Optical coherence tomography device for optimizing aberration digital compensation and imaging method

Abstract

The invention discloses an optical coherence tomography device for optimizing aberration digital compensation and an imaging method, the device comprises an optical coherence tomography device and an image processing module, and the optical coherence tomography device comprises an illumination module, an interference module, an area array detection module and a synchronous control module. The imaging method comprises the following steps: setting device light source parameters according to coherent gating, and determining a sampling matching relation of the area array detection module according to a sampling theorem; selecting a frequency sweeping mode or a time domain mode, electrifying to activate a corresponding light source and a synchronous time sequence, and adjusting the position of a reference mirror; collecting a full-field interference sequence of the area-array camera, performing digital aberration compensation, performing compensation and denoising on volume data of different depths, and exporting a three-dimensional diffraction limit image; and outputting a three-dimensional tomographic image with transverse diffraction limit and axial micron-level resolution. According to the invention, the imaging resolution and the imaging speed of optical coherence tomography are improved under the conditions that the camera specification is reduced, the GPU dependence is reduced and the system cost is reduced.

Description

Technical Field

[0001] This invention relates to the field of biomedical optical imaging technology, and in particular to an optical coherence tomography apparatus and imaging method with optimized digital aberration compensation. Background Technology

[0002] Optical coherence tomography (OCT) is primarily used for three-dimensional, label-free imaging of transparent or semi-transparent tissues at the micrometer level, such as the corneal epithelium and the intima of veins. It can also be extended to intraoperative biopsy navigation and the detection of transparent materials such as optical films and fiber optic endfaces. With its advantages of high-speed, full-field acquisition and low cost, it can simultaneously provide three-dimensional structural and blood flow images in clinical, research, surgical, and industrial quality control scenarios.

[0003] In this technical field, existing solutions are mainly divided into three categories: point / line scanning OCT, time-domain full-field OCT, and frequency-sweep full-field OCT. Among them, point / line scanning OCT generally uses a high-speed galvanometer in conjunction with a line-scanning CCD / CMOS sensor to obtain lateral information by scanning point by point or line by line. To obtain imaging depth, a small numerical aperture microscope objective is used, with a lateral resolution of about 10 μm. This type of solution has relatively mature resolution and sensitivity, but its mechanical components are complex and costly, and it is prone to motion artifacts in live imaging.

[0004] To meet the demands of non-scanning and high-resolution 3D imaging, the Langevin Institute in France invented temporal full-field OCT. This technology utilizes short-coherence white light and high-speed phase-shifting interference, capturing the entire interference field in a single frame on a camera, followed by multi-frame phase-shifting demodulation. Peng Xiao et al. reviewed the device configuration and performance bottlenecks of temporal full-field OCT, pointing out that deep scattering and aberrations under large numerical apertures lead to a sharp decrease in resolution and contrast, necessitating numerical denoising or iterative phase correction to clearly visualize corneal details. The limitations are mainly threefold: 1. The axial coherence gate width is fixed, making it difficult to balance depth of focus and imaging depth; 2. A high dynamic range camera (≥70dB) is required to maintain the signal-to-noise ratio; 3. Aberration correction is mostly performed after acquisition, resulting in time-consuming algorithms and limited compensation range.

[0005] Non-scanning swept-frequency full-field OCT has developed rapidly in recent years. On the one hand, it is based on the hardware framework of time-domain full-field OCT; on the other hand, it uses swept-frequency light source beam expansion, combined with an area array camera to capture the interference field in one go, enabling high-speed volume imaging. Similarly, addressing the trade-off between depth of focus and resolution, to suppress aberrations, the University of Lübeck in Germany designed hardware correction relying on off-axis interferometry. This involves introducing the reference light at an off-axis angle to increase the spatial envelope, and then using an area array camera and swept-frequency light source for parallel acquisition. In recent years, with the emergence of computational refocusing methods, the mainstream approach uses iterative correction with digital adaptive algorithms. By iteratively optimizing the wavefront through phase reconstruction, it can recover details of the living cornea.

[0006] Currently, the Intalight commercial prototype DREAM OCT is... TM Employing full-field scanning OCT and integrating a multimodal blood flow algorithm, it can obtain a 130° field of view in one go, stitched together into a wide field of view of about 20 mm. It has obtained CE certification from the European Union.

[0007] However, the full-field sweep OCT of Dubois 2004 and Hillmann 2017 did not provide an analytical expression for the lower sampling limit, and currently commercial OCT relies on high dynamic range cameras. The following are the technical problems that urgently need to be solved:

[0008] (1) Depth of focus – imaging depth contradiction. In a wide field of view full-field scheme, without real-time aberration compensation, the accumulation of wavefront error causes the axial resolution and contrast to drop sharply with depth.

[0009] (2) Strong dependence on high-specification cameras. Existing non-scanning full-field OCT requires a dynamic range of ≥70dB and a frame rate of ≥800fps to maintain the signal-to-noise ratio, which leads to a significant increase in hardware costs and data bandwidth. Cameras with a frame rate higher than 500fps that meet the needs of biological tissue detection cost millions of RMB.

[0010] (3) Aberration compensation is only a post-hoc algorithm. Iterative phase correction methods are prone to getting trapped in local extrema when the initial aberrations are large or the signal-to-noise ratio is low, and the computational cost is high, making them unsuitable for real-time applications.

[0011] (4) The overall system cost and integration complexity remain high. Commercial prototype DreamOCT TM Although the galvanometer has been eliminated, it requires a large-aperture scanning head, a high-end camera, and a dedicated GPU cluster, which is difficult for conventional clinical departments to afford.

[0012] In summary, existing solutions cannot achieve real-time full-field, μm-level resolution and high-contrast images under conditions of low dynamic range and low frame rate area array cameras through design-level coherence gate-wavefront coupling optimization. The system cost is also difficult to reduce to a level acceptable for routine clinical use. Therefore, it is urgent to study an OCT system that calculates the back-inference optical path and conjugate lens parameters using coherence gates and realizes aberration optimization digital compensation in a hardware-algorithm collaborative framework to solve the core challenges of depth of focus and imaging depth, and to balance cost reduction and real-time performance. Summary of the Invention

[0013] The purpose of this invention is to provide an optical coherence tomography system and imaging method that features high imaging resolution, low hardware cost, simple algorithm, high computational efficiency, small size, low power consumption, and ease of use.

[0014] The technical solution to achieve the purpose of this invention is: an optical coherence tomography device for optimizing digital aberration compensation, comprising an optical coherence tomography device and an image processing module, wherein the optical coherence tomography device includes an illumination module, an interference module, an area array detection module and a synchronization control module;

[0015] The illumination module is used to generate swept light or broadband white light source to provide uniform illumination for the array detection module that meets the Nyquist sampling requirements.

[0016] The interference module uses the Linnik interference optical path to generate interference fringes, which are then received by the area array detection module.

[0017] The area array detection module is used to perform lateral diffraction-limited sampling and send the data to the image processing module in real time;

[0018] The synchronization control module outputs a frequency sweep time base and camera trigger in frequency sweep mode, and outputs a phase shift signal and camera trigger in time domain mode.

[0019] The image processing module is used to perform single phase compensation on the acquired interference sequence, reconstruct the three-dimensional tomography, and output a three-dimensional image.

[0020] An optical coherence tomography method with optimized digital aberration compensation, the method being based on the aforementioned optical coherence tomography apparatus with optimized digital aberration compensation, the method comprising the following steps:

[0021] Step 1: During the device manufacturing stage, set the device light source parameters according to the coherent gating setting, and determine the sampling matching relationship of the array detection module according to the sampling theorem;

[0022] Step 2: Select the sweep frequency mode or time domain mode, and power on to activate the corresponding light source and synchronization timing;

[0023] Step 3: Adjust the position of the reference mirror so that the optical path difference between the sample arm and the reference arm is less than the coherence length;

[0024] Step 4: Acquire the full-field interference sequence of the area array camera;

[0025] Step 5: Perform digital aberration compensation on the full field-of-view interference sequence, compensate and denoise the volume data at different depths, and derive the three-dimensional diffraction-limited image.

[0026] Step 6: Output a three-dimensional tomographic image with transverse diffraction limit and axial micron-level resolution.

[0027] Compared with the prior art, the significant advantages of this invention are: (1) By precisely setting the light source bandwidth Δλ through an adjustable filter and coordinating with the numerical aperture NA of the microscope objective, the time gate and spatial gate are determined to dominate in the design stage, achieving micron-level axial resolution and improving the imaging resolution; (2) Taking the diffraction limit resolution of the objective's numerical aperture, i.e., the lateral resolution, as the target, the magnification of the objective and the tube lens is designed according to the object size and the object-image relationship of the camera pixels, thus solving the contradiction between focal depth and imaging depth; Based on the Nyquist sampling theorem, the acquired lateral spectrum falls completely into the sampling bandwidth, without the need to increase the bit depth or frame rate to compensate for undersampling noise; In the time domain full-field OCT mode (Mode-T), an economical 8-bit, 340fps mode is used. The camera can achieve the same signal-to-noise ratio as high-end cameras such as 16-bit and 800fps, reducing hardware costs; (3) After the magnification of the objective lens and the tube lens is matched, the size of the diffraction spot is determined. Based on the forward propagation model, only one phase compensation is needed to correct the value to the diffraction limit, eliminating multiple iterations of phase solving, reducing the amount of computation by two orders of magnitude, and reducing the algorithm delay from ≥200ms to <20ms, providing space for real-time imaging; (4) The total material list of the whole machine is reduced by about 75%, the number of calibration points is reduced from 18 to 7, and the cabinet volume and power consumption are halved. While maintaining micron-level resolution and high contrast imaging performance, the system cost and integration complexity are reduced, and the usage cost of conventional clinical and scientific research laboratories is reduced. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the structure of an optical coherence tomography apparatus for optimizing digital aberration compensation according to the present invention.

[0029] Figure 2 This is a schematic diagram of the Mode-S lighting module in this invention.

[0030] Figure 3 This is a schematic diagram of the Mode-T lighting module in this invention.

[0031] Figure 4 This is a block diagram of the Mode-S synchronous control system in this invention.

[0032] Figure 5 This is a block diagram of the Mode-T synchronous control system in this invention.

[0033] Figure 6 This is a flowchart of the Mode-S signal triggering process in this invention.

[0034] Figure 7 This is a flowchart of the Mode-T signal triggering process in this invention.

[0035] Figure 8This is a flowchart of the Mode-S signal demodulation process in this invention.

[0036] Figure 9 This is a flowchart of the Mode-T signal demodulation process in this invention.

[0037] Figure 10 This is a response curve of axial strength in an embodiment of the present invention.

[0038] Figure 11 This is a response curve diagram of the lateral edge in an embodiment of the present invention.

[0039] The following are the labeling elements in the diagram: 1. Optical coherence tomography apparatus; 2. Image processing module; 3. Illumination module; 4. Interference module; 5. Area array detector module; 6. Synchronization control module; 7. Light source; 8. First lens; 9. Filter; 10. Second lens; 11. Beam splitter; 12. First microscope objective; 13. Second microscope objective; 14. Sample; 15. Reference mirror; 16. Sample axial displacement stage; 17. Piezoelectric ceramic phase shifter; 18. Tube mirror; 19. Area array detector. Detailed Implementation

[0040] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0041] Combination Figure 1 The present invention provides an optical coherence tomography apparatus for optimizing digital aberration compensation, comprising an optical coherence tomography apparatus 1 and an image processing module 2, wherein the optical coherence tomography apparatus 1 comprises an illumination module 3, an interference module 4, an area array detection module 5, and a synchronization control module 6.

[0042] The illumination module 3 is used to generate swept light or broadband white light source to provide uniform illumination for the array detection module 5 that meets the Nyquist sampling requirements.

[0043] The interference module 4 uses the Linnik interference optical path to generate interference fringes, which are received by the area array detection module 5;

[0044] The array detection module 5 is used to perform lateral diffraction-limited sampling and send the data to the image processing module 2 in real time;

[0045] The synchronization control module 6 outputs a frequency sweep time base and camera trigger in frequency sweep mode, and outputs a phase shift signal and camera trigger in time domain mode.

[0046] The image processing module 2 is used to perform single phase compensation on the acquired interference sequence, reconstruct the three-dimensional tomography, and output a three-dimensional image.

[0047] The center wavelength λ0 and bandwidth Δλ are set according to the target imaging depth and resolution requirements, and the parameters of the optical system satisfy the Nyquist sampling and coherence gate optimization principles.

[0048] As a specific example, the lighting module 3 includes a light source 7, a first lens 8, a filter 9, and a second lens 10;

[0049] The light source 7 is equipped with two modes: a frequency sweep mode (Mode-S) and a time-domain mode (Mode-T). Mode-S generates frequency sweep light and can be used with a tunable frequency sweep laser, such as a MEMS-VCSEL. The output is a fiber optic end face connected to the first lens 8. Figure 2 As shown; Mode-T generates a broadband white light source, such as LED, SLED, halogen tungsten lamp, etc., which is directly incident on the first lens 8, such as... Figure 3 As shown; the internal drive circuit maintains linear output according to the preset wavenumber table or phase shift table, and receives the trigger signal from the synchronization control module 6;

[0050] The first lens 8 collimates the divergent light emitted by the light source 7 to form a uniform parallel beam, reducing coupling losses of subsequent components; the filter 9 precisely defines the spectral bandwidth Δλ; by replacing or fine-tuning the filter 9, the coherence gate length can be set during the assembly stage to match the axial resolution of the system with the target; the second lens 10 reshapes or moderately expands the beam after it has been shaped by the filter 9, so that the numerical aperture of the outgoing beam matches the input optical path of the interference module 4, and the power is uniformly distributed to the detection surface, providing uniform illumination that meets the Nyquist sampling requirements for the array detection module 5, and achieving high signal-to-noise ratio sampling.

[0051] As a specific example, the lighting module 3 uses only one type of light source during operation, and the center wavelength λ0 and bandwidth Δλ remain constant after being designed according to the coherence gate width during the manufacturing stage.

[0052] As a specific example, the filter 9 is an adjustable bandwidth filter with an adjustment range of 10–300 nm.

[0053] As a specific example, the interference module 4 includes a beam splitter 11, a first microscope objective 12, a second microscope objective 13, a sample 14, a reference mirror 15, a sample axial displacement stage 16, and a piezoelectric ceramic phase shifter 17.

[0054] The beam splitter 11 is a 45° dielectric beam-splitting cube, which splits the collimated beam from the illumination module 3 into the sample arm and the reference arm at a 1:1 ratio, and recombines the beams in the reverse path to form interference fringes. The first microscope objective 12 is mounted on the sample arm and focuses the incident light onto the sample 14. Its NA and magnification are designed and matched according to the sampling principle to ensure transverse diffraction-limited sampling. The second microscope objective 13 is mounted on the reference arm and has the same specifications as the first microscope objective 12 to maintain wavefront symmetry between the two arms, and forms conjugate imaging with the reference mirror 15. The sample 14 is a transparent or translucent tissue or material to be examined and is fixed on the sample axial displacement stage 16. The reference mirror 15 is a plane mirror, which is flush with the sample 14. The surface conjugation provides a reference wave for interference; the sample axial displacement stage 16 adopts a nanometer-positioning direct-drive platform, which can accurately move the sample 14 along the optical axis to achieve three-dimensional tomographic scanning, with step resolution better than axial resolution; the piezoelectric ceramic phase shifter 17 is fixedly mounted with the reference mirror 15 in the same frame, providing subwavelength-level phase stepping without changing the maximum optical path difference. A conventional stepping method or a carrier frequency method can be selected. The stepping method refers to applying discrete phase modulation sequentially at four phase points (0, π / 2, π, 3π / 2), completing one four-step phase shift cycle in no more than 10ms; the carrier frequency method refers to applying sinusoidal modulation to the reference mirror; finally, an interference sequence is obtained through the stepping or carrier frequency method for time-domain mode demodulation.

[0055] The interference module 4 uses the Linnik interference optical path to generate interference fringes in both Mode-T and Mode-S. The beam emitted by the illumination module 3 is incident on the beam splitter 11 and converges onto the first microscope objective 12 and the second microscope objective 13, respectively, uniformly illuminating the sample 14 and the reference mirror 15. The backscattered light returning from the sample 14 and the reference mirror 15 converges at the beam splitter 11 and forms interference fringes after satisfying the coherence condition, which are received by the area array detection module 5. Only in Mode-T is the piezoelectric ceramic phase shifter 17 used to fine-tune the phase between the interference arms to form interference sequences with different phase shifts, preparing for demodulation. The sample axial displacement stage 16 is used to move the sample 14 axially to measure the three-dimensional structure.

[0056] The interferometer module 4 employs aberration symmetry and compensability. Dual-objective symmetry and sampling matching cancel out the initial wavefront errors of the two arms, allowing for single-stage phase template correction to the diffraction limit. Separation of single-step phase shifting and axial scanning decouples phase modulation from volume scanning. The piezoelectric ceramic phase shifter 17 only handles interferometric demodulation, while the sample axial displacement stage 16 is responsible for volume sampling, reducing motion crosstalk. Through this structure, the interferometer module 4 ensures extremely low residual aberrations while creating stable and repeatable interferometric input conditions for subsequent single-stage digital compensation.

[0057] As a specific example, the piezoelectric ceramic phase shifter 17 is driven by a sinusoidal voltage in the phase range of 0 to 2π, with a driving voltage frequency of 0.5kHz–20kHz and a displacement amplitude of 50nm–300nm, to achieve continuous carrier frequency phase modulation; the piezoelectric ceramic phase shifter 17 is a conventional piezoelectric ceramic actuator with a phase noise of no more than 0.02πrms.

[0058] As a specific example, the area array detection module 5 includes a tube mirror 18 and an area array camera 19;

[0059] The tube lens 18 and the first microscope objective 12 form an infinity optical path, which conjugates the sample 14 planar image onto the target surface of the area array camera 19 according to the magnification. The focal length, objective numerical aperture NA, and pixels of the area array camera 19 are designed according to the sampling matching principle to meet the Nyquist sampling condition and achieve lateral diffraction-limited sampling. The image captured by the area array camera 19 is output by the interferometry module 4, and the interferometric sequence frame stream data after imaging by the tube lens 18 is sent to the image processing module 2 in real time. Under Mode-T conditions, the pixel size p of the area array camera 19, the beam waist radius w0 of the incident light at the back focal plane of the objective, the objective numerical aperture NA, and the objective focal length f are all specified. MO , focal length f of the tube lens tl ,satisfy The sampling matching relationship can maintain a signal-to-noise ratio comparable to that of high-end cameras under low dynamic range of 48-60dB and medium frame rate of around 340fps.

[0060] The scattered light from sample 14 is combined with the reference wave at beam splitter 11 after passing through the first microscope objective 12, forming interference fringes under coherent conditions. The fringes are then imaged onto the target surface of area array camera 19 through tube lens 18. The plane of sample 14 and the image plane of area array camera 19 form a strict conjugate relationship. Due to the sampling matching relationship between the objective lens, the focal length of tube lens 18 and the pixels, the diffraction-limited transverse spectrum falls completely within the Nyquist bandwidth of area array camera 19, without the need to increase the bit depth or frame rate to compensate for undersampling noise. Area array camera 19 receives the exposure trigger pulse from synchronization control module 6 and acquires two-dimensional interference fringes at each phase step or each axial position. The data is transmitted to image processing module 2 via a high-speed interface to complete phase-shift demodulation and three-dimensional reconstruction. Through the integrated design and sampling matching of tube lens 18 and area array camera 19, area array detection module 5 achieves transverse diffraction-limited sampling without increasing hardware costs, providing high signal-to-noise interference data for backend one-time aberration compensation.

[0061] As a specific example, the focal length of the tube lens 18 is 150–300 mm.

[0062] As a specific example, the area array camera 19 is a mid-depth area array camera with a single-frame exposure time of no more than 2ms, a dynamic range of no less than 56dB, and a frame rate of no less than 300fps.

[0063] As a specific example, the synchronization control module 6 includes an FPGA and a multi-channel driver board;

[0064] The synchronization control module 6 adopts a "two-choice" mutually exclusive timing design, selecting between the frequency sweep mode (Mode-S) and the time domain mode (Mode-T). During operation, the modes cannot be activated simultaneously or switched online, ensuring that a single timing source synchronizes the entire system. A schematic diagram of the Mode-S synchronization control system is shown below. Figure 4 The schematic diagram of the Mode-T synchronous control system is shown below. Figure 5 ;

[0065] like Figure 6 As shown, in the sweep frequency mode (Mode-S), the synchronization control module 6 sends a trigger signal, then sends a sweep frequency pulse to the light source 7, sends a square wave signal to the sample axial displacement stage 16, and sends a TTL signal to trigger the area array camera 19 to acquire data; the sweep frequency clock triggers the light source 7 to complete N light source spectral scans; at the beginning of each λ sweep frequency cycle, the TTL triggers the area array camera 19 to perform a single-frame exposure; volume data can be acquired in a single scan, and volume sampling of "N scans and N frames" can be achieved simultaneously; afterwards, the axial displacement stage 16 can be driven.

[0066] like Figure 7 As shown, in the time-domain Mode-T, the synchronous control module 6 sends a trigger signal, transmits a square wave signal to the sample axial displacement stage 16, sends an analog voltage signal to the piezoelectric ceramic phase shifter 17, and sends a TTL signal to trigger the area array camera 19 to acquire data, generating segmented or continuous voltage waveforms. The piezoelectric ceramic can be subjected to 0, π / 2, π, 3π / 2, or sinusoidal modulation phase shifts using either a conventional stepping method or a carrier frequency method. After each predetermined phase sequence is completed, a TTL signal is sent to trigger the area array camera 19 to acquire the interference sequence at the corresponding depth. Subsequently, a square wave is generated to drive the axial displacement stage 16 to obtain the four-step or N-step interference sequence of the next depth layer, achieving volume sampling.

[0067] Through a "two-choice" mutually exclusive timing design, the synchronization control module 6 achieves hardware sharing while maintaining functional independence between the frequency sweep type FF-SS-OCT and the time domain type FF-TD-OCT, ensuring timing accuracy and avoiding redundancy in electronic resources and cables.

[0068] As a specific example, the image processing module 2 includes a high-speed acquisition card, a GPU / CPU processing unit, a display terminal, solid-state storage, a chassis power supply, and heat dissipation components;

[0069] The high-speed acquisition card is used to write the data generated by the area array camera 19 into the video memory using direct memory access; the GPU / CPU processing unit is used for interferometric demodulation, aberration compensation and reconstruction in various modes, performs single phase compensation on the acquired interferometric sequence, reconstructs three-dimensional tomographic data, and obtains a three-dimensional image with a horizontal diameter of 1.3μm, an axial diameter of 1.05μm and a processing delay of less than 20ms; the display terminal presents the processing results in real time and writes them to solid-state storage simultaneously.

[0070] The chassis power supply and heat dissipation components are used to provide an installation environment, power system and heat dissipation system for high-speed acquisition cards, GPU / CPU processing units, display terminals and solid-state storage.

[0071] The image processing module includes, in sequence, an interferometric demodulation unit, an aberration compensation unit, an image reconstruction unit, and an image display unit, making the technical solution more modular and feasible.

[0072] This invention also provides an optical coherence tomography method for optimizing digital aberration compensation, comprising the following steps:

[0073] Step 1: During the device manufacturing stage, set the parameters of the device light source 7 according to the coherent gating setting, and determine the sampling matching relationship of the area array detection module 5 according to the sampling theorem;

[0074] Step 2: Select the sweep frequency mode or time domain mode, and power on to activate the corresponding light source 7 and synchronization timing;

[0075] Step 3: Adjust the position of the reference mirror 15 so that the optical path difference between the sample arm and the reference arm is less than the coherence length;

[0076] Step 4: Acquire the full-field interferometric sequence of the area array camera 19;

[0077] Step 5: Perform digital aberration compensation on the full field-of-view interference sequence, compensate and denoise the volume data at different depths, and derive the three-dimensional diffraction-limited image.

[0078] Step 6: Output a three-dimensional tomographic image with transverse diffraction limit and axial micron-level resolution.

[0079] As a specific example, in step 1, during the device manufacturing stage, the parameters of the device light source 7 are set according to the coherent gating system, and the sampling matching relationship of the array detection module 5 is determined according to the sampling theorem, as follows:

[0080] Step 1.1, Setting the coherence gate width:

[0081] Select the center wavelength λ0 of the light source, the bandwidth Δλ of the light source, and the numerical aperture NA, and calculate the coherence gate width L. c The formula is:

[0082] L c=min{l tc ,l sc ,l Lsc} (1)

[0083] Among them l tc l sc l Lsc These represent the widths of the time coherence gate, the lateral spatial coherence gate, and the axial spatial coherence gate, respectively.

[0084] Given the center wavelength λ0 and bandwidth Δλ of the light source, and the numerical aperture NA, calculate and derive the coherence gate width L. c Coherence gates can be divided into time coherence gates:

[0085]

[0086] Lateral spatial coherence gate:

[0087]

[0088] Longitudinal spatial coherence gate:

[0089]

[0090] In the formula, θ satisfies the geometric relationship n sinθ = NA. Among the three coherence gates mentioned above, the smallest is the OCT total coherence gate, therefore L is obtained. c The value,

[0091] L c =min{l tc ,l sc ,l Lsc} (1-4)

[0092] By adjusting the light source bandwidth Δλ, i.e., the sweep frequency integration bandwidth or the filter bandwidth, L c Meets the target axial resolution.

[0093] Based on this condition, the light source parameters are determined, and light source 7 is selected. Mode-S uses a swept frequency light source and a swept frequency full-field OCT as the main device; Mode-T uses a white light source and a filter and a time-domain full-field OCT as the main device. The two share the OCT main device and can be quickly switched by replacing light source 7 and area scan camera 19.

[0094] Step 1.2, Sampling and Matching:

[0095] To ensure that all out-of-focus areas are corrected to the diffraction limit, it is necessary to obtain signals with spatial frequency components close to the diffraction limit while also ensuring that the phase difference between adjacent pixels is within a specific range to avoid aliasing. According to the Nyquist sampling theorem, the phase change between adjacent pixels should be less than π. Therefore, the full-field OCT sampling interval Δx should satisfy:

[0096]

[0097] Where w0 is the beam waist radius of the Gaussian beam focal plane, n is the sample refractive index, and M is the magnification between the sample and the pixel size p;

[0098] Calculate the focal length f of the tube lens based on the object-image relationship. tl The formula is:

[0099]

[0100] Based on this condition, a tube lens with a focal length of 18 f was selected. tl This ensures that the transverse spectrum falls completely within the sampling bandwidth, thereby obtaining the optical design of the optical coherence tomography device that satisfies the optimal digital aberration compensation.

[0101] Step 1.3: Complete the center wavelength λ0, bandwidth Δλ, objective back focal plane waist radius w0, objective numerical aperture NA, and objective focal length f. MO , tubular endoscope tl Matching and shaping the optical path based on pixel size p.

[0102] As a specific example, step 5 involves performing digital aberration compensation on the full-field interference sequence, compensating and denoising volume data at different depths, and deriving a three-dimensional diffraction-limited image. The frequency sweep mode is as follows: Mode-S... Figure 8 As shown, the time-domain mode Mode-T is as follows: Figure 9 As shown, the details are as follows:

[0103] Step 5.1, Solving for the frontal complex signal, as follows:

[0104] (1) In the sweep frequency mode Mode-S, the three-dimensional interference spectrum signal acquired by the area array camera 19 is represented as I(λ; x, y). The interference sequence of each pixel is resampled according to the wavelength interval of the sweep frequency light source to obtain the wavenumber linearized spectral interference signal I(k; x, y), where k represents the spatial frequency. For each pixel (x, y) in the area array camera 19, the Fourier transform is solved to obtain the initial three-dimensional volume data. The frontal complex signal V(x, y; z) of this data at the sample depth z is extracted.

[0105] (2) In the time-domain Mode-T mode, the area array camera 19 acquires a two-dimensional interferometric sequence I. N The light intensity expression for (x, y) is:

[0106]

[0107] Analytical demodulation is performed on the four-step phase-shifting or carrier-frequency phase-shifting interferogram; the full-field OCT amplitude signal A(x,y) of the four-step phase-shifting interferogram is:

[0108]

[0109] The full-field OCT phase signal φ(x,y) is:

[0110]

[0111] After phase unpacking, the generated two-dimensional complex signal V(x,y) is:

[0112] V(x,y)=A(x,y)exp(-iφ(x,y)) (7)

[0113] Step 5.2, 2D FFT, is as follows:

[0114] (1) In the sweep frequency mode (Mode-S), the front complex signal V(x,y;z) is extracted, and the spatial frequency domain spectrum of the signal is solved. The spectral domain signal V(f) can be obtained by two-dimensional Fourier transform. x ,f y ;z), where (f x ,f y () represents the spectral coordinate system;

[0115] (2) In the time-domain mode (Mode-T), the spatial frequency domain spectrum of the complex signal V(x,y) can be obtained by solving the two-dimensional Fourier transform, which yields the spectral domain signal V(f). x ,f y ), where (f x ,f y () represents the spectral coordinate system;

[0116] Step 5.3, digital aberration correction, is as follows:

[0117] (1) In the swept-frequency mode (Mode-S), a forward diffraction model is adopted, and a digital refocusing algorithm or a multi-focus complex averaging algorithm to suppress multiple scattering light is used to analyze the complex interference spectrum V(f). x ,f y ;z) Perform phase retrieval; In the digital refocusing algorithm, the deconvolution operator for phase retrieval is H -1 (f x ,f y ;z d This can be expressed as:

[0118]

[0119] Where z d This indicates the amount of defocus that needs to be estimated;

[0120] The three-dimensional reflectivity signal S(x,y,z) after refocusing can be expressed as:

[0121]

[0122] (2) In the time-domain Mode-T, angular spectrum propagation or aperture filtering methods are used to apply V(f) to the frequency domain. x ,f y Phase retrieval is performed, restoring the lateral diffraction-limited resolution in the depth direction. The deconvolution operator for phase retrieval is H. -1 (f x ,f y ;z d After performing deconvolution using this operator, the phase-recovered two-dimensional structure signal T(x,y) is obtained by performing a two-dimensional inverse Fourier transform.

[0123]

[0124] Step 5.4, depth signal acquisition, as detailed below:

[0125] (1) In the sweep frequency mode Mode-S, the sweep frequency full-field OCT principle does not require the use of an axial displacement stage. The depth signal can be obtained after performing a Fourier transform on the interference spectrum. At this time, the diffraction-limited resolution three-dimensional tomographic structure S(x,y,z) can be directly output. The sample axial displacement stage 16 can be selected to obtain other depths, or blood flow or phase function imaging can be selected as signal supplements.

[0126] (2) In the time-domain mode Mode-T, the sample axial displacement stage 16 is used to stack the corresponding two-dimensional tomographic images in depth order to form a three-dimensional tomographic data signal T(x,y,z), and blood flow or phase function maps can also be superimposed.

[0127] As a specific example, in the frequency sweep mode Mode-S described in step 5.1, wavenumber linearization is based on k-clock triggering for resampling, and the resampling error is no greater than the width of one sampling point. Those skilled in the art can use existing mature hardware modules to implement the above functions, and the specific implementation method is not limited.

[0128] As a specific example, in the time-domain mode Mode-T described in step 5.1, the predetermined phase modulation of the carrier frequency method is a sinusoidal continuous carrier frequency modulation of 0.5–20kHz, and the carrier frequency component is extracted in the digital domain by lock-in amplification.

[0129] As a specific example, in the sweep frequency mode Mode-S described in step 5.3, when the depth exceeds 0.5mm, a multi-focus complex averaging algorithm is used to suppress multiple scattering noise.

[0130] The optical coherence tomography device with optimized digital aberration compensation proposed in this invention can be used for three-dimensional microstructure imaging and quantitative assessment in the following clinical or research applications:

[0131] (1) Based on high-order aberration correction and diffraction-limited resolution design, it is applied to non-invasive imaging of corneal nerve fiber layer to assist in the diagnosis of dry eye and postoperative recovery monitoring.

[0132] (2) It can be applied to non-blocking imaging of superficial and deep retinal vascular networks, and combined with micro-blood flow analysis to assist in the assessment of diseases such as diabetic retinopathy.

[0133] (3) In interventional treatment, transparent tissue structure imaging and material interface judgment are performed on the balloon dilation or vascular stent implantation area to assist in preoperative selection and intraoperative positioning.

[0134] Example

[0135] In one embodiment, parameters such as λ0 = 660nm, Δλ = 225nm, NA = 0.25, n = 1.38, and camera pixel size p = 5.5μm were selected for performance testing.

[0136] The coherent gate L is obtained based on the coherent gate calculation formula. c Length is:

[0137] L c =min{l tc ,l sc ,l Lsc}=0.85μm (11)

[0138] Therefore, the axial resolution δ is estimated. z =0.85μm.

[0139] According to Rayleigh's criterion, the theoretical lateral resolution formula for the diffraction limit is:

[0140]

[0141] Therefore, the lateral resolution δ is estimated. x =1.61μm.

[0142] A Gaussian beam with a beam waist radius of w0 = 0.54 μm was selected as the incident light. The sampling matching relationship was determined based on the sampling theorem. The upper limit of the sampling theorem is the amount of data stored, which is not a requirement for optical design; the lower limit of the sampling theorem is that the phase interval of each pixel in the camera is less than π, satisfying the formula... The minimum focal length of the tube lens is 170mm, and the aforementioned f tl It should fall within the range of 100–300 mm; f can be selected. tl =250mm commercial infinity-corrected telescope.

[0143] An apparatus is constructed according to the present invention, and the apparatus is used to perform actual resolution verification.

[0144] Four-step phase-shifting TD-FF-OCT was used to inspect the USAF standard resolution board, such as... Figure 10 As shown, depth-direction data was acquired on a standard resolution plate, and tomographic images were obtained using a four-step shifting method. The actual axial resolution of the system was calculated based on the intensity distribution values ​​of the tomographic images at different depths. The intensity response curve of the resolution plate along the axial depth direction was fitted, and the abscissas at the full width at half maximum (FWHM) were approximately 5.95 μm and 7.00 μm for the peak relative intensity of 162 and the relative intensity at half width at half maximum (FWHM), respectively, differing by 1.05 μm. Therefore, the axial resolution of the system is approximately 1.05 μm.

[0145] The actual lateral resolution of the device was calculated by performing a four-step phase-shifting method on the same sample and based on the line pairs on the resolution target tomography. Figure 11 The intensity response curve is fitted to the yellow line of the resolution plate tomography. The width of the edge response, which falls from 80% to 20%, is calculated. This span corresponds to about 1.3μm for two pixels. Therefore, the actual horizontal resolution is about 1.3μm.

[0146] In summary, due to the coherent gating design and high-contrast fringes of this invention, the actual axial resolution of the device is approximately 1.05 μm, which is very close to the theoretical axial resolution. The actual lateral resolution is approximately 1.3 μm, exceeding the theoretical resolution by 20% in detail due to the high-contrast fringes and aberration compensation of this invention.

[0147] This embodiment demonstrates that, under coherent gating and sampling matching conditions, the device can obtain diffraction-limited three-dimensional tomographic images through a single digital compensation without the need for a high dynamic range or high-speed camera, and the overall procurement cost is reduced by approximately 75% compared to commercial full-field OCT.

[0148] The above specific embodiments are used to explain and illustrate the present invention, and not to limit the present invention. Any modifications and changes made to the present invention within the spirit and scope of the claims fall within the protection scope of the present invention.

Claims

1. An optical coherence tomography apparatus for optimizing digital aberration compensation, characterized in that, It includes an optical coherence tomography device (1) and an image processing module (2), wherein the optical coherence tomography device (1) includes an illumination module (3), an interference module (4), an area array detection module (5), and a synchronization control module (6); The illumination module (3) is used to generate swept light or broadband white light source to provide uniform illumination for the array detection module (5) that meets the Nyquist sampling requirements; The interference module (4) uses the Linnik interference optical path to generate interference fringes, which are received by the area array detection module (5); The array detection module (5) is used to perform transverse diffraction-limited sampling and send the data to the image processing module (2) in real time; The synchronization control module (6) outputs a frequency sweep time base and camera trigger in frequency sweep mode, and outputs a phase shift signal and camera trigger in time domain mode. The image processing module (2) is used to perform single phase compensation on the acquired interference sequence, reconstruct the three-dimensional tomography, and output a three-dimensional image.

2. The optical coherence tomography apparatus for optimized digital aberration compensation according to claim 1, characterized in that, The lighting module (3) includes a light source (7), a first lens (8), a filter (9), and a second lens (10); The light source (7) is equipped with a sweep frequency mode (Mode-S) and a time domain mode (Mode-T). Mode-S generates sweep frequency light, tunes the sweep frequency laser, and outputs an optical fiber end face that is connected to the first lens (8). Mode-T generates a broadband white light source that is directly incident on the first lens (8). The internal driving circuit maintains linear output according to a preset wavenumber table or phase shift table and receives the trigger signal from the synchronization control module (6). The first lens (8) collimates the diverging light emitted by the light source (7) to form a uniform parallel beam, reducing the coupling loss of subsequent components; the filter (9) limits the spectral bandwidth Δλ; by replacing or fine-tuning the filter (9), the coherence gate length is set during the assembly stage to match the axial resolution of the system with the target; the second lens (10) reshapes or expands the beam after it has been shaped by the filter (9) to match the numerical aperture of the outgoing beam with the input optical path of the interference module (4) and distributes the power uniformly to the detection surface, providing uniform illumination that meets the Nyquist sampling requirements for the array detection module (5) and achieving high signal-to-noise ratio sampling; The lighting module (3) only uses one type of light source during operation, and the center wavelength λ0 and bandwidth Δλ remain constant after being designed according to the coherence gate width during the manufacturing stage; The filter (9) is a filter with adjustable bandwidth.

3. The optical coherence tomography apparatus for optimized digital aberration compensation according to claim 1, characterized in that, The interference module (4) includes a beam splitter (11), a first microscope objective (12), a second microscope objective (13), a sample (14), a reference mirror (15), a sample axial displacement stage (16), and a piezoelectric ceramic phase shifter (17). The beam splitter (11) is a medium beam splitter cube, which splits the collimated beam from the illumination module (3) into the sample arm and the reference arm at a 1:1 ratio, and recombines the beams in the reverse path to form interference fringes; the first microscope objective (12) is mounted on the sample arm and focuses the incident light onto the sample (14). Its NA and magnification are designed and matched according to the sampling principle to ensure transverse diffraction-limited sampling; the second microscope objective (13) is mounted on the reference arm and has the same specifications as the first microscope objective (12) to maintain the wavefront symmetry of the two arms and form conjugate imaging with the reference mirror (15); the sample (14) is a transparent or translucent material to be tested and is fixed on the sample axial displacement stage (16); the reference mirror (15) is a plane mirror and is used in conjunction with the sample. The sample (14) is planar conjugate, providing a reference wave for interference; the sample axial displacement stage (16) adopts a nano-positioning direct drive platform, moving the sample (14) along the optical axis to achieve three-dimensional tomographic scanning, with step resolution better than axial resolution; the piezoelectric ceramic phase shifter (17) is fixed in the same frame as the reference mirror (15), providing subwavelength phase stepping without changing the maximum optical path difference; either the stepping method or the carrier frequency method is selected: the stepping method refers to applying discrete phase modulation sequentially at four phase points of 0, π / 2, π, 3π / 2, with a time of no more than 10ms to complete one four-step phase shift cycle; the carrier frequency method refers to applying sinusoidal modulation to the reference mirror; finally, the interference sequence is obtained through the stepping method or the carrier frequency method, which is used for time-domain mode demodulation; The interference module (4) uses the Linnik interference optical path to generate interference fringes in both Mode-T and Mode-S. The beam emitted by the illumination module (3) is incident on the beam splitter (11) and converges on the first microscope objective (12) and the second microscope objective (13) respectively, uniformly illuminating the sample (14) and the reference mirror (15). The backscattered light returning from the sample (14) and the reference mirror (15) converges at the beam splitter (11) and forms interference fringes after satisfying the coherence condition, which are received by the area array detection module (5). Only in Mode-T, the piezoelectric ceramic phase shifter (17) is used to fine-tune the phase between the interference arms to form interference sequences with different phase shifts, in preparation for demodulation. The sample axial displacement stage (16) is used to move the sample (14) axially to measure the three-dimensional structure. The piezoelectric ceramic phase shifter (17) is driven by a sinusoidal voltage in the phase range of 0 to 2π to achieve continuous carrier frequency phase modulation; the piezoelectric ceramic phase shifter (17) is a piezoelectric ceramic actuator.

4. The optical coherence tomography apparatus for optimized digital aberration compensation according to claim 1, characterized in that, The area array detection module (5) includes a tube mirror (18) and an area array camera (19); The tube lens (18) and the first microscope objective (12) form an infinity optical path. The sample (14) is conjugately imaged onto the target surface of the area array camera (19) according to the magnification. The focal length, objective numerical aperture NA, and the pixels of the area array camera (19) are designed according to the sampling matching principle to meet the Nyquist sampling condition and achieve transverse diffraction-limited sampling. The image captured by the area array camera (19) is output by the interference module (4). The interference sequence frame stream data after imaging by the tube lens (18) is sent to the image processing module (2) in real time. Under Mode-T conditions, the pixel size p of the area array camera (19), the waist radius w0 of the incident light on the back focal plane of the objective, the objective numerical aperture NA, and the objective focal length f are determined. MO , focal length f of the tube lens tl ,satisfy The sampling matching relationship; The scattered light from the sample (14) is combined with the reference wave by the beam splitter (11) after passing through the first microscope objective (12), forming interference fringes under coherent conditions, and is imaged onto the target surface of the area array camera (19) through the tube lens (18). The plane of the sample (14) and the image plane of the area array camera (19) form a conjugate relationship. By matching the sampling relationship between the focal length and pixels of the objective lens and the tube lens (18), the diffraction-limited transverse spectrum falls completely within the Nyquist bandwidth of the area array camera (19), without the need to increase the bit depth or frame rate to compensate for undersampling noise. The area array camera (19) receives the exposure trigger pulse sent by the synchronization control module (6) and collects two-dimensional interference fringes at each phase step or each axial position. The data is transmitted to the image processing module (2) via a high-speed interface to complete phase-shift demodulation and three-dimensional reconstruction. Through the integrated design and sampling matching of the tube lens (18) and the area array camera (19), the area array detection module (5) realizes transverse diffraction-limited sampling, providing interference data for back-end one-time aberration compensation.

5. The optical coherence tomography apparatus for optimized digital aberration compensation according to claim 1, characterized in that, The synchronization control module (6) includes an FPGA and a multi-channel driver board; The synchronization control module (6) adopts a two-choice mutually exclusive timing design, selecting one between the sweep frequency mode Mode-S and the time domain mode Mode-T. During operation, the modes cannot be activated simultaneously or switched online, ensuring that a single timing source synchronizes the entire system. In the sweep frequency mode Mode-S, the synchronous control module (6) sends a trigger signal, then sends a sweep frequency pulse to the light source (7), sends a square wave signal to the sample axial displacement stage (16), and sends a TTL signal to trigger the area array camera (19) to acquire data; the sweep frequency clock triggers the light source (7) to complete N light source spectrum scans; at the rising or falling edge of each λ sweep frequency cycle, the TTL triggers the area array camera (19) to perform single-frame exposure; a single scan acquires volume data, and at the same time, N scans and N frames of volume sampling are achieved; Then, a drive axial displacement stage (16) was selected; In the time-domain mode (Mode-T), the synchronous control module (6) sends a trigger signal, sends a square wave signal to the sample axial displacement stage (16), sends an analog voltage signal to the piezoelectric ceramic phase shifter (17), and sends a TTL signal to trigger the area array camera (19) to collect data, generating segmented or continuous voltage waveforms. The piezoelectric ceramic is subjected to 0, π / 2, π, 3π / 2 or sinusoidal modulation phase shift using the stepping method or carrier frequency method. After each predetermined phase sequence is completed, the TTL signal is sent to trigger the area array camera (19) to collect the interference sequence at the corresponding depth. Subsequently, a square wave is generated to drive the axial displacement stage (16) to obtain the four-step or N-step interference sequence of the next depth layer, realizing volume sampling. Through a two-choice mutually exclusive timing design, the synchronous control module (6) achieves hardware sharing while maintaining functional independence between the frequency sweeping FF-SS-OCT and the time-domain FF-TD-OCT.

6. The optical coherence tomography apparatus for optimized digital aberration compensation according to claim 1, characterized in that, The image processing module (2) includes a high-speed acquisition card, a GPU / CPU processing unit, a display terminal, solid-state storage, a chassis power supply and heat dissipation components; The high-speed acquisition card is used to write the data generated by the area array camera (19) into the video memory in a direct memory access manner; The GPU / CPU processing unit is used for interferometric demodulation, aberration compensation and reconstruction in each mode. It performs single phase compensation on the acquired interferometric sequence, reconstructs three-dimensional tomographic data, and obtains three-dimensional images. The display terminal presents the processing results in real time and writes them to solid-state storage simultaneously. The chassis power supply and heat dissipation components are used to provide an installation environment, power system and heat dissipation system for high-speed acquisition cards, GPU / CPU processing units, display terminals and solid-state storage.

7. An optical coherence tomography method for optimizing digital aberration compensation, characterized in that, This method is based on the optical coherence tomography apparatus with optimized digital aberration compensation as described in any one of claims 1 to 6, and the method includes the following steps: Step 1: During the device manufacturing stage, set the parameters of the device light source (7) according to the coherent gating setting, and determine the sampling matching relationship of the array detection module (5) according to the sampling theorem; Step 2: Select the frequency sweep mode or time domain mode, power on to activate the corresponding light source (7) and synchronization timing; Step 3: Adjust the position of the reference mirror (15) so that the optical path difference between the sample arm and the reference arm is less than the coherence length; Step 4: Acquire the full-field interference sequence of the area array camera (19); Step 5: Perform digital aberration compensation on the full field-of-view interference sequence, compensate and denoise the volume data at different depths, and derive the three-dimensional diffraction-limited image. Step 6: Output a three-dimensional tomographic image with transverse diffraction limit and axial micron-level resolution.

8. The optical coherence tomography method with optimized digital aberration compensation according to claim 7, characterized in that, In step 1, during the device manufacturing stage, the parameters of the light source (7) of the device are set according to the coherent gating setting, and the sampling matching relationship of the array detection module (5) is determined according to the sampling theorem, as follows: Step 1.1, Setting the coherence gate width: Select the center wavelength λ0 of the light source, the bandwidth Δλ of the light source, and the numerical aperture NA, and calculate the coherence gate width L. c The formula is: L c =min{l tc ,L sc ,L Lsc } (1) Among them l tc l sc l Lsc These represent the widths of the time coherence gate, the lateral spatial coherence gate, and the axial spatial coherence gate, respectively. The light source parameters are determined according to formula (1), and the light source (7) is selected. Mode-S uses a swept frequency light source and a swept frequency full-field OCT as the main device; Mode-T uses a white light source and a filter and a time-domain full-field OCT as the main device. The two share the OCT main body and switch by replacing the light source (7) and the area array camera (19). Step 1.2, Sampling and Matching: To ensure that all defocusing is corrected to the diffraction limit, according to the Nyquist sampling theorem, the phase change between adjacent pixels should be less than π. Therefore, the full-field OCT sampling interval Δx should satisfy: Where w0 is the beam waist radius of the Gaussian beam focal plane, n is the sample refractive index, and M is the magnification between the sample and the pixel size p; Calculate the focal length f of the tube lens (18) based on the object-image relationship. tl The formula is: Select the focal length f of the tube lens (18) according to formula (3). tl This ensures that the transverse spectrum falls completely within the sampling bandwidth, thereby obtaining the optical design of the optical coherence tomography device that satisfies the optimal digital aberration compensation. Step 1.3: Complete the center wavelength λ0, bandwidth Δλ, objective back focal plane waist radius w0, objective numerical aperture NA, and objective focal length f. MO , tubular endoscope tl Matching and shaping the optical path based on pixel size p.

9. The optical coherence tomography method with optimized digital aberration compensation according to claim 8, characterized in that, Step 5 describes digital aberration compensation of the full-field interference sequence, compensation and denoising of volume data at different depths, and deriving a three-dimensional diffraction-limited image, which includes the following steps: Step 5.1, Solving for the frontal complex signal, as follows: (1) In the sweep frequency mode Mode-S, the three-dimensional interference spectrum signal acquired by the area array camera (19) is represented as I(λ; x, y). According to the wavelength interval of the sweep frequency light source, the interference sequence of each pixel is resampled to obtain the wavenumber linearized spectral interference signal I(k; x, y), where k represents the spatial frequency. For each pixel (x, y) in the area array camera (19), the Fourier transform is solved to obtain the initial three-dimensional volume data. The frontal complex signal V(x, y; z) of this data at the sample depth z is extracted. (2) In the time-domain Mode-T mode, the area array camera (19) acquires a two-dimensional interference sequence l N (x, y), the light intensity expression is: Analytical demodulation of four-step phase-shifting or carrier frequency phase-shifting interferograms; The full-field OCT amplitude signal A(x,y) of the four-step phase-shifting interferogram is: The full-field OCT phase signal φ(x,y) is: After phase unpacking, the generated two-dimensional complex signal V(x,u) is: V(x,y)=A(x,y)exp(-iφ(x,y)) (7) Step 5.2, Two-dimensional FFT, as follows: (1) In the sweep frequency mode (Mode-S), the front complex signal V(x,y;z) is extracted, and the spatial frequency domain spectrum of the signal is solved. The spectral domain signal V(f) can be obtained through a two-dimensional Fourier transform. x ,f y ;z), where (f x ,f y () represents the spectral coordinate system; (2) In the time-domain mode (Mode-T), the spatial frequency domain spectrum of the complex signal V(x,y) can be obtained by solving the two-dimensional Fourier transform, which yields the spectral domain signal V(f). x ,f y ), where (f x ,f y () represents the spectral coordinate system; Step 5.3, digital aberration correction, is as follows: (1) In the swept-frequency mode (Mode-S), a forward diffraction model is adopted, and a digital refocusing algorithm or a multi-focus complex averaging algorithm to suppress multiple scattering light is used to analyze the complex interference spectrum V(f). x ,f y ;z) Perform phase retrieval; In the digital refocusing algorithm, the deconvolution operator for phase retrieval is H -1 (f x ,f y ;z d ), expressed as: Where z d This indicates the amount of defocus that needs to be estimated; The three-dimensional reflectivity signal S(x,y,z) after refocusing is expressed as: (2) In the time-domain Mode-T, angular spectrum propagation or aperture filtering methods are used to apply V(f) to the frequency domain. x ,f y Phase retrieval is performed, restoring the lateral diffraction-limited resolution in the depth direction. The deconvolution operator for phase retrieval is H. -1 (f x ,f y ;z d After performing deconvolution using this operator, the phase-recovered two-dimensional structure signal T(x,y) is obtained by performing a two-dimensional inverse Fourier transform. Step 5.4, depth signal acquisition, as detailed below: (1) In the sweep frequency mode Mode-S, the sweep frequency full field OCT principle does not require the use of an axial displacement stage. After performing Fourier transform on the interference spectrum, the depth signal can be obtained. At this time, the diffraction-limited resolution three-dimensional tomographic structure S(x,y,z) is directly output. Select the sample axial displacement stage (16) to obtain other depths, or select blood flow or phase function imaging as signal supplement. (2) In the time-domain mode Mode-T, the sample axial displacement stage (16) is used to stack the corresponding two-dimensional tomographic images in depth order to form a three-dimensional tomographic data signal T(x,y,z), and blood flow or phase function map is superimposed.

10. The aberration-optimized digital compensation optical coherence tomography method according to claim 9, characterized in that, In the frequency sweep mode Mode-S described in step 5.1, wavenumber linearization is based on k-clock triggering for resampling, and the resampling error is no greater than the width of one sampling point; In the sweep frequency mode Mode-S described in step 5.3, when the depth exceeds 0.5mm, a multi-focus complex averaging algorithm is used to suppress multiple scattering noise.

Images