Image-based laser Brillouin scattering spectrum high-precision reconstruction system

Abstract

The invention discloses an image-based high-precision reconstruction system for a laser Brillouin scattering spectrum. The system comprises an echo beam expanding and shaping unit, a frequency discrimination unit, a light path switching device, an image sensor and an image acquisition and data processing unit. The echo beam expanding and shaping unit expands and collimates the scattered echo and then emits the scattered echo into the frequency discrimination unit; the frequency discrimination unit performs multi-beam interference on the scattered echoes after beam expansion and collimation to form linear interference fringes and then emits the linear interference fringes; the light path switching device reflects the linear interference fringes to the image sensor and senses light on the image sensor; the image acquisition and data processing unit acquires a photosensitive result, and the photosensitive result is processed by an interference spectrum extraction algorithm to obtain a device function and a wedge angle corresponding to interference fringes. According to the invention, the image sensor is adopted to replace a multi-channel photoelectric detector to receive interference fringes for calibration, the spatial resolution of calibration receiving is improved, and the calibration precision is improved.

Description

Technical Field

[0001] The invention belongs to the field of laser hyperspectral detection and relates to a spectrum reconstruction system for laser, which is used to provide calibration input for a spectrum identification system and improve the measurement accuracy of underwater Brillouin scattering spectrum. Background Art

[0002] The blue-green band with a wavelength of around 470 to 580 nm is a transmission window with a small attenuation coefficient of seawater. The blue-green laser in this band can undergo a variety of scattering processes with the water body. Among them, by receiving and analyzing the inelastic scattering echoes between the laser and the seawater molecules, namely the Brillouin scattering and Raman scattering of water molecules, it can be used to invert the two seawater environmental parameters of temperature and salinity. In addition, since the bandwidth of Brillouin scattering is narrower and less susceptible to background signals, it has a higher detection potential. There are many types of spectral identification systems that invert seawater temperature and salinity through Brillouin scattering. They invert the temperature and salinity information of seawater by obtaining the frequency shift of Brillouin scattering relative to the frequency of the emitted laser and the two spectral line parameters of the Brillouin scattering spectrum.

[0003] The spectrum identification system of Fertz interferometer combined with multi-channel photoelectric detector has the characteristics of strong real-time performance and can obtain complete Brillouin spectral lines. However, in practical applications, the resolution of the spectrum identification system of Fertz interferometer combined with multi-channel photoelectric detector is limited by many factors. Among them, if the prior information of Fertz interferometer itself, such as interferometer wedge angle and interferometer device function, cannot be accurately determined, it will introduce systematic errors in the process of extracting Brillouin scattering spectrum line parameters.

[0004] In the existing Brillouin spectroscopy identification system using Fertz interferometer, only a multi-channel photodetector for detecting echo signals can be used to perform system self-calibration. Although the multi-channel photodetector can detect weak signals and has high time resolution, it has a small number of channels and low spatial resolution. Summary of the invention

[0005] The technical problem solved by the present invention is: to overcome the shortcomings of the prior art and provide an interferometer parameter calibration system for a spectral identification system of a Fertz interferometer combined with a multi-channel photoelectric detector, by adding an image sensor receiving channel to the calibration process of the spectral identification system, collecting high spatial resolution image information of interference fringes, improving the calibration accuracy of the wedge angle and device function of the Fertz interferometer, thereby obtaining more reliable Brillouin scattering frequency shift and line width extraction accuracy.

[0006] The technical solution of the present invention is: a high-precision reconstruction system of laser Brillouin scattering spectrum based on image, including an echo beam expansion shaping unit, a frequency discrimination unit, an optical path switching device, an image sensor, an image acquisition and data processing unit, wherein:

[0007] Echo beam expansion and shaping unit: expands and collimates the scattered echo and then injects it into the frequency discrimination unit;

[0008] Frequency discrimination unit: multi-beam interference is performed on the scattered echo after beam expansion and collimation to form linear interference fringes and then emit;

[0009] An optical path switching device: used for controlling the reflection of the linear interference fringes to the image sensor and sensing light on the image sensor;

[0010] Image acquisition and data processing unit: collects and stores the photosensitivity results of the image sensor, and obtains the device function and wedge angle corresponding to the interference fringes through the interference spectrum extraction algorithm.

[0011] Furthermore, a beam shaping unit is included between the frequency discrimination unit and the optical path switching device to expand or shrink the light beam passing through the frequency discrimination unit to match the size of the image sensor.

[0012] Preferably, after the echo beam expansion and shaping unit expands and collimates the scattered echo, the divergence angle entering the frequency discrimination unit is in the order of milliradians.

[0013] Preferably, the frequency discrimination unit is a Fezzo interferometer.

[0014] Furthermore, the direction of the linear interference fringes is parallel to the column direction of the image sensor pixels.

[0015] Furthermore, the image acquisition and data processing unit obtains the device function and wedge angle corresponding to the interference fringes through the interference spectrum extraction algorithm, specifically:

[0016] (61) averaging the brightness values ​​of pixels in each row of the image collected by the image sensor, and compressing the two-dimensional interference fringes into a one-dimensional interference spectrum; the image collected by the image sensor is obtained by irradiating a diffuse reflection target with a fixed-frequency laser emitted by a laser, and the echo signal of the diffuse reflection target is obtained after passing through a frequency discrimination unit and an optical path switching device in sequence;

[0017] (62) near the brightness peak of the one-dimensional interference spectrum, obtain the center position x0 of the interference fringes in pixels;

[0018] (63) changing the frequency of the laser emitting laser light, repeating steps (61) to (62), obtaining relationship data with the center position of the interference fringes as the independent variable and the frequency offset of the laser emitting laser light as the dependent variable, and obtaining the fitting slope value k by linear function fitting;

[0019] (64) The wedge angle α of the frequency discrimination unit is calculated using the absolute value of the fitting slope value |k|.

[0020]

[0021] Where n is the refractive index of the medium between the two reflection planes of the frequency discrimination unit, λ0 is the laser wavelength, and v FSR is the free spectral range of the frequency discrimination unit;

[0022] (65) The one-dimensional interference spectrum is normalized, and then the horizontal axis of the one-dimensional interference spectrum is adjusted to take the relative frequency as the independent variable, and the relative frequency corresponding to the center position of the fringe is made 0; the one-dimensional interference spectrum is fitted with a function to obtain the fitted spectral line width Γ x , using the line width Γ x The device function Γ of the frequency discrimination unit is calculated by using the absolute value of the fitting slope value |k| v =Γ x *|k|.

[0023] Furthermore, the central position x0 of the interference fringes in pixels is obtained by: the signal intensity vector of each column in the one-dimensional interference spectrum selection area is expressed as Record the pixel position of each column as Calculate the centroid position as

[0024]

[0025] The centroid x p The dimension is length, x0=d0x p , d0 is the pixel size of the image sensor.

[0026] Furthermore, the acquisition of the center position x0 of the interference fringes in pixels is specifically as follows: the Lorentz function is used to fit the centroid x p , the Lorentz function is in the form

[0027]

[0028] where x p Represents the pixel row position corresponding to the peak intensity of the one-dimensional interference spectrum, Γ x is the half-height full width occupied by the function fitting on the interference spectrum, c0 represents the lift of the dark level and background signal, I x Indicates the strength of the signal; x0 = d0x p , d0 is the pixel size of the image sensor.

[0029] Preferably, the frequency of laser emission by the laser is changed in steps of 0.2 to 1.0 GHz.

[0030] Preferably, the image sensor adopts a CCD or CMOS device, and the size of the photosensitive surface is larger than the coverage range of the linear interference fringes.

[0031] The advantages of the present invention compared with the prior art are:

[0032] (1) The system of the present invention aims at the problem that the calibration accuracy of the wedge angle and device function information of the Fertz interferometer is not high in the spectral identification system of the traditional Fertz interferometer combined with a multi-channel photoelectric detector. It proposes to use an image sensor instead of a multi-channel photoelectric detector to receive the interference fringes for calibration, thereby improving the spatial resolution of the calibration reception and the calibration accuracy. Higher calibration information input reduces the systematic error of the inverted Brillouin scattering spectral line parameters when the Fertz interferometer is combined with a multi-channel photoelectric detector;

[0033] (2) The system of the present invention introduces an optical path switching device, so that the interferometer transmission signal can be switched between the image sensor used in the present invention and the original multi-channel photodetector in the spectral identification system, and can be integrated into the existing mature spectral identification system without destroying its original detection function. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is the composition principle diagram of the system of the present invention;

[0035] Figure 2 The present invention is a schematic diagram of a spectrum identification system including the system of the present invention;

[0036] Figure 3 Schematic diagram of two-dimensional interference fringes and one-dimensional interference spectrum in an embodiment of the present invention;

[0037] Figure 4 Schematic diagram of the photosensitive surface of a multi-channel image sensor in an embodiment;

[0038] Figure 5 Flowchart of generation and processing of simulation signals in the embodiment;

[0039] Figure 6 It is a comparison diagram of the fitting effect of the center position of the stripes in the embodiment;

[0040] Figure 7 It is a comparison diagram of the stripe line width fitting effect in the embodiment;

[0041] Figure 8 It is a comparison diagram of the frequency-space mapping slope |k| fitting results in the embodiment. DETAILED DESCRIPTION

[0042] In the calibration process of the Fizeau interferometer, the echo energy is significantly stronger than the scattered signal in the detection application, and there is no requirement for time resolution. Therefore, image sensors with high spatial resolution, lower time resolution, and poorer weak signal detection capabilities are more suitable for signal acquisition and processing in the calibration process.

[0043] The present invention enables the original Brillouin scattering spectrum identification system to have a stronger self-calibration capability by adding an optical path switching device and an image monitoring channel for calibration of relevant parameters of the Fertz interferometer to the system, and adding a corresponding image processing algorithm. After performing the calibration measurement, by adjusting the optical path switching device, the interference fringes are switched from the image monitoring channel to be received by the multi-channel photoelectric detector, and the original measurement process can be continued. Therefore, the system of the present invention can work in coordination with the existing Brillouin spectrum identification system of the Fertz interferometer combined with the multi-channel detector.

[0044] The pixel size of common image sensors is on the order of several microns. Compared with multi-channel photodetectors with pixel sizes on the order of millimeters, the spatial resolution is improved by two orders of magnitude, and more detailed interference fringe information can be received, thereby improving the calibration accuracy of the interferometer parameters.

[0045] like Figure 1 As shown, the system of the present invention comprises an echo beam expansion shaping unit 1, a frequency discrimination unit 2, an optical path switching device 3, an image sensor 4, an image acquisition and data processing unit 5. In addition, there is an optional beam shaping unit 6.

[0046] The echo beam expansion and shaping unit 1 expands and collimates the scattered echo used for Fezzo interferometer calibration, and then injects it into the discrimination unit 2 at a divergence angle of the order of milliradian. The echo enters the discrimination unit 2, and after the multi-beam interference process, linear interference fringes are formed, and then introduced into the image sensor 4 after being reflected by the optical path switching device 3. The interference fringes are photosensitive on the image sensor 4, and after being collected and stored by the image acquisition and data processing unit 5, the device function, wedge angle and other information corresponding to the interference fringes are obtained through the interference spectrum extraction algorithm. When the image sensor 4 does not match the input beam size, that is, the size of the interference fringes formed is significantly larger or smaller than the size of the photosensitive surface of the image sensor 4, an additional beam shaping unit 6 can be inserted between the discrimination unit 2 and the optical path switching device 3 to expand or shrink the beam passing through the discrimination unit 2 to match the size of the image sensor 4. If the beam size does not need to be adjusted, the beam shaping unit 6 does not need to be introduced.

[0047] The echo beam expansion and shaping unit 1 can adopt different configurations according to different input signals. When the echo beam expansion and shaping unit 1 receives the echo signal of the fiber coupling input, the unit is composed of a fiber flange, a collimating beam expansion optical lens, and an aperture. When the echo beam expansion and shaping unit 1 receives the free space input, the unit is composed of a collimating beam expansion optical lens and an aperture. After passing through the echo beam expansion and shaping unit 1, the laser spot just covers the aperture of the frequency discrimination unit 2, and the exit divergence angle should be as small as possible, which needs to reach the milliradian level.

[0048] The reflection cavity of the demodulator unit 2 is composed of two reflection planes with a small angle, and the angle is in the order of micro-radians. This structure is usually called a Fizeau Interferometer. The reflection plane is plated with a partial reflection film composed of multiple layers of dielectrics or metals, and the space between the two planes is composed of gas or solid optical glass. The cavity length and angle of the reflection cavity can be designed as a non-adjustable structure, or a structure that can be adjusted by mechanical motion. When the scattered echo excited by the narrow linewidth laser is shaped into a beam of light close to parallel light and vertically enters the Fizeau interferometer, it is reflected multiple times in the reflection cavity and then emitted from the rear surface of the Fizeau interferometer and coherently superimposed to form linear interference fringes with periodic changes in light and dark, and the direction of the stripes is perpendicular to the direction of the small angle. The striped laser spot generated by the demodulator unit 2, the direction parallel to the striped spot represents the energy distribution, and the direction perpendicular to the striped spot corresponds to the spectral distribution of the incident light. By adjusting the angle direction of the front and rear reflection planes of the reflection cavity of the frequency discrimination unit 2, the direction of the linear stripe spot is parallel to the column direction of the 4 pixels of the image sensor. At this time, the positions of different columns of the same row of pixels correspond to the frequency spectrum of the echo signal.

[0049] The optical path switching device 3 is composed of a reflector and a mechanical adjustment mechanism, and its function is to allow the interference fringes formed after passing through the frequency discrimination unit 2 to be respectively projected into different subsequent units by moving the reflector into or out of the optical path.

[0050] The image sensor 4 is a CCD or CMOS device, and the size of its photosensitive surface should be larger than the coverage of the interference fringes obtained by the system.

[0051] The image acquisition and data processing unit 5 is calibrated to obtain the interferometer device function and wedge angle. The specific calibration process is as follows:

[0052] In the condition preparation stage, the diffuse reflection target needs to be placed in the center of the receiving system's field of view and illuminated by the light beam emitted by the monostable laser. After the diffuse reflection signal is received, it will pass through the echo beam expansion shaping unit 1, the frequency discrimination unit 2, the optical path switching device 3 and the image sensor 4 in sequence, and the linear interference fringes formed by the diffuse reflection echo will be collected by the image sensor 4. Adjust the angle direction of the front and rear reflection planes of the reflection cavity of the frequency discrimination unit 2 so that the direction of the interference fringes is parallel to the column direction of the image sensor 4 and the long side direction of the multi-channel detector 7. Since the echo signal comes from the laser directly diffusely reflected by the hard target plate, its spectral distribution is consistent with the emitted laser. Then, data processing is performed according to the following process.

[0053] S1. Average the brightness values ​​of pixels in each row of the image collected by the image sensor 4, and compress the two-dimensional interference fringes into a one-dimensional interference spectrum. Select an area near the brightness peak of the one-dimensional interference spectrum, and obtain the center position x0 of the fringes in pixels by finding the centroid position or fitting the Lorentz function.

[0054] (a) The method for finding the centroid is as follows:

[0055] The signal intensity vector of each column in the one-dimensional interference spectrum selection area is expressed as Record the pixel position of each column as

[0056] The centroid position is expressed as

[0057]

[0058] The centroid x p The dimension of is length, and the unit length 1 corresponds to the pixel size of the image sensor, that is, the side length d0 of each pixel.

[0059] (b) The Lorentz function is in the following form:

[0060]

[0061] where x p Represents the pixel row position corresponding to the peak intensity of the one-dimensional interference spectrum, Γ x is the half-height full width occupied by the function fitting on the interference spectrum, c0 represents the lift of the dark level and background signal, I x Indicates the strength of the signal.

[0062] Then, the center position of the stripe in pixels is x according to the pixel size d0 of the image sensor. p The center position of the stripe converted to length units is x0, x0 = d0x p .

[0063] S2. Continue the calibration process and change the frequency of the laser emission. Tune the frequency of the laser emission in steps of about 0.2 to 1.0 GHz. The center position of the stripes will shift accordingly. Repeat the acquisition of interference fringe images and execute step S1 to obtain the center position of the stripes. Finally, with the center position of the stripes (mm) as the independent variable and the offset of the laser output frequency (GHz) as the dependent variable, perform a function fitting to obtain the fitting slope value k. The absolute value of the fitting slope value |k| gives the corresponding relationship between the relative movement v of the laser frequency and the spatial position x of the stripes:

[0064] v=x*|k|

[0065] Due to the periodicity of interference fringes, when the frequency of the monochromatic light incident on the interference device is changed, the linear interference fringe maximum will repeat periodically. The light frequency difference between two adjacent fringe maximums is the free spectral range (FSR) of the interference device, denoted as v FSR, whose size can be calculated by measuring the distance between the two reflecting surfaces of the frequency discrimination unit 2.

[0066] When the free spectral range (FSR) of the interferometer is known, the relationship between the fitting slope |k| and the interferometer wedge angle α can be given by the following relationship:

[0067]

[0068] Where n is the refractive index of the medium between the reflection planes of the discrimination unit 2, and λ0 is the laser wavelength.

[0069] S3. Collect a two-dimensional interference image in the manner described in step S1, average each row to obtain a one-dimensional interference spectrum, and normalize its maximum value to 1.

[0070] Substitute the pixel size d0 of the image sensor and the slope value obtained by fitting in step S2, and adjust the horizontal coordinate of the one-dimensional interference spectrum to take the relative frequency as the independent variable. Shift the horizontal coordinate so that the relative frequency corresponding to the center position of the fringe is 0.

[0071] Then, the one-dimensional interference spectrum is fitted with a function, where the Lorentz line type is used to obtain the fitted spectral line width Γ x

[0072]

[0073] Since the spectral line width of the laser itself is much smaller than the transmission line width of the interferometer device function, the fitted line width can be approximately regarded as the line width value of the Ferzzo interferometer device function.

[0074] The line width Γ obtained by fitting the one-dimensional interference spectrum x With the unit of length, the interference spectrum line width Γ can be obtained by using the slope |k| obtained in step S2. x Mapped to the device function Γ in units of frequency v .

[0075] Γ v =Γ x *|k|

[0076] like Figure 2As shown, the link composed of the discrimination unit 2, the beam shaping unit 6 (optional), the optical path switching device 3, the image sensor 4, the image acquisition and data processing unit 5 is the calibration link of the spectrum identification system designed by the present invention. The link composed of the discrimination unit 2, the beam shaping unit 6 (optional), the multi-channel detector 7, and the acquisition and processing unit 8 is the existing Brillouin spectrum identification system. When the two work together, first adjust the optical path switching device 3 so that the echo enters the image sensor 4, and obtain the wedge angle and device function calibration results of the discrimination unit 2 according to the aforementioned condition preparation method and data processing flow S1-S3. After that, adjust the optical path switching device 3 so that the echo enters the multi-channel detector 7, and at this time, the Brillouin scattering signal of the water body can be collected on the photosensitive surface of the photodetector 7. The image sensor 4 introduced in the present invention does not work when the system performs the scattered signal acquisition function.

[0077] Example

[0078] In order to explain that the present invention can provide more accurate Fertz interferometer wedge angle and device function calibration information when the received signal has noise interference compared to the original spectrum identification system, the calibration results of the present invention and the original spectrum identification system are compared by designing a simulation experiment.

[0079] The simulation condition setting method, simulation process and simulation results are described as follows.

[0080] The echo signal 1 injected into the calibration system is approximately regarded as monochromatic light, with a wavelength λ0 of 532nm. The frequency discrimination unit 2 uses air as the filling medium between the reflective mirrors, and the refractive index is regarded as 1.0. The free spectrum range of the Fezzo interferometer is known to be 25GHz.

[0081] After the monochromatic light passes through the frequency discrimination unit 2, linear interference fringes are formed, and the illumination area can completely fill the Figure 2 The image sensor 4 and the multi-channel photodetector 7 are shown in FIG. The shapes of the two-dimensional interference fringes formed on the photosensitive surfaces of the image sensor 4 and the multi-channel photodetector 7 are as follows: Figure 3 As shown on the left. The outer dimensions of the photosensitive surface of the image sensor 4 and the multi-channel photodetector 7 are set to be 15.8 mm wide and 16 mm long. In this area, it is assumed that the light intensity incident on the frequency discrimination unit 2 is uniform. The two-dimensional interference fringes are averaged along the column direction, and the one-dimensional interference spectrum obtained is as follows Figure 3 Shown on the right.

[0082] Under noise-free conditions, the line shape of the one-dimensional interference spectrum is set to the Lorentz line shape. Since the simulation approximates the calibration echo as reflected light with negligible line width, its line width in the frequency domain is equivalent to the real line width of the device function of the interferometer Γ0 = 1.0 GHz, and the maximum amplitude of the interference spectrum is set to 1.

[0083]

[0084] The multi-channel photodetector structure in the simulation setting is as follows Figure 4 As shown, the unit is mm. The specifications of the photosensitive surface used in the simulation refer to the relevant structure of the common multi-anode photomultiplier tube on the market. The multi-channel photodetector 7 has 16 effective photosensitive surfaces. Each photosensitive area is 0.8 mm wide and 16 mm long. There is a dead zone with a width of 0.2 mm between adjacent photosensitive areas, which is considered to have no response to incident light.

[0085] The image sensor 4 pixel size in the simulation setting is 3.45um. To cover a width of 15.8mm, 4580 columns of pixels are required, assuming that all pixels have the same response characteristics to the incident signal.

[0086] The simulation signal generation process is as follows: Figure 5 shown.

[0087] ① Generation of one-dimensional interference spectrum: First, a Lorentz line shape with a real line width of 1.0 GHz is generated as the one-dimensional interference spectrum for input simulation. The center position of the spectrum line is set to a specific position. Since the relative frequency coordinates of the interference spectrum have a linear relationship with the spatial position coordinates of the interference pattern, the one-dimensional interference spectrum can have two sets of horizontal coordinates at the same time. The range of the relative frequency coordinates is set to -12.5 GHz to 12.5 GHz, corresponding to the complete free spectrum range of the demodulation unit 2. The range of the spatial coordinates is set to 0 to 15.8 mm, corresponding to the width of the photosensitive surface of the image sensor 4 and the multi-channel photodetector 7. The frequency-space mapping slope determined by the wedge angle of the interferometer is therefore set to |k| = 25 GHz / 15.8 mm.

[0088] The purpose of the simulation is to recover the following three calibration parameters from the noisy signal input by processing the simulated signal and evaluate the statistical difference between the recovered results and the true values:

[0089] The real line width of the interferometer device function is Γ0 = 1.0 GHz, the center position of the interference fringe is x0 = 7.9 mm, and the frequency-space mapping slope |k| = 25 GHz / 15.8 mm.

[0090] ②Calculation of theoretical transmittance

[0091] The theoretical transmittance of each channel of the multi-channel photodetector 7 and the image sensor 4 can be modeled as a piecewise integration of a specific interval on the one-dimensional interference spectrum.

[0092] The multi-channel photodetector 7 divides the horizontal axis of the one-dimensional interference spectrum 0-15.8 mm into 16 separate segmented integrals, and the image sensor 4 divides the horizontal axis of the one-dimensional interference spectrum 0-15.8 mm into 4580 closely adjacent segmented integrals, and the integral interval length of each pixel is 3.45 um corresponding to the pixel width.

[0093] ③Generate a noisy detector signal

[0094] Then, a multi-channel incident signal with the same length as the 7 channels of the multi-channel photodetector is generated. The signal is composed of two parts. One part is a DC signal with a constant amplitude of 1. The other part is a random number that obeys a Gaussian distribution, and the signal-to-noise ratio SNR is set to 20. The variance of the random noise is 1 / SNR, and the mean is the same as the variance.

[0095] ④ Inversion of interference spectrum parameters-multi-channel detector signal

[0096] Next, the generated noisy detector signal is used to estimate the center position and line width of the original one-dimensional interference spectrum, and the spectrum line parameters are inverted using a nonlinear least squares problem. The principle is to set the line width and center position of the interference spectrum as the quantity to be fitted, and calculate the 16-channel detector signal vector under noise-free conditions. The calculated result is compared with the noisy signal input Subtract, and use nonlinear optimization method to search for The minimum interference spectrum center position and line width are used as estimates of the true spectral line parameters of the one-dimensional interference spectrum.

[0097] ⑤ Inversion of interference spectrum parameters - image sensor signal

[0098] When processing the signal of the image sensor 4, since the number of pixels is large enough, a nonlinear curve fitting method is used to fit the input signal.

[0099] The fitted curve is set to a Lorentzian line type with an additional offset:

[0100]

[0101] Among them, p1 fits the amplitude of the spectrum line, p2 fits the center position of the spectrum line, p3 fits the half-height full width of the spectrum line, and p4 fits the offset introduced by the noise. p2 and p3 are the parameter values ​​required for inversion, with length units (mm), representing the spatial characteristics of the one-dimensional interference spectrum. Assuming that the frequency-space mapping slope |k| is known, the center position and half-height full width of the spectrum line can be mapped to the frequency domain.

[0102] ⑥Result analysis

[0103] The simulation described in steps ③-⑤ was repeated 100 times under the condition that the signal-to-noise ratio was set to 20, and 100 sets of interference spectrum line parameters obtained by inverting the signals of the multi-channel detector and the image sensor were generated respectively.

[0104] The true value of the frequency-space mapping slope |k|=25GHz / 15.8mm is taken as known information, and the true line width of the device function Γ0=1.0GHz is taken as unknown information for analysis.

[0105] The inversion results of the interference fringe center position and line width are as follows: Figure 6 and Figure 7 shown.

[0106] The mean center position deviation inverted using multi-channel detector signals is -6.2937MHz, and the standard deviation is 99.9709MHz.

[0107] The mean of the center position deviation inverted using the image sensor signal is -0.4769MHz and the standard deviation is 9.8423MHz.

[0108] Taking the standard deviation as the evaluation index of the stability of the device function line width calibration, under the condition of a signal-to-noise ratio of 20, the error is reduced by one order of magnitude when the image sensor introduced by the present invention is used compared to the calibration using a multi-channel detector.

[0109] When calibrating the frequency-space mapping slope |k|, it is necessary to change the emission frequency of the calibration monochromatic light so that the center position of the bright stripes on the photosensitive surface changes accordingly.

[0110] Under the condition that the frequency change of the calibrated monochromatic light is known, the center position of the interference fringe is inverted according to the method described in the above steps ④-⑤, and a linear least squares fit is performed with the frequency change of the monochromatic light as the ordinate and the inversion result of the fringe center position as the abscissa. The slope value obtained by fitting is the estimate of the frequency-space mapping slope |k|.

[0111] In order to calibrate the frequency-space mapping slope |k|, simulation is performed as follows. The signal-to-noise ratio is still set to 20, the true line width is still Γ0 = 1.0 GHz, the relative center frequency position of the one-dimensional interference spectrum is changed, and 5 sets of simulations under different conditions are performed, and the relative center frequency position is set to 0, 0.5 GHz, 1 GHz, 1.5 GHz and 2 GHz respectively. Under each set of conditions, the experiment is repeated 100 times according to steps ①-⑤, and a total of 500 sets of data are obtained.

[0112] At 5 different center frequencies, the test data were taken once respectively, and the estimated value of the frequency-space mapping slope |k| was calculated using the fitting results of the fringe center spatial position (in mm), which constituted 100 repeated independent estimates of the slope |k|. The results are shown in Figure 8 shown.

[0113] The mean deviation of the slope inverted using the multi-channel detector signal from the true value is -0.02476 (GHz / mm), and the standard deviation is 0.1016 (GHz / mm).

[0114] The mean value of the center position deviation inverted using the image sensor signal is -0.001300 (GHz / mm) and the standard deviation is 0.009512 (GHz / mm).

[0115] Taking the standard deviation as the evaluation index of the device function line width calibration stability, under the condition of a signal-to-noise ratio of 20, the calibration error of the frequency-space mapping slope |k| is also reduced by one order of magnitude when the image sensor introduced by the present invention is used compared to the calibration using a multi-channel detector.

[0116] After obtaining the frequency-space mapping slope |k|, the current wedge angle of the interferometer reflection plane can be calculated according to the following relationship.

[0117]

[0118] ⑦Simulation summary

[0119] The above simulations show that, under the set conditions, the method of using image sensors to calibrate the Fezzo interferometer device parameters introduced in the present invention reduces the calibration errors of the interferometer device function line width, the interference fringe center position and the frequency-space mapping slope |k| by about one order of magnitude compared to using the original multi-channel detector as the calibration data source, thus proving the effectiveness of the present invention.

[0120] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A high-precision reconstruction system of laser Brillouin scattering spectrum based on image, characterized by: It comprises an echo beam expansion shaping unit (1), a frequency discrimination unit (2), an optical path switching device (3), an image sensor (4), and an image acquisition and data processing unit (5), wherein: The echo beam expansion and shaping unit (1) expands and collimates the scattered echo and then injects it into the frequency discrimination unit (2); Frequency discrimination unit (2): performs multi-beam interference on the scattered echo after beam expansion and collimation, forms linear interference fringes and then emits them; An optical path switching device (3) is used to control the reflection of the linear interference fringes to the image sensor (4) and to sense light on the image sensor (4); Image acquisition and data processing unit (5): collects and stores the photosensitivity results of the image sensor (4), and obtains the device function and wedge angle corresponding to the interference fringes through interference spectrum extraction algorithm processing.

2. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 1, characterized in that: A beam shaping unit (6) is also included between the frequency discrimination unit (2) and the optical path switching device (3) for expanding or contracting the light beam passing through the frequency discrimination unit (2) to match the size of the image sensor (4).

3. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 1 or 2, characterized in that: After the echo beam expansion and shaping unit (1) expands and collimates the scattered echo, the divergence angle of the echo entering the frequency discrimination unit (2) is in the order of milliradian.

4. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 1 or 2, characterized in that: The frequency discrimination unit (2) is a Fezzo interferometer.

5. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 4, characterized in that: The direction of the linear interference fringes is parallel to the column direction of pixels of the image sensor (4).

6. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 5, characterized in that: The image acquisition and data processing unit (5) obtains the device function and wedge angle corresponding to the interference fringes through the interference spectrum extraction algorithm, which are specifically: (61) averaging the brightness values ​​of pixels in each row of the image collected by the image sensor (4) to compress the two-dimensional interference fringes into a one-dimensional interference spectrum; the image collected by the image sensor (4) is obtained by emitting a fixed-frequency laser beam from a laser to illuminate a diffuse reflection target, and the echo signal of the diffuse reflection target is sequentially obtained by passing through a frequency discrimination unit (2) and an optical path switching device (3); (62) near the brightness peak of the one-dimensional interference spectrum, obtain the center position x0 of the interference fringes in pixels; (63) changing the frequency of the laser emitting laser light, repeating steps (61) to (62), obtaining relationship data with the center position of the interference fringes as the independent variable and the frequency offset of the laser emitting laser light as the dependent variable, and obtaining the fitting slope value k by linear function fitting; (64) The wedge angle α of the frequency discrimination unit (2) is calculated using the absolute value of the fitting slope value |k|. Where n is the refractive index of the medium between the two reflection planes of the frequency discrimination unit (2), λ0 is the laser wavelength, and v FSR is the free spectral range of the frequency discrimination unit (2); (65) The one-dimensional interference spectrum is normalized, and then the horizontal axis of the one-dimensional interference spectrum is adjusted to take the relative frequency as the independent variable, and the relative frequency corresponding to the center position of the fringe is made 0; the one-dimensional interference spectrum is fitted with a function to obtain the fitted spectral line width Γ x , using the line width Γ x The device function Γ of the frequency discrimination unit (2) is calculated by using the absolute value of the fitting slope value |k| v =Γ x *|k|.

7. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 6, characterized in that: The method for obtaining the center position x0 of the interference fringes in pixels is as follows: the signal intensity vector of each column in the one-dimensional interference spectrum selection area is expressed as Record the pixel position of each column as Calculate the centroid position as The centroid x p The dimension is length, x0=d0x p , d0 is the pixel size of the image sensor (4).

8. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 6, characterized in that: The method of obtaining the center position x0 of the interference fringes in pixels is as follows: using the Lorentz function to fit the centroid x p , the Lorentz function is in the form where x p Represents the pixel row position corresponding to the peak intensity of the one-dimensional interference spectrum, Γ x is the half-height full width occupied by the function fitting on the interference spectrum, c0 represents the lift of the dark level and background signal, I x Indicates the strength of the signal; x0 = d0x p , d0 is the pixel size of the image sensor (4).

9. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 6, characterized in that: The frequency of the laser emitting laser is changed in steps of 0.2 to 1.0 GHz.

10. The high-precision reconstruction system of laser Brillouin scattering spectrum based on image according to claim 1 or 2, characterized in that: The image sensor (4) adopts a CCD or CMOS device, and the size of the photosensitive surface is larger than the coverage range of the linear interference fringes.

Images