A spectrometer calibration method and system based on low-coherence interferometry

Abstract

The present invention discloses a spectrometer calibration method and system based on low-coherence interference. The present invention places two gauges of different thicknesses on the same plane, and respectively measures the interference signals between the gauge surface and the reference plane mirror, thereby obtaining the relative phase difference distribution. By coupling the light of the He-Ne laser into the interferometer, the characteristic wavelength and the corresponding pixel position can be obtained, and the absolute value of the phase difference can be obtained based on the relative phase difference distribution by adding the difference between the optical path differences of the two measurements; then, the exact value of the difference between the optical path differences of the two measurements can be calculated according to the wavelength and position of the characteristic spectrum line, and the wavelength corresponding to each pixel can be directly obtained from the absolute value of the phase difference, thereby completing the calibration of the spectrometer. The present invention can obtain the correspondence between all pixels and wavelengths of the camera at one time. At the same time, the high sensitivity of the phase of the interference signal avoids the fitting error of the traditional multi-characteristic spectrum line fitting method, so that the method can realize wide-spectrum, fast and high-precision calibration of the spectrometer.

Description

Technical Field

[0001] The present invention relates to optical coherence tomography (OCT) technology and spectrometer calibration technology, and in particular to a spectrometer calibration method and system based on low-coherence interference. Background Art

[0002] Spectrometers can be divided into dispersive spectrometers and Fourier spectrometers according to their working principles. Dispersive spectrometers use the dispersion phenomenon of light to disperse the incident light of the continuous spectrum to different positions in space according to different wavelengths and detect them. The detected data is the spectrum. The Fourier spectrometer performs spectral analysis in the frequency domain after Fourier transforming the measured time domain interference signal. The present invention focuses on and is only applicable to dispersive spectrometers, so the spectrometer in the present invention specifically refers to a dispersive spectrometer. When light interacts with the analyzed substance, each element will leave a characteristic emission or absorption line in the spectrum, and analyzing the spectrum can infer the composition of the substance. Therefore, spectrometers are widely used in particle physics, chemical analysis, astronomy and other fields.

[0003] Spectrometers are designed to linearly disperse light in the band of interest according to different wavelengths. However, the normal operation of spectrometers is limited by many factors. On the one hand, due to component equipment errors, focal mismatch, optical distortion and aberration, the incident light cannot be linearly dispersed, resulting in a nonlinear relationship between the camera pixels and the detected wavelength inside the spectrometer. Before performing spectral analysis, the nonlinear relationship between pixels and wavelengths must be obtained, and its accuracy directly affects the accuracy of spectral analysis.

[0004] On the other hand, during the actual use of the spectrometer, environmental factors such as temperature, humidity, pressure, vibration, etc. will affect the hardware performance, and the wavelength dispersion distribution of the spectrometer will change to varying degrees, directly affecting the spectral detection performance of the spectrometer. Therefore, the normal operation of the spectrometer under long-term and special environments requires recalibration steps to ensure.

[0005] Generally speaking, calibration lamps and tunable lasers can be used to calibrate the correspondence between pixels and wavelengths. The calibration lamp spectrum contains multiple narrow linewidth characteristic spectral lines, while the tunable laser can output a single spectral line with a narrow linewidth and can repeatedly change the wavelength of the output spectral line. After the spectrometer detects multiple characteristic spectral lines of the calibration lamp or tunable laser, the pixel-wavelength fitting curve can be obtained by polynomial fitting between the peak position of the spectrum and the wavelength of the characteristic spectral line. The corresponding wavelength of the pixel points outside the peak position is obtained by the fitting curve, so that the calibration of the wavelength relationship of all pixels of the spectrometer is realized. However, the tunable wavelength range of the tunable laser is small, the price is expensive, and the output wavelength also needs to be calibrated; although the calibration lamp does not require additional calibration, its light source output power is small, and the number of output spectral lines is insufficient when the spectrometer is calibrated with a wide spectrum. At the same time, when the calibration lamp and the tunable laser cannot cover the entire spectrum to be calibrated, the interpolation value will cause errors in the fitting curve of the pixel-wavelength relationship, thereby reducing the calibration accuracy.

[0006] In order to achieve wide-spectrum, fast, and high-precision spectrometer calibration, many foreign research institutions have proposed a variety of solutions. These methods are applied in optical coherence tomography (OCT) technology and are used for spectrometer calibration of spectral domain OCT systems.

[0007] The SABoppart research team at Illinois State University in the United States added a wavelength scanning filter module before the light source entered the interferometer. After the broadband light passed through the circulator and entered the module, only narrow-spectrum light was output, and the output light was tunable. An optical spectrum analyzer (OSA) was added to the detection end to calibrate the output wavelength. Although this method solves the problem of full calibration of the entire spectral range, the wavelength tuning module requires precise control to ensure the accuracy of the calibration, and the broadband light source needs to have a sufficiently high power output.

[0008] Ji-Hyun Kim and others from Korea University in South Korea obtained the wavelength distribution of all sampling points by detecting the zero-crossing point of the interference signal and fitting a polynomial, and then introduced a characteristic spectrum line to obtain the absolute wavelength. The relative wavelength distribution of this method depends on the measurement of the maximum imaging depth of the OCT system. It places a plane mirror on the sample arm as a sample and drives it with a motor. The maximum imaging depth of the system is obtained by observing the confusion caused by undersampling of the interference signal. However, this method of measuring the maximum imaging depth is limited by the errors caused by human judgment, and the accuracy of the spectrometer calibration is difficult to guarantee.

[0009] Maciej Wojtkowski's research group at the Nicolaus Copernicus University in Poland moved the sample arm reference plane mirror, added the Doppler frequency shift to the interference signal, and then measured the Doppler shift to obtain the wavelength of each sampling point. Assuming the moving speed is known, the Doppler frequency shift is related to the wavelength, so there is no need to introduce additional characteristic spectral line light sources, and the absolute wavelength corresponding to the sampling point can be calibrated. However, this method still requires measuring the maximum imaging depth of the system, and the ideal linear movement cannot be guaranteed when the motor drives the reference plane mirror to move, so the accuracy of the spectrometer calibration is also difficult to guarantee.

[0010] All of the above methods have inherent disadvantages, requiring the introduction of more complex devices, and the calibration accuracy of the spectrometer is limited by the large error of parameter measurement. Therefore, it is necessary to study a method for calibrating the absolute wavelength of a spectrometer that is easy to achieve wide spectrum, fast, and high precision. Summary of the invention

[0011] The purpose of the present invention is to provide a spectrometer calibration method and system based on low-coherence interference. The present invention is based on low-coherence interference technology and uses the interference spectrum collected by the spectrometer to calibrate the pixel-wavelength relationship of the spectrometer. The spectral domain optical coherence tomography system of the present invention measures the interference signals of two gauge surfaces of different thicknesses respectively, and the relative wavelength distribution can be obtained by using the phase difference of the two measurements. At the same time, a characteristic spectral line light source is introduced to obtain the calibration result of the absolute wavelength of the spectrometer.

[0012] The objective of the present invention is achieved through the following technical solutions:

[0013] A spectrometer calibration method based on low-coherence interference is to simultaneously couple the light emitted by a broadband light source and a He-Ne laser light source into an interferometer through a beam splitter; two gauges of different thicknesses are placed on the same plane and the interference signals between the gauge surfaces and the reference plane mirror are measured respectively. The interference signals are collected by the spectrometer and then synchronously collected by a data acquisition card. The interference signal is processed by inverse Fourier transform, filtering, Fourier transform, etc. in a computer to separate the AC term in the interference signal. The AC term in the interference signal can be obtained by Hilbert transform to obtain the wrapped phase of the interference signal. After unwrapping the wrapped phase of the interference signals of the two gauges, the relative phase distribution of all pixel points of the spectrometer can be obtained by subtracting them. The light of the He-Ne laser is coupled into the interferometer to obtain the characteristic wavelength and the corresponding pixel position. The absolute value of the phase difference can be obtained based on the relative phase difference distribution by adding the difference in the optical path difference of the two measurements. The difference in the optical path difference of the two measurements is theoretically twice the thickness difference of the two gauges. In practice, the rough value of the thickness difference of the two gauges is obtained by subtracting their nominal values. When the rough value error is small enough (accuracy is higher than 0.0002mm, i.e. national standard level 1 and above), it can be ensured that the error of the phase difference between two measurements of each pixel does not exceed 2π, then the 2π integer multiple information of the phase difference can be accurately restored. Finally, through the transformation relationship between phase and wavelength, the absolute wavelength calibration result of each pixel of the spectrometer can be obtained. The specific steps of this method are as follows:

[0014] (1) In the spectral domain optical coherence tomography system, the output light of the broadband light source and the output light of the He-Ne laser light source are simultaneously coupled into the interferometer through a beam splitter. Among them, the output light of the broadband light source is broadband light, which serves as the broadband light source for generating the interference signal; the output light of the He-Ne laser is quasi-monochromatic light, which serves as the characteristic spectrum line;

[0015] (2) A plane reflector is fixed on the reference arm of the spectral domain optical coherence tomography system as the reference surface of the interferometer, and two metal gauges are used as high-reflection samples as sample surfaces in the sample arm. The thicknesses of the two metal gauges are d1 and d2, respectively, and the thickness difference satisfies 0<|d2-d1|<5mm. The optical path differences of the interference signals on the surfaces of the two gauges are 2z1 and 2z2 respectively;

[0016] (3) The two interference spectrum signals are detected by a spectrometer and collected by a data acquisition card, and then transmitted to a computer memory for data processing;

[0017] (4) Since the OCT system has a depth-resolved tomographic effect, the spectral interference signal of the interferometer is subjected to inverse Fourier transform and window filtering to obtain the spatial spectrum of the interference signal with an optical path difference of 2z1 and 2z2. The AC term of the interference signal after removing the DC component can be obtained by Fourier transform.

[0018] (5) Performing Hilbert transform on the AC term of the interference signal can obtain the complex form of the AC term of the interference signal, and its phase can be directly obtained from the complex form. In this way, the wrapped phases of the interference signals on the surfaces of the two metal gauges are: and The difference in the package phase is

[0019] (6) The He-Ne laser light source emits quasi-monochromatic light, and its spectral line width is much smaller than the resolution of the spectrometer, so its light will be focused within a pixel, thus determining the wavelength corresponding to the pixel. The package phase difference of the pixel point plus the difference between the two measured optical path differences can obtain the 2π integer multiple information of the package phase difference, and then obtain the absolute phase difference. Among them, the difference between the two measured optical path differences is twice the difference in the nominal thickness of the two sets of gauges. The accuracy of the gauge nominal value is high enough, and the phase error caused by the accuracy error is less than 2π, so the 2π integer multiple information of the package phase difference can be accurately obtained;

[0020] (7) Since the two sets of gauges are measured separately, it is actually impossible to ensure that the gauge surfaces are placed in the same position. At the same time, the disturbance caused by environmental factors also causes the difference in the optical path difference between the two measurements to be not equal to twice the difference in the thickness of the two sets of metal gauges (considering the round-trip light reflection and the refractive index of air being 1). Therefore, directly substituting the nominal values ​​of the thickness of the two sets of gauges into the calculated wavelength may result in a large error in the calculated result. The exact value Δz of the difference in the optical path difference between the two measurements of the pixel obtained in (6) and the corresponding wavelength can be calculated. Based on the absolute phase difference and Δz of the two measurements, the wavelength corresponding to each pixel of the spectrometer can be calculated to complete the calibration of the spectrometer.

[0021] The present invention also provides a spectral domain optical coherence tomography system for realizing the self-calibration of the spectrometer of the above method, and the system includes a broadband light source, a He-Ne laser light source, an OCT interferometer, a spectrometer and a computer. The broadband light source and the He-Ne laser light source are coupled into the interferometer through a beam splitter via a reflector; the light coupled into the interferometer is expanded and shaped by a spatial filter, and then split by a beam splitter, and enters the reference arm and the sample arm respectively; the reference arm includes an adjustable attenuation plate, a focusing lens and a plane reflector, and the incident light is attenuated by the adjustable attenuation plate, and then focused by the focusing lens onto the surface of the plane reflector and reflected; the sample arm includes a two-dimensional galvanometer, an objective lens and two sets of gauge samples, and the incident light is controlled by the two-dimensional galvanometer to realize two-dimensional scanning, and is focused by the objective lens onto the surface of the gauge and reflected. Reflection; after the reflected light from the reference arm and the sample arm interferes, it enters the detection arm through a beam splitter and passes through another spatial filter to achieve beam expansion and stray light removal; the interference light after spatial filtering enters the spectrometer and is detected. The spectrometer includes a diffraction grating, a focusing lens and a linear array CCD camera; after the interference light passes through the diffraction grating, the collimated light of different wavelengths is dispersed in different directions and focused onto the linear array CCD camera through a focusing lens; the interference spectrum data detected by the linear array CCD camera is collected and processed by a computer, and finally the reconstructed image is displayed.

[0022] It should be pointed out that the scanning of the two-dimensional galvanometer and the triggering acquisition of the linear array CCD camera in the imaging system need to be performed synchronously. Therefore, the function generator card on the computer host generates three synchronous analog signals as the scanning drive signal of the two-dimensional scanning galvanometer and the external trigger signal of the linear array CCD camera.

[0023] Compared with the background technology, the present invention has the following beneficial effects:

[0024] 1. The present invention adopts a low-coherence interference detection method and introduces a coherence gate on the basis of a confocal gate. Therefore, the signal-to-noise ratio is high and the interference signal phase sensitivity is extremely high, so that the wavelength calibration accuracy of the spectrometer is very high; the high sensitivity of the interference signal phase avoids the fitting error of the traditional multi-feature spectral line fitting method, so that the method can achieve wide-spectrum, fast and high-precision calibration of the spectrometer;

[0025] 2. The phase difference of the interference signal on the two gauge surfaces is used to add a characteristic spectrum line to calculate the absolute phase. The error of the nominal value of the gauge thickness is small enough to recover the 2π integer multiple information lost in the 2π confusion problem, thus ensuring the accuracy of the calibration;

[0026] 3. No additional complex optical path is required, and the spectrometer can be calibrated by itself using the imaging function of the spectral domain optical coherence tomography system, which has high repeatability. The present invention can obtain the correspondence between all pixels and wavelengths of the camera at one time. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention is a schematic diagram of a spectral domain optical coherence tomography system for automatic calibration of absolute wavelength of a spectrometer.

[0028] Figure 2 It is a schematic diagram of data acquisition and hardware control of the spectral domain optical coherence tomography system of the present invention.

[0029] Figure 3 It is a schematic diagram of hardware control signals of the spectral domain optical coherence tomography system of the present invention.

[0030] Figure 4 It is a spectrometer calibration flow chart and interference spectrum data processing chart of the present invention.

[0031] In the figure: 1. broadband light source, 2. He-Ne laser light source, 3. beam splitter 1, 4. reflector, 5. focusing lens 1, 6. pinhole 1, 7. collimating lens 1, 8. beam splitter 2, 9. adjustable attenuation plate, 10. focusing lens 2, 11. plane reflector; 12. two-dimensional galvanometer, 13. objective lens, 14. gauge sample, 15. focusing lens 3, 16. pinhole 2, 17. collimating lens 2, 18. diffraction grating, 19. focusing lens 4, 20. linear array CCD camera, 21. computer, 22. data acquisition card, 23. function generator card. DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0033] like Figure 1As shown, the spectral domain optical coherence tomography system for spectrometer self-calibration of the present invention comprises a broadband light source, a He-Ne laser light source, an OCT interferometer, a spectrometer and a computer. The broadband light source (1) and the He-Ne laser light source (2) are coupled into the OCT interferometer through a beam splitter 1 (3) via a reflector (4). The wavelength range of the broadband light source (1) is 650-900nm. The light coupled into the interferometer is expanded and shaped by a spatial filter 1. The spatial filter 1 comprises a focusing lens 1 (5), a 50um pinhole 1 (6) and a collimating lens 1 (7). The light after collimation and expansion is then split by a beam splitter 2 (8) and enters the reference arm and the sample arm respectively. A plane reflector is fixed as the reference surface of the interferometer, and two metal gauges are used as the reflection surfaces of the sample arm. The thicknesses of the two metal gauges are d1 and d2 respectively, and the thickness difference satisfies 0<|d2-d1|<5mm, and the accuracy is higher than 0.0002mm. The reference arm comprises an adjustable attenuation plate (9), a focusing lens 2 (10) and a plane reflector (11). After the collimated light is attenuated by the attenuation plate, it is focused by the focusing lens 2 onto the plane mirror surface and reflected. The sample arm comprises a two-dimensional galvanometer (12), an objective lens (13) and two sets of gauge samples (14). The incident light is controlled by the two-dimensional galvanometer to achieve two-dimensional scanning, and is focused by the objective lens onto the gauge surface and reflected. The reflected light of the reference arm and the sample arm enters the detection arm after interference by the beam splitter, and is expanded and stray light is removed by another spatial filter 2. The spatial filter 2 comprises a focusing lens 3 (15), a 50um pinhole 2 (16) and a collimating lens 2 (17). The collimated light enters the spectrometer and is detected. The spectrometer comprises a diffraction grating (18), a focusing lens 4 (19) and a linear array CCD camera (20). After the collimated light passes through the diffraction grating, the collimated light of different wavelengths is dispersed in different directions and focused onto the CCD camera by the focusing lens. The interference spectrum data detected by the linear array CCD camera is collected by a data acquisition card (22), processed in a computer (21), and finally the reconstructed image is displayed. The solid line in the figure represents the data transmission path, and the dotted line represents the control signal transmission path.

[0034] Figure 2 The data acquisition and hardware control signal path diagram of the spectral domain optical coherence tomography system. The spectral data of the linear array CCD camera (20) in the spectrometer is transmitted to the data acquisition card (22) through the CameraLink cable and processed and displayed by the computer. It should be pointed out that the scanning of the two-dimensional galvanometer (12) and the triggering acquisition of the linear array CCD camera (20) need to be performed synchronously. Therefore, the function generator card (23) on the computer generates three synchronous analog signals as the scanning drive signal of the two-dimensional galvanometer (12) and the external trigger signal of the camera. The waveform of the three synchronous signals is as follows: Figure 3As shown in the figure, the fast axis drive signal of the galvanometer scanning is a sawtooth wave signal, the slow axis drive signal is a step signal, and the image acquisition card acquisition trigger signal is a square wave signal. The three-way synchronization signal realizes the frame scanning of the sample and the frame triggering of the camera data.

[0035] Figure 4 The figure shows the spectrometer calibration flow chart and data processing diagram based on low coherence interference according to the present invention, including characteristic spectrum line diagram and phase distribution diagram. Figure 4 The curve in is used to illustrate this.

[0036] The spectral domain optical coherence tomography interference spectrum signal detected by single axial scanning (A-scanning) can be expressed as:

[0037]

[0038] Where η is the sensitivity of the detector, q is the single charge, hν is the single photon energy, P r is the optical power returned from the reference arm to the detector, P o is the light power irradiated on the sample, z is half of the optical path difference between the reflection surface and the reference surface, r(z) and represent the amplitude and phase of the reflection coefficient in the depth direction of the sample, Γ(z) is the coherence function of the instantaneous output of the light source, k i represents the wave number at the i-th pixel. When z = 0, the optical path difference between the reference arm and the sample arm is 0. On the right side of the equation (1), the first term represents the light intensity returned from the reference arm, and the second term is the light intensity returned from each layer of the sample and the interference light intensity between each layer of the sample. The first two terms represent the noise signal that will interfere with the background of the final image. The first term is usually called the DC term, the second term is called the autocorrelation term, and the third term is the interference light intensity between the reference reflected light and each layer of the sample, that is, the effective interference spectrum signal detected by the spectral domain optical coherence tomography system.

[0039] The inverse Fourier transform of the interference spectrum signal in equation (1) can obtain the complex axial space signal, and then filter it to eliminate the mirror axial space signal, DC term, self-correlation term signal and separate the two cross-correlation terms. The effective interference spectrum signal can be obtained by Fourier transforming the two cross-correlation terms into the spectral space. Without loss of generality, only one reflection surface is considered, and the reference arm light intensity P r =P0, the spectral distribution function of the light source power spectrum is recorded as The coherence function Γ(z) of the instantaneous output of the light source is 1, then equation (1) can be simplified to:

[0040]

[0041] like Figure 4 (b) is the effective interference spectrum signal of the two sets of gauge surfaces. is the corresponding spectral domain phase discrete distribution. Applying Hilbert transform to equation (2) can obtain the wrapped phase distribution limited to the [-π, +π] principal value interval

[0042]

[0043] Here represents the absolute phase distribution in the spectral domain, floor represents the rounding operation towards negative infinity, and different discrete sampling wave numbers k i There are different orders of phase wrapping, and there is a 2π confusion or phase wrapping problem.

[0044] The wrapping phase is based on the starting sampling point Relative phase distribution can be obtained by continuous unwrapping, but the signal-to-noise ratio in the center area of ​​the general interference spectrum is high (small phase noise) and the signal-to-noise ratio in the edge area is low (large phase noise). The phase noise in the edge area will be accumulated and amplified during the continuous unwrapping process, thus affecting the accuracy of the recovered phase. Therefore, a bidirectional continuous unwrapping process with the phase principal value at the center position as the starting point is used to obtain the relative phase distribution.

[0045]

[0046] In the formula is a fixed value, indicating the central sampling position (corresponding to the central wave number k c ) is the number of phase wraps.

[0047] In order to eliminate the unbalanced dispersion between the two interference arms in equation (3), the phases of the two sets of gauge surface interference signals are extracted as follows: Figure 4 As shown in (c), the phases of the two positions are subtracted to obtain the corresponding phase difference distribution under the approximate optical path difference (2Δz = 2(d2-d1)) Figure 4 (d) Figure below):

[0048]

[0049] He-Ne light source provides characteristic spectral lines for spectrometer calibration, such as Figure 4 (a) shows the absolute phase 2k corresponding to this characteristic spectral line. HeNe Δz, the relative phase wrapping number N2-N1 can be calculated, expressed as:

[0050]

[0051] Here, the round operation means taking the nearest integer value. When the error phase generated by the characteristic spectral line does not exceed π, the accurate relative phase wrapping number can be obtained. The absolute phase distribution of the spectrum when the optical path difference is 2Δz is thus restored as ( Figure 4 (d) above):

[0052]

[0053] In this way, the exact value of the difference between the two measured optical path differences can be obtained by The absolute wavelength calibration value of the spectrometer is obtained as follows:

[0054]

[0055] Among them, i is the sampling point of the spectrometer and M is the maximum number of sampling points.

[0056] The present invention discloses a method for calibrating the absolute wavelength of a spectrometer based on low-coherence interference. The method uses the phase of the interference signal to match the characteristic wavelength to calibrate the pixel-wavelength relationship, and can obtain the corresponding relationship between all pixels and wavelengths of the camera at one time. The nominal values ​​of the two sets of gauge thicknesses are accurate enough to solve the 2π confusion problem of spectral phase jump calibration, and ensure the calibration accuracy and precision of spectral phase jump. At the same time, the high sensitivity of the interference phase avoids the fitting error of the traditional multi-feature spectral line fitting method, so that the method can achieve wide-spectrum, fast, and high-precision calibration of the spectrometer.

Claims

1. A spectrometer calibration method based on low-coherence interferometry, characterized in that: The specific steps are as follows: (1) The output light of the broadband light source and the output light of the He-Ne laser light source are simultaneously coupled into the interferometer through a beam splitter; wherein the output light of the broadband light source produces low-coherence interference; the output light of the He-Ne laser is quasi-monochromatic light, which serves as a characteristic spectrum line; (2) A plane reflector is fixed as the reference surface of the interferometer, and two metal gauges are used as the reflection surfaces of the sample arms. The thicknesses of the two metal gauges are d1 and d2, respectively, and the optical path differences of the interference signals on the surfaces of the two gauges are 2z1 and 2z2, respectively; (3) The interference signals are detected twice by the spectrometer and collected by the data acquisition card, and then transmitted to the computer memory for data processing; where the interference signal I(k i ) is: Where η is the sensitivity of the detector, q is the single charge, hν is the single photon energy, P r is the optical power returned from the reference arm to the detector, P o is the light power irradiated on the sample, z is half of the optical path difference between the reflection surface and the reference surface, r(z) and represent the amplitude and phase of the reflection coefficient in the depth direction of the sample, Γ(z) is the coherence function of the instantaneous output of the light source, k i represents the wave number at the i-th pixel. When z = 0, the optical path difference between the reflection surface of the sample arm and the reference surface is 0; (4) performing inverse Fourier transform and window filtering on the interference signal to obtain the spatial spectrum of the interference signal with the optical path difference of 2z1 and 2z2, and then performing Fourier transform to obtain the AC term of the interference signal with the DC component removed; the AC term I AC (k i ) is: Among them, S(k i ) is the spectral distribution function of the light source power spectrum; (5) Performing Hilbert transform on the AC term of the interference signal and taking the phase to obtain the wrapped phase of the AC term of the interference signal; (6) The wrapped phases of the interference signals on the two metal gauge surfaces are: and The calculation method of the wrapping phase is as follows: Among them, floor means rounding to negative infinity; The phase main value at the center position is used as the starting point for bidirectional continuous unwrapping processing to obtain the relative phase distribution In the formula is a fixed value, indicating the number of phase wrapping at the center sampling position, k c is the central wave number; The phases of the two measured interference signals are subtracted to obtain the corresponding phase difference distribution under the approximate optical path difference: Among them, the approximate optical path difference 2Δz=2(d2-d1); (7) Using the absolute phase 2k corresponding to the characteristic spectrum of the He-Ne light source HeNe Δz, calculate the relative phase wrapping number N2-N1, expressed as Among them, the round operation means taking the nearest integer value, k HeNe is the wave number corresponding to the characteristic spectrum line of He-Ne light source; Calculate the absolute phase distribution of the spectrum when the optical path difference is 2Δz: (8) The exact value of the difference in optical path difference between the two measurements is obtained by The absolute wavelength calibration value of the spectrometer is obtained as follows: Among them, i is the sampling point of the spectrometer and M is the maximum number of sampling points.

2. A spectral domain optical coherence tomography system for spectrometer self-calibration implementing the method of claim 1, characterized in that: It includes a broadband light source, a He-Ne laser light source, an interferometer, a spectrometer and a computer; the broadband light source and the He-Ne laser light source are coupled into the interferometer through a beam splitter and a reflector; the light coupled into the interferometer is expanded and shaped by a spatial filter, and then split by a beam splitter, and enters a reference arm and a sample arm respectively; the reference arm includes an adjustable attenuation plate, a focusing lens and a plane reflector, and the incident light is attenuated by the adjustable attenuation plate and then focused by the focusing lens onto the surface of the plane reflector and reflected; The sample arm includes a two-dimensional galvanometer, an objective lens and two sets of gauge samples. The direction of the incident light is controlled by the two-dimensional galvanometer to achieve two-dimensional scanning, and is focused by the objective lens onto the gauge surface and reflected; the reflected light from the reference arm and the sample arm interferes, enters the detection arm through a beam splitter, and passes through another spatial filter to achieve beam expansion and stray light removal; the interference light after spatial filtering enters the spectrometer and is detected. The spectrometer includes a diffraction grating, a focusing lens and a linear array CCD camera; after the interference light passes through the diffraction grating, the collimated light of different wavelengths is dispersed in different directions and focused onto the linear array CCD camera through a focusing lens; the interference spectrum data detected by the linear array CCD camera is collected and processed by a computer, and finally the reconstructed image is displayed.

Images