a fixed threshold is set to segment the spectrum, and finally the precise peak position of the single-peak spectrum is calculated through the symmetry-corrected Gaussian function model. However, there are still the following problems to be solved.

[0007] (1) Patent CN115717906A does not perform noise reduction processing before spectral signal processing, which easily leads to high noise in subsequent spectral signal processing, thereby causing spectrum segmentation failure or increased peak search error.

[0008] (2) Patent CN115717906A requires manual selection of a fixed threshold before segmenting a multi-peak spectrum. When faced with light noise interference or when the spectrum fluctuates greatly, segmentation failure is prone to occur and the fixed threshold is not suitable for real-time detection.

[0009] Patent CN113358239A discloses a wavelength feature recognition method based on FBG. In its implementation case, the multi-peak spectrum is first smoothed and filtered to reduce noise, then the spectrum after noise reduction is Hilbert transformed and first-order derivative, and the spectrum is segmented by selecting a fixed threshold, and finally the central wavelength is found by the Gaussian-LM algorithm. However, there are still the following problems to be solved.

[0010] (1) Although patent CN113358239A uses a five-point sliding average filter method to reduce noise on the original signal, the sliding average filter method is not easy to eliminate the sampling value deviation caused by interference and is not suitable for real-time detection that requires a faster data calculation speed.

[0011] (2) Patent CN113358239A's Hilbert transform peak splitting method requires manual selection of appropriate thresholds for spectrum segmentation, which is inefficient and not suitable for situations where the spectrum changes during real-time measurement. Summary of the invention

[0012] The present application provides an adaptive threshold peak-finding method based on a distributed FBG sensor system, which can solve the problem that the fixed threshold for segmenting the spectrum peak in the peak-finding algorithm based on Hilbert transform in the related art fails under the influence of certain noise. The technical solution is as follows:

[0013] According to one aspect of the present application, an adaptive threshold peak finding method based on a distributed FBG sensor system comprises the following steps:

[0014] The wavelet threshold denoising method is used to denoise the collected and normalized spectral signal to obtain a denoised signal;

[0015] Spectral signal pre-segmentation, building an adaptive threshold calculation model to pre-segment the spectrum;

[0016] Fine segmentation of fixed bandwidth regions of the spectrum;

[0017] Peak finding processing, peak location through Gaussian fitting algorithm.

[0018] In an exemplary embodiment, the wavelet threshold denoising method includes:

[0019] Threshold function:

[0020]

[0021] Among them, d j,k represents the wavelet coefficients in scalar j, represents the quantized wavelet coefficient, j represents the decomposition scale, γ is the threshold, and is defined as σ is the standard deviation of the noise, estimated by the high-frequency coefficient: σ = median(D j ) / 0.6754, where D j is the coefficient of the diagonal detail component of layer j, a and b are two adjustment parameters, and the value of a and b can be adjusted by adjusting the value of a and b. The degree of convergence to the soft threshold function;

[0022] In an exemplary embodiment, the fixed value bandwidth of the optical spectrum is a 3 dB bandwidth.

[0023] In an exemplary embodiment, the method of pre-segmenting a spectral signal and constructing an adaptive threshold calculation model to pre-segment the spectrum includes:

[0024] First, the spectral reflection signal after noise reduction is Hilbert transformed. After Hilbert transform, the amplitude and frequency components of the signal remain unchanged, but the phase will have a 90° shift. The Hilbert transform of the real-valued function signal x(t) is shown in formula (2):

[0025]

[0026] Calculate the first-order derivative and build an adaptive threshold calculation model:

[0027]

[0028] hsort =[h'] max→min (4)

[0029] γ auto =h sort [l] (5)

[0030] Where: l represents the length of the array for solving the adaptive threshold, n represents the number of FBG sensors, Δλ i represents the 3 dB bandwidth value of the i-th FBG, Δλ represents the acquisition bandwidth setting value of the spectrometer, N represents the number of sampling points set by the spectrometer, β represents the adaptive factor, h' represents the first-order derivative data of the Hilbert transform, and h sort represents the array after h' is sorted from large to small, γ auto Represents the adaptive threshold. As shown in formula (3), we estimate the 3dB bandwidth effective information length by multiplying the ratio of the 3dB total bandwidth of the FBG array to the spectrum acquisition bandwidth by N and β, where β is generally 2. In order to find the adaptive threshold, the data is arranged from large to small to obtain the array h sort As shown in formula (4). Finally, the adaptive threshold γ is obtained by formula (5): auto , and determine whether the threshold is met.

[0031] In an exemplary embodiment, the threshold determination method includes:

[0032] It is determined whether the threshold is satisfied to segment n sub-spectra (segmentation failure occurs if the number of sub-spectra is greater than n or less than n). If so, the pre-segmentation is completed.

[0033] In an exemplary embodiment, if it is not satisfied, β is reduced by 0.1, and the threshold is calculated again for judgment until the pre-segmentation is completed.

[0034] In an exemplary embodiment, the power peak of each sub-spectral region is obtained through pre-segmentation, and then the half-power point is obtained, and the spectrum is expanded based on this point to achieve the segmented sub-spectrum being stable in the 3dB bandwidth region.

[0035] In an exemplary embodiment, the peak finding process, through the Gaussian fitting algorithm, the peak location method includes:

[0036] Since the collected FBG reflection signal has Gaussian distribution characteristics, the segmented FBG sub-spectrum can be approximately represented by a Gaussian function, as shown in formula (6):

[0037]

[0038] Where A is the amplitude of the FBG reflection spectrum, λ b is the Bragg wavelength, Δλ b is 3dB bandwidth;

[0039] Peak finding processing, through Gaussian fitting algorithm, the peak location methods include:

[0040] The spectral fitting error is shown in formula (7):

[0041]

[0042] Where X = (A, λ b ,Δλ b ) T and (x i ,y i ) is the sampling point;

[0043] According to the least squares method, the sum of squares of the deviations is as shown in formula (8):

[0044]

[0045] When S reaches the minimum value, the central wavelength λ of the FBG reflection spectrum signal can be calculated. b .

[0046] In an exemplary embodiment, it also includes spectral signal acquisition and normalization processing; the FBG array sensor reflection spectrum is acquired by a spectrometer and normalized.

[0047] In an exemplary embodiment, a plurality of FBG sensors connected in series are placed in a constant temperature environment, and then an ASE broadband light source is used to send light to the FBG sensor array through a circulator, and then through an optical coupler, the reflected signal is transmitted to a demodulator to read the central wavelength of the above-mentioned spectral signal as a theoretical wavelength value, and the reflected signal is transmitted to a spectrometer to obtain a reflected spectrum and transmitted to a computer for peak finding processing.

[0048] In summary, in some exemplary embodiments, the use of improved threshold wavelet transform and the two-segmentation method based on adaptive threshold can greatly improve the anti-noise performance to better meet the real-time multi-peak spectrum demodulation in various situations.

[0049] The beneficial effects of the technical solution provided by this application are:

[0050] 1. This application utilizes a wavelet denoising algorithm with an improved threshold function and a two-segmentation algorithm based on an adaptive threshold to process the FBG array reflection spectrum, which can quickly and stably achieve spectrum segmentation and peak demodulation, achieving a real-time demodulation effect.

[0051] 2. This application can effectively reduce the noise in the spectral information, and can quickly and stably segment the 3dB bandwidth area of ​​all sub-spectra, and has realized the fast and stable demodulation of the FBG array reflection spectrum. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can obtain other drawings based on these drawings without creative work.

[0053] Figure 1 is a system diagram of an exemplary embodiment of the present application;

[0054] Figure 2 is an algorithm flow chart according to an exemplary embodiment of the present application;

[0055] Figure 3 is a spectrum diagram of an exemplary embodiment of the present application;

[0056] Figure 4 is a pre-segmented graph of an exemplary embodiment of the present application;

[0057] Figure 5 is a refined segmentation diagram of an exemplary embodiment of the present application;

[0058] Figure 6 It is a flowchart of an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0059] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present application.

[0060] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present disclosure refers to the presence of the features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0061] See also Figure 6, an adaptive threshold peak finding method based on a distributed FBG sensor system includes the following steps:

[0062] S1, using a wavelet threshold denoising method to perform denoising on the collected and normalized spectral signal to obtain a denoised signal;

[0063] S2, pre-segmentation of spectral signals, building an adaptive threshold calculation model to pre-segment the spectrum;

[0064] S3, fine segmentation of fixed bandwidth region of the spectrum;

[0065] S4: Peak finding processing: peak location is performed through Gaussian fitting algorithm.

[0066] In an exemplary embodiment, if Figure 1 As shown, four FBG sensors connected in series are placed in a constant temperature environment, and then the ASE broadband light source sends light to the FBG sensor array through a circulator, and then through an optical coupler, the reflected signal is transmitted to the demodulator to read the central wavelength of the above spectral signal as the theoretical wavelength value, and the reflected signal is transmitted to the spectrometer AQ6370D to obtain the reflected spectrum and transmit it to the computer for peak searching.

[0067] In an exemplary embodiment, if Figure 2 As shown in the figure, suppose the FBG array sensor has n FBG sensors. First, the FBG array spectrum is obtained by the spectrometer for normalization, and then wavelet denoising is performed. The signal after denoising is Hilbert transformed and its first-order derivative is calculated. The adaptive threshold is calculated and set to determine whether the threshold satisfies the requirement of segmenting n sub-spectra (the number of sub-spectra greater than n or less than n is considered segmentation failure). If the segmentation fails, the adaptive factor is automatically adjusted to recalculate the threshold until the condition is met. If the segmentation is successful, the next stage of spectrum expansion is performed to achieve the stability of the segmented spectrum in the 3dB bandwidth area. Finally, the peak values ​​of all sub-spectra are searched through the Gaussian fitting algorithm.

[0068] In an exemplary embodiment, see Figure 2 :Adaptive threshold peak finding method based on distributed FBG sensing system:

[0069] Step 1: Spectral signal acquisition and normalization processing. The FBG array sensor reflection spectrum is collected by the spectrometer and normalized. After normalization, the spectrum is as follows Figure 3 (a) shown.

[0070] Step 2: Use the improved wavelet transform to perform noise reduction on the normalized data. The improved threshold function is shown in formula (1):

[0071]

[0072] where d j,k represents the wavelet coefficients in scalar j, represents the quantized wavelet coefficient, j represents the decomposition scale, γ is the threshold, and is defined as σ is the standard deviation of the noise, estimated by the high-frequency coefficient: σ = median(D j ) / 0.6754, where D j is the coefficient of the diagonal detail component of layer j, a and b are two adjustment parameters, and the value of a and b can be adjusted by adjusting the value of a and b. The degree of convergence to the soft threshold function. The effect of wavelet transform denoising after improving the threshold is shown in 3(b).

[0073] Step 3: Reflection spectrum segmentation. First, perform Hilbert transform on the denoised spectral reflection signal. After Hilbert transform, the amplitude and frequency components of the signal remain unchanged, but the phase will have a 90° shift. The Hilbert transform of the real-valued function signal x(t) is shown in formula (2):

[0074]

[0075] After Hilbert transformation Figure 3 (c) and then take its first-order derivative, the result is Figure 3 (d) It can be observed that although most of the signal noise is filtered out by the noise reduction algorithm, there is still a small amount of noise that affects the spectrum shape, resulting in Figure 3 (d) There is "burr" interference, and this "burr" is not fixed in different environments and at different times, so it is difficult to perform real-time measurement by using a fixed threshold or manually selecting a threshold. Figure 3 (b) and Figure 3 (d) The similarity of the spectrum shape can be deduced from the total bandwidth of the FBG array sensor. Figure 3 The adaptive threshold of (d) is shown in equation (3), equation (4) and equation (5):

[0076]

[0077] h sort =[h'] max→min (4)

[0078] γ auto =h sort [l] (5)

[0079] Where l represents the length of the array for solving the adaptive threshold, n represents the number of FBG sensors, Δλ irepresents the 3 dB bandwidth value of the i-th FBG, Δλ represents the acquisition bandwidth setting value of the spectrometer, N represents the number of sampling points set by the spectrometer, β represents the adaptive factor, h' represents the first-order derivative data of the Hilbert transform, and h sort represents the array after h' is sorted from large to small, γ auto Represents the adaptive threshold.

[0080] As shown in formula (3), the ratio of the 3 dB total bandwidth of the FBG array to the spectrum acquisition bandwidth is multiplied by N and β to estimate Figure 3 (d) is the effective information length of 3dB bandwidth, where β is generally set to 2. In order to find the adaptive threshold, Figure 3 The data in (d) are arranged from large to small to get the array h sort As shown in formula (4). Finally, the adaptive threshold γ is obtained by formula (5): auto , and judge whether the threshold meets the requirement of segmenting n sub-spectra (the number of sub-spectra is greater than n or less than n, which means segmentation failure). If not, β is reduced by 0.1, and the threshold is calculated again for judgment. If it is satisfied, the pre-segmentation is completed. Figure 4 Then, through pre-segmentation, we can find the power peak of each sub-spectral region, and then find the half-power point, and use this point as the standard to expand the spectrum, so that the segmented sub-spectrum is stable in the 3dB bandwidth area. The fine segmentation is as follows: Figure 5 As shown in (a), (b), (c) and (d).

[0081] Step 4: Peak search. Since the collected FBG reflection signal has Gaussian distribution characteristics, the segmented FBG sub-spectrum can be approximately represented by a Gaussian function, as shown in formula (6):

[0082]

[0083] Where A is the amplitude of the FBG reflection spectrum, λ b is the Bragg wavelength, Δλ b is 3dB bandwidth.

[0084] The spectral fitting error is shown in formula (7):

[0085]

[0086] Where X = (A, λ b ,Δλ b ) T and (x i ,y i ) are sampling points.

[0087] According to the least squares method, the sum of squares of the deviations is as shown in formula (8):

[0088]

[0089] When S reaches the minimum value, the central wavelength λ of the FBG reflection spectrum signal can be calculated. b .

[0090] In summary, by improving the threshold function wavelet transform denoising and the adaptive threshold calculation model and the two-segmentation method, the optimal segmentation threshold point can be quickly found and the sub-spectral 3dB bandwidth segmentation can be achieved, thereby improving the effectiveness of real-time detection of FBG array sensors.

[0091] In some exemplary embodiments, the method may also be applicable to all other grating types with similar spectral shapes;

[0092] Therefore, compared with the artificial threshold selection method designed in patent CN115717906A, the present application has faster speed and more universal characteristics, and has obvious advantages when facing real-time measurement.

[0093] Compared with the five-point sliding filter denoising method in patent CN113358239A, this application has a better denoising effect and a wider application scenario than the five-point sliding filter method.

[0094] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0095] The above description is only a partial implementation method of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An adaptive threshold peak finding method based on a distributed FBG sensor system, characterized in that: The steps include: The wavelet threshold denoising method is used to denoise the collected and normalized spectral signal to obtain a denoised signal; Spectral signal pre-segmentation, building an adaptive threshold calculation model to pre-segment the spectrum; Fine segmentation of fixed bandwidth regions of the spectrum; Peak finding processing, peak location through Gaussian fitting algorithm.

2. The adaptive threshold peak finding method based on the distributed FBG sensor system according to claim 1 is characterized in that: The wavelet threshold denoising method comprises: Threshold function: Among them, d j,k represents the wavelet coefficients in scalar j, represents the quantized wavelet coefficient, j represents the decomposition scale, γ is the threshold, and is defined as σ is the standard deviation of the noise, estimated by the high-frequency coefficient: σ = median(D j ) / 0.6754, D j is the coefficient of the diagonal detail component of layer j, a and b are two adjustment parameters, which are adjusted by adjusting the size of a and b. The degree of convergence to the soft threshold function.

3. The adaptive threshold peak finding method based on the distributed FBG sensor system according to claim 1 is characterized in that: The fixed value bandwidth of the spectrum is a 3dB bandwidth.

4. The adaptive threshold peak finding method based on the distributed FBG sensor system according to claim 3 is characterized in that: The method for pre-segmenting the spectral signal and constructing an adaptive threshold calculation model to pre-segment the spectrum includes: First, perform Hilbert transform on the denoised spectral signal. The Hilbert transform of the spectral signal x(t) is shown in formula (2): Calculate the first-order derivative and build an adaptive threshold calculation model: h sort =[h'] max→min (4) c auto =h sort [l] (5) Where: l represents the length of the array for solving the adaptive threshold, n represents the number of FBG sensors, Δλ i represents the 3 dB bandwidth value of the i-th FBG, Δλ represents the acquisition bandwidth setting value of the spectrometer, N represents the number of sampling points set by the spectrometer, β represents the adaptive factor, h' represents the first-order derivative data of the Hilbert transform, and h sort represents the array after h' is sorted from large to small, γ auto Represents the adaptive threshold.

5. The adaptive threshold peak finding method based on the distributed FBG sensor system according to claim 4 is characterized in that: The threshold determination method includes: It is determined whether the threshold is satisfied to segment n sub-spectra. If the number of sub-spectra is greater than n or less than n, the segmentation fails. If it is satisfied, the pre-segmentation is completed.

6. The adaptive threshold peak finding method based on the distributed FBG sensor system according to claim 5 is characterized in that: If it is not satisfied, β is reduced by 0.1, and the threshold is calculated again for judgment until the pre-segmentation is completed.

7. The adaptive threshold peak finding method based on a distributed FBG sensor system according to any one of claims 3 to 6, characterized in that: The power peak of each sub-spectral area is obtained through pre-segmentation, and then the half-power point is obtained, and the spectrum is expanded based on this point to achieve the stability of the segmented sub-spectrum in the 3dB bandwidth area.

8. The adaptive threshold peak finding method based on the distributed FBG sensor system according to claim 7 is characterized in that: The peak finding process, through the Gaussian fitting algorithm, the peak location method includes: The segmented FBG sub-spectrum is represented by a Gaussian function, as shown in formula (6): Where A is the amplitude of the FBG reflection spectrum, λ b is the Bragg wavelength, Δλ b is 3dB bandwidth; The spectral fitting error is shown in formula (7): Where X = (A, λ b ,Δλ b ) T and (x i ,y i ) is the sampling point; According to the least squares method, the sum of squares of the deviations is as shown in formula (8): When S reaches the minimum value, the central wavelength λ of the FBG spectrum signal is calculated b .

9. The adaptive threshold peak finding method based on a distributed FBG sensor system according to claim 8, characterized in that: It also includes spectral signal acquisition and normalization processing; the FBG array sensor reflection spectrum is acquired through a spectrometer and normalized.

10. The adaptive threshold peak finding method based on a distributed FBG sensor system according to claim 8, characterized in that: It also includes placing several FBG sensors connected in series in a constant temperature environment, and then sending light from a broadband light source to the FBG sensor array through a circulator, and then transmitting the signal through an optical coupler to a demodulator to read the central wavelength of the spectral signal as a theoretical wavelength value, and then transmitting the signal to a spectrometer to obtain a spectrum and transmit it to a computer for peak searching.

Images