Method for eliminating autocorrelation artifacts of polarization sensitive optical coherence tomography image

Abstract

The invention discloses a method for eliminating autocorrelation artifacts of a polarization sensitive optical coherence tomography image. Comprising the following steps: carrying out sample detection by using a PS-OCT (Polarization Sensitive Optical Coherence Tomography) system to obtain interference signals of two polarization channels; performing inverse Fourier transform on the interference signal of one channel to obtain an airspace signal of the interference signal; finding a self-correlation horizontal fringe artifact signal with the maximum intensity in the interference signals of the airspace, and setting the intensities of other positions to be zero; calculating an unwrapping phase of the self-correlation artifact in each A-scan, and taking a minimum value of the unwrapping phase as a reference phase; interpolating the spectral phases of the original interference signals of the two polarization channels according to a reference phase; and performing inverse Fourier transform and polarization parameter calculation to obtain an OCT intensity image and a polarization image with autocorrelation artifacts eliminated. According to the image phase information-based self-correlation artifact removal method, the self-correlation artifacts in the PS-OCT image can be effectively eliminated, additional hardware facilities do not need to be introduced, and the method is simple and effective.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of polarization sensitive optical coherence tomography, and particularly relates to a method for eliminating self-correlation artifacts of a polarization sensitive optical coherence tomography image. BACKGROUND

[0002] Polarization sensitive optical coherence tomography (PS-OCT) is a non-invasive imaging method that detects and quantifies the polarization properties of tissue by analyzing the changes in the polarization state of the backscattered light beam, obtaining information such as intensity, phase retardation and birefringence axis orientation of the sample. In the PS-OCT system based on polarization maintaining optical fiber, light is transmitted in different polarization modes (horizontal and vertical) in the polarization maintaining optical fiber, and the two produce a constant optical path difference and interfere to form horizontal stripe artifacts across the entire image, which are called self-correlation artifacts. Self-correlation artifacts can appear anywhere in the PS-OCT image, seriously affecting the quality of the PS-OCT image, especially for samples that need to be imaged at a large depth.

[0003] It is of great significance to remove the autocorrelation artifact for improving the quality of PS-OCT images. The commonly used method to eliminate the autocorrelation artifact is the background subtraction method. This method calculates the ensemble average of the spectrum in a single B-scan OCT image as the background signal, and then subtracts the background signal from all A-scans of the B-scan OCT image before performing the fast Fourier transform. However, the background subtraction method requires correction of the light source phase stability. In order to correct the phase of the light source, there are two methods at present. One method is to connect a fiber Bragg grating (FBG) in the system to ensure the consistency of the starting wavelength of each A-scan (W. Choi, et. al. "Phase-sensitive swept-source optical coherence tomography imaging of the human retina with a vertical cavity surface-emitting laser light source," Opt. Lett. 38, 338-340 (2013)). The other method is to add a Mach-Zehnder interferometer to the original system, and use an additional analog-to-digital conversion channel to record the interference signal, which can provide equidistant optical frequency space sampling and jitter amplitude to compensate for the phase error (S. Song, et. al. "Robust numerical phase stabilization for long-range swept-source optical coherence tomography," J. Biophotonics. 10, 1398-1410 (2017)). However, the FBG can only solve the phase alignment between A-scans, and cannot solve the problem of phase instability within a single A-scan; the second method requires an additional data acquisition channel, which increases the complexity of the system hardware and software processing.

[0004] In addition, Byers et al. proposed a method of filtering the autocorrelation artifact based on wavelet transform (R. Byers, et. al. "Attenuation of stripe artifacts in optical coherence tomography images through wavelet-FFT filtering," Biomed. Opt. Express, 10, 4179-4189 (2019)), which is a post-processing method based on intensity images. The information of the artifact position in the image cannot be completely recovered. SUMMARY

[0005] The present application aims to solve the problem of how to effectively eliminate the autocorrelation artifact existing in the PS-OCT image and the above-mentioned problems existing in the prior art, and provide an autocorrelation artifact elimination method applied to a PS-OCT system, which can effectively eliminate the autocorrelation artifact and restore the original sample information.

[0006] The technical scheme of the present application is as follows:

[0007] The present application discloses a method for eliminating autocorrelation artifact of a polarization-sensitive optical coherence tomography (PS-OCT) image, which utilizes the autocorrelation artifact information existing in the PS-OCT image, and is based on an interpolation method of a wave number domain spectrum phase to eliminate the autocorrelation artifact of the PS-OCT image, and comprises the following steps:

[0008] Step 1: using a PS-OCT system to detect a sample to obtain interference spectra of two polarization components perpendicular to each other, denoted as I H_i (k) and I V_i (k), wherein k represents a wave number, and subscripts H_i and V_i represent a horizontal polarization direction H and a vertical polarization direction V of an i-th A-scan in a B-scan OCT image, respectively;

[0009] Step 2: performing inverse Fourier transform on the interference spectrum with the horizontal polarization direction to obtain a complex amplitude of a spatial domain signal thereof, denoted as:

[0010]

[0011] wherein A H_i and represent an amplitude and a phase of the complex amplitude, respectively, and z represents a depth coordinate along a longitudinal direction;

[0012] performing inverse Fourier transform on the interference spectrum I H_i (k) to obtain an interference signal in a spatial domain, finding a strongest autocorrelation horizontal stripe artifact signal in the interference signal in the spatial domain, and setting signal intensities of other positions in the longitudinal direction to zero to obtain a spatial domain signal containing only the strongest autocorrelation horizontal stripe artifact wherein A a_i and represent an amplitude and a phase of the spatial domain signal containing only the strongest autocorrelation horizontal stripe artifact, respectively, and z a represents a position of the strongest autocorrelation horizontal stripe artifact;

[0013] Step 3: performing Fourier transform on the obtained spatial domain signal containing only the strongest autocorrelation horizontal stripe artifact to obtain a wave number domain signal of the autocorrelation horizontal stripe artifact wherein Ia_i and Let the intensity and phase of the wavenumber domain signal of the autocorrelation horizontal fringe artifact be represented, respectively; calculate the unwrapping phase of the autocorrelation artifact. This calculation is described as follows:

[0014]

[0015] Step 4: Unwrapping phase of all autocorrelation artifacts in the B-scan OCT image Find the minimum value, and get As a reference phase;

[0016] Step 5: Interferometric signal I in the wavenumber domain H_i (k) performs interpolation; specifically, the interference signal I is interpolated according to the wavenumber k. H_i (k) Unwrapping phase using autocorrelation artifacts This mapping is described as follows:

[0017]

[0018] Based on reference phase right Interpolation is performed, represented as:

[0019]

[0020] This yields the phase-corrected spectral signal;

[0021] Step 6: Use the reference phase obtained in step 4. Interference spectrum I for another polarization component V_i (k) Perform step 5 to obtain

[0022] Step 7: Analyze the corrected interference spectrum. and Inverse Fourier transform and polarization parameter calculations are performed to obtain the PS-OCT image after autocorrelation artifact removal.

[0023] Furthermore, for steps 2 to 4, the calculation of the reference phase can be based on the interference spectrum I with the polarization direction in the horizontal direction. H_i (k) can also be based on the interference spectrum I with the polarization direction in the vertical direction. V_i (k);

[0024] Furthermore, the interpolation method described in step 5 may include: linear interpolation, nearest neighbor interpolation, or cubic spline interpolation;

[0025] Furthermore, the applicable PS-OCT imaging system can be spectral domain OCT or swept-source OCT;

[0026] Further, the PS-OCT imaging system is applicable to PS-OCT systems with any central wavelength, as long as a polarization maintaining optical fiber, a polarization maintaining device or other devices causing the autocorrelation artifact are used in the construction of the PS-OCT system;

[0027] Further, the device causing the autocorrelation artifact is a lens, an attenuating sheet or a wave plate;

[0028] Further, the application is not only applicable to the PS-OCT imaging system, but also applicable to conventional OCT systems with autocorrelation artifacts. For the conventional OCT system, only one interference spectrum is obtained, and thus the autocorrelation artifact can be eliminated by performing steps 1-5 and then directly performing the inverse Fourier transform of step 7 to obtain the OCT image without the autocorrelation artifact.

[0029] The application has the advantages and beneficial effects that:

[0030] 1. The application eliminates the autocorrelation artifact in the PS-OCT system, improves the image quality of the PS-OCT system and the effective imaging range of the image;

[0031] 2. The application uses a software processing method to eliminate the autocorrelation artifact, which can reduce the complexity and development cost of the system compared with the hardware-based method;

[0032] 3. The application is based on spectral phase interpolation, which can effectively eliminate the autocorrelation artifact in the PS-OCT image and completely recover the information of the artifact position in the image compared with the existing software processing method;

[0033] 4. The application uses the phase interpolation method to eliminate the autocorrelation artifact, which does not need to perform complex iteration, has a fast calculation speed and can realize the artifact removal in real-time imaging. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 The flowchart of the PS-OCT image autocorrelation artifact elimination method provided in the embodiments of the application is shown in the figure;

[0035] Figure 2 The structural schematic diagram of the swept source PS-OCT imaging system provided in the embodiments of the application is shown in the figure;

[0036] In the figure, 101 is a broadband light source, 102 is a polarization controller, 103 is a circulator, 104 is a polarization beam splitter I, 105 is a polarization maintaining coupler, 106 is a collimator I, 107 is a quarter-wave plate I, 108 is a dispersion compensator, 109 is an aperture, 110 is a mirror, 111 is a collimator II, 112 is a quarter-wave plate II, 113 is a polarization beam splitter II, 114 is a balanced detector I, 115 is a balanced detector II, 116 is a signal processing device, and 117 is a sample detection module.

[0037] Figure 3 (a) is a PS-OCT intensity image without self-correlation artifact elimination, and (b) is a PS-OCT intensity image after using the self-correlation artifact elimination method.

[0038] Figure 4 (a) is a PS-OCT cumulative optical axis image without self-correlation artifact elimination, and (b) is a PS-OCT cumulative optical axis image after using the self-correlation artifact elimination method.

[0039] Figure 5 (a) is a PS-OCT cumulative phase delay image without self-correlation artifact elimination, and (b) is a PS-OCT cumulative phase delay image after using the self-correlation artifact elimination method. Specific embodiments

[0040] The technical solutions of the present application will be described clearly and completely in combination with the accompanying drawings in the embodiments of the present application.

[0041] The PS-OCT system described in the present application is suitable for any center wavelength PS-OCT system, which can be a spectral domain OCT or a swept source OCT. As long as the polarization maintaining fiber, the device with polarization maintaining function or other devices that cause self-correlation artifacts are used when building, such as lenses, attenuators and wave plates.

[0042] The embodiment provides a method for removing self-correlation artifacts of PS-OCT images, and proposes a solution for the self-correlation artifact problem existing in the PS-OCT system based on polarization maintaining fiber. The method of spectral phase interpolation is used to eliminate the self-correlation artifacts in the PS-OCT images.

[0043] The embodiment demonstrates the effect of the method on an ex vivo pig eye.

[0044] The flowchart of the method for removing self-correlation artifacts of PS-OCT images provided by the embodiment is shown in Figure 1 The specific steps are as follows:

[0045] Step 1, using a PS-OCT system to detect a sample, obtaining two mutually perpendicular polarization components of interference spectra, denoted as I H_i (k) and IV_i (k), where k represents wave number, and subscripts H_i and V_i represent horizontal polarization direction H and vertical polarization direction V of the i-th A-scan in the B-scan OCT image, respectively;

[0046] In this embodiment, a self-made PS-OCT system is used to detect an ex vivo pig eye to obtain interference signals of two mutually perpendicular polarization components of the ex vivo pig eye. The system is a PS-OCT system based on a polarization-maintaining optical fiber, as shown in Figure 2 , and the system comprises a wide-spectrum light source 101, a polarization controller 102, a circulator 103, a polarization beam splitter 104, a polarization-maintaining coupler 105, a collimator 106, a quarter-wave plate 107, a dispersion compensator 108, an aperture 109, a mirror 110, a collimator 111, a quarter-wave plate 112, a polarization beam splitter 113, a balanced detector 114, a balanced detector 115, a signal processing device 116, and a sample detection module 117. Specifically, the wide-spectrum light source 101 uses a swept light source with a wavelength band of 1300 nm.

[0047] The light output from the wide-spectrum light source 101 is output as linearly polarized light along the vertical direction after passing through the polarization controller 102. The linearly polarized light enters the circulator 103, and the light emitted from the circulator 103 is divided into a reference beam and a sample beam after passing through the polarization beam splitter 104 and the a end of the 50:50 polarization-maintaining coupler 105. The reference beam emitted from the c end passes through the collimator 106, the quarter-wave plate 107 with an angle of 22.5° between the fast-axis direction and the vertical direction, the dispersion compensator 108, and the aperture 109, and is reflected by the mirror 110 back to the c end of the polarization-maintaining coupler 105. The sample beam emitted from the d end passes through the collimator 111 and the quarter-wave plate 112 with an angle of 45° between the fast-axis direction and the vertical direction, so that the sample beam is converted from linear polarization to circular polarization. The sample beam is then detected by the sample detection module 117 to detect the ex vivo pig eye, and the light scattered from the ex vivo pig eye tissue returns to the d end of the polarization-maintaining coupler 105 along the original path. The reference light entering the c end of the polarization-maintaining coupler 105 and the sample light entering the d end interfere with each other, and the light after the interference is emitted from the polarization-maintaining coupler 105. The light emitted from the a end of the polarization-maintaining coupler 105 enters the polarization beam splitter 104, and the light along the vertical direction passes through the circulator 103 and enters the balanced detector 114, while the light along the horizontal direction enters the balanced detector 115. The light emitted from the b end of the polarization-maintaining coupler 105 passes through the polarization beam splitter 113, and the light along the vertical direction enters the balanced detector 114, while the light along the horizontal direction enters the balanced detector 115. The balanced detector 114 and the balanced detector 115 receive the light and convert it into an electrical signal, which is transmitted to the signal processing device 116.

[0048] The broadband light source 101 used in this embodiment is a swept light source (SL131160, Thorlabs) based on a micro-electromechanical vertical cavity surface emitting laser, the center wavelength of the light source is 1300 nm, the bandwidth is 101.8 nm, the number of sampling points of an A-scan is 4352, and the longitudinal imaging range is 18 mm (in air). The lateral resolution and longitudinal resolution of the system are 22 μm and 10 μm, respectively;

[0049] Step 2, the signal I with the polarization direction being the horizontal direction in the collected ex vivo pig eye interference signal H_i (k) performing inverse Fourier transform to obtain the horizontal component of the spatial domain complex amplitude signal, denoted as:

[0050]

[0051] wherein A H_i and respectively represent the amplitude and phase of the complex amplitude, and z represents the depth coordinate along the longitudinal direction; the autocorrelation horizontal stripe artifact signal with the maximum intensity is found in the spatial domain interference signal. For the system shown in this embodiment, the autocorrelation artifact with the maximum intensity is located at the longitudinal depth of 0.33 mm, and the signal intensity at other longitudinal positions is set to zero to obtain the spatial domain signal containing only the autocorrelation horizontal stripe artifact with the maximum intensity wherein A a_i and respectively represent the amplitude and phase of the autocorrelation horizontal stripe artifact with the maximum intensity, and z a represents the position of the autocorrelation horizontal stripe artifact with the maximum intensity; in this embodiment, z a = 0.33 mm;

[0052] Step 3, Fourier transform is performed on the obtained spatial domain signal containing only the autocorrelation horizontal stripe artifact with the maximum intensity to obtain the wave number domain signal of the autocorrelation horizontal stripe artifact wherein I a_i and respectively represent the intensity and phase of the wave number domain signal of the autocorrelation horizontal stripe artifact; the unwrapping phase of the autocorrelation artifact is calculated The calculation is described as:

[0053]

[0054] Step 4, the unwrapping phase of all autocorrelation artifacts in the B-scan OCT image is taken as the reference phase;

[0055] ​In this embodiment, the lateral scanning range of one B-scan is 12 mm, containing 500 A-scans, i.e. the minimum value of the unwrapped phase of the autocorrelation artifact of 500 A-scans is obtained as the reference phase

[0056] Step 5, interpolation is performed on the wave number domain interference signal I H_i (k) ; specifically, the interference signal I H_i (k) is interpolated according to the wave number k The mapping is described as:

[0057]

[0058] Based on the reference phase Interpolation is performed on , which is described as:

[0059]

[0060] Thus, the phase-corrected spectral signal is obtained; wherein the interpolation method can use linear interpolation, nearest neighbor interpolation or cubic spline interpolation;

[0061] In this embodiment, the interpolation method uses cubic spline interpolation;

[0062] Step 6, the reference phase obtained in step 4 is used to obtain the interference signal I (k) of the other polarization component V_i (k) is obtained by performing step 5

[0063] Step 7, inverse Fourier transform is performed on the corrected interference signal and , and the polarization parameter is calculated, to obtain the PS-OCT image after the autocorrelation artifact is eliminated.

[0064] For steps 2 to 4, the calculation of the reference phase can be based on the interference spectrum I H_i (k) with the polarization direction being the horizontal direction, or based on the interference spectrum I V_i (k) with the polarization direction being the vertical direction. This embodiment uses the interference spectrum I H_i (k) with the polarization direction being the horizontal direction.

[0065] The PS-OCT image obtained in this embodiment is calculated by the following formula respectively:

[0066]

[0067] Wherein, I is the intensity of PS-OCT, θ is the cumulative optical axis, δ is the cumulative phase delay, and AH and A V respectively represent the complex amplitudes of the spatial signals with polarization direction being horizontal and vertical direction respectively, is the phase difference of the two interference signals, the measurement range of θ is [-π / 2, π / 2], and is used to describe the cumulative polarization state change of the sample from the surface to the measurement depth.

[0068] As Figure 3 (a), Figure 4 (a), Figure 5 (a) are respectively the intensity image, the cumulative optical axis image and the cumulative phase retardation image of the pig eye without treatment, and it can be seen that the autocorrelation artifact horizontal lines appear in the PS-OCT intensity image and the polarization image, which run through the whole sample and seriously affect the image quality; Figure 3 (b), Figure 4 (b), Figure 5 (b) are respectively the intensity image, the cumulative optical axis image and the cumulative phase retardation image of the pig eye after the autocorrelation artifact elimination by using the method of the present application, and it can be seen that after the data is processed by using the method proposed in the present application, the autocorrelation artifact horizontal lines covering the structure of the pig eye in the PS-OCT intensity image and the polarization image are completely eliminated, and the sample information is not affected, and the image quality is significantly improved.

[0069] The above examples are only for illustrating the technical concept and characteristics of the present application, the purpose is to enable the personnel familiar with the art to understand the content of the present application and implement it, and cannot limit the protection scope of the present application, and any equivalent changes or modifications made according to the spirit and essence of the present application should be covered within the protection scope of the present application.

Claims

1. A method of eliminating autocorrelation artifacts from a polarization sensitive optical coherence tomographic image, the method comprising: The autocorrelation artifact in the polarization sensitive optical coherence tomography (PS-OCT) image is eliminated by using the autocorrelation artifact information existing in the PS-OCT image and an interpolation method based on the wave number domain spectrum phase; the specific steps include: Step 1, probing the sample with a PS-OCT system to obtain the interference spectra of two perpendicular polarization components, denoted as I H_i (k) and I V_i (k), where k represents the wave number, and the subscripts H_i and V_i represent the horizontal polarization direction H and the vertical polarization direction V of the i-th A-scan in the B-scan OCT image, respectively. Step 2, performing inverse Fourier transform on the interference spectrum with the polarization direction being the horizontal direction to obtain the complex amplitude of the spatial domain signal, denoted as: where A H_i and denote the amplitude and phase of the complex amplitude, respectively, and z denotes the depth coordinate along the longitudinal direction. Interferometric spectrum I H_i (k) after inverse Fourier transform, the strongest autocorrelation horizontal fringe artifact signal is found in the spatial domain interferometric signal, and the signal intensity of the longitudinal other positions is set to zero, to obtain the spatial domain signal containing only the strongest autocorrelation horizontal fringe artifact where A a_i and respectively represent the amplitude and phase of the strongest autocorrelation horizontal fringe artifact spatial domain signal, z a represents the position of the strongest autocorrelation horizontal fringe artifact; Step 3, Fourier transform of the obtained spatial domain signal containing only the intensity maximum of the autocorrelation horizontal stripe artifact Step 4, calculation of the intensity and phase of the autocorrelation horizontal stripe artifact in wave number domain where I a_i and represent the intensity and phase of the autocorrelation horizontal stripe artifact in wave number domain, respectively This calculation is described as: Step 4, unwrap the phase of all autocorrelation artifacts within the B-scan OCT image minimizing, to obtain as the reference phase; Step 5, Interference signal I in wave number domain H_i (k) performing an interpolation; in particular, an interpolation of the interference signal I according to the wave number k H_i (k) unwrapping the phase of the autocorrelation artifact The mapping is described as: Based on the reference phase To Interpolation is performed, expressed as: Thus, the spectrum signal after phase correction is obtained; Step 6, using the reference phase obtained in step 4 Interference spectrum I for the other polarization component V_i (k) performing step 5, obtaining Step 7, Corrected interferometric spectrum and Inverse Fourier transform and polarization parameter calculation are performed to obtain the PS-OCT image with autocorrelation artifact removed.

2. The method of claim 1, wherein the method is performed on a polarization sensitive optical coherence tomographic image. For steps 2 to 4, the calculation of the reference phase can be based on the interference spectrum I H_i (k) for which the direction of polarization is horizontal V_i (k).

3. The method of claim 1, wherein the method further comprises: determining a polarization state of the sample; and determining a polarization state of the reference sample. The interpolation method can be linear interpolation, nearest neighbor interpolation or cubic spline interpolation.

4. The method of claim 1, wherein the method is performed on a polarization sensitive optical coherence tomographic image. The applicable PS-OCT imaging system can be a spectral domain OCT or a swept source OCT.

5. The method of claim 1, wherein the method is performed on a polarization sensitive optical coherence tomographic image. The PS-OCT imaging system is applicable to any center wavelength PS-OCT system as long as a polarization maintaining optical fiber, a polarization maintaining device or other devices causing autocorrelation artifacts are used when the system is built.

6. The method of claim 5, wherein the method further comprises: determining a polarization state of the sample; and determining a polarization state of the reference sample. The device causing the autocorrelation artifact is a lens, an attenuating sheet or a wave plate.

7. The method of claim 1, wherein the method is performed on a polarization sensitive optical coherence tomographic image. The autocorrelation artifact elimination method is not only applicable to the PS-OCT imaging system, but also applicable to the conventional OCT system with autocorrelation artifacts.

8. The method of claim 7, wherein the method further comprises: determining a polarization state of the sample; and determining a polarization state of the reference sample. For the conventional OCT system, since there is only one interference spectrum, the autocorrelation artifact elimination only needs to perform steps 1-5 in claim 1, and then perform the inverse Fourier transform of step 7 to obtain the OCT image with the autocorrelation artifact eliminated.

Images