Efficient non-line-of-sight imaging method based on polarization differential correlation

Through polarization difference correlation technology, the coherence matrix and phase recovery algorithm of polarized speckle are used to solve the calculation and hardware cost challenges of hidden objects reconstruction in non-sight imaging, achieving stable and efficient imaging, providing millimeter-level resolution and Lu Great.

CN118518592BActive Publication Date: 2024-11-29NANJING UNIV OF SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411000555.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-11-29
Estimated Expiration
2044-07-25

AI Technical Summary

Technical Problem

The existing non-field-of-sight imaging methods have high computational requirements, high hardware cost, limited imaging field of view, time-consuming sampling iteration and generalization challenges across application scenarios, especially steady-state NLOS imaging in mechanical scanning and multiple Instability and complexity problems are faced in spectral modulation.

Method used

Using high-efficiency non-field-of-sight imaging method based on polarization difference correlation, the polarization difference is used to encode and reconstruct hidden objects through polarization speckle illumination by laser source, avoid mechanical scanning, use a single light source for indirect illumination, and use the coherence matrix of polarization speckle The and phase recovery algorithm realizes autocorrelation estimation and reconstruction of hidden objects.

Benefits of technology

A stable hidden object imaging quality is achieved, providing millimeter-level imaging resolution, reducing instability brought about by mechanical scanning, and improving the robustness and efficiency of the system through a single-shot polarization speckle illumination strategy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118518592B_ABST
    Figure CN118518592B_ABST
Patent Text Reader

Abstract

The present invention discloses an efficient non-field of view imaging method based on polarization differential correlation. The output light field of a virtual light source is polarized laser speckle, which is transmitted to the surface of a hidden object through free space, and generates an outgoing light field through reflection and diffuse reflection; the outgoing light field continues to propagate to another position of the relay wall, serving as a virtual detector VD; a light field is generated at VD, and a camera captures an intensity image; the autocorrelation estimation value of the hidden object is calculated based on the autocorrelation estimation method; and the hidden object is reconstructed by adopting a phase recovery algorithm. The present invention utilizes the polarization difference of laser speckle to design a single polarization speckle illumination strategy; the fast and stable real-time encoding of the hidden object ensures the stable imaging quality of the PDC‑NLOS system; the proposed method exhibits millimeter-level imaging resolution when imaging horizontal and vertical striped objects.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of non-line-of-sight imaging, and in particular relates to a high-efficiency non-line-of-sight imaging method based on polarization differential correlation. Background Art

[0002] Non-line-of-sight (NLOS) imaging uses relay walls to detect, identify, and reconstruct hidden objects around corners. It has applications in key areas such as autonomous driving, remote sensing, and medical diagnosis. However, since the hidden object signal must undergo at least three diffuse reflections before being collected through the relay wall, and the Fourier amplitude is not fully sampled, the reconstruction of the hidden object is an ill-posed inverse problem.

[0003] Existing NLOS imaging methods are divided into transient and steady-state types according to the temporal resolution of the hardware used. Transient NLOS imaging uses high-frequency pulsed lasers and high-temporal resolution detectors to encode the three-dimensional information of hidden objects. It reconstructs hidden objects by calculating the flight time of multiple scattered photons. Initially, Velten et al. proposed a transient NLOS imaging solution based on the back-projection method of a streak camera (Velten A, Willwacher T, Gupta O, et al. Recovering three-dimensional shape around a corner using ultrafast time-of-flight imaging [J]. Nature communications, 2012, 3 (1): 745). In order to improve the efficiency and accuracy of hidden object reconstruction, many researchers use single-photon avalanche diodes as detectors and introduce reconstruction methods based on light cone transformation, fk migration and virtual wave phasor fields. Various scanning methods have also been proposed to optimize the acquisition and reconstruction efficiency of the system. Despite this, transient methods still face challenges such as high computational requirements and high hardware costs.

[0004] Reconstruction methods based on speckle correlation, computational periscope, wavefront modulation, edge camera and neural network have been developed for steady-state NLOS imaging. These methods need to be further improved to address challenges such as limited imaging field of view, time-consuming sampling iterations and generalization across application scenarios.

[0005] Indirect imaging correlogram (IIC) has emerged as a promising method to enhance the performance of steady-state NLOS imaging. This technique exploits the intensity correlation of multiple speckles on the wall in spatial patterns to effectively image hidden objects. To further improve the process, researchers established a noise model for IIC and developed a deep inverse correlation convolutional neural network. This method can reconstruct hidden objects using only two short exposure images of 0.125 seconds. However, this method requires mechanical scanning of the wall using a laser source to achieve independent speckle illumination. This scanning process introduces instability to the imaging optical path. To address this issue, researchers explored the color axis difference of the speckle to achieve single-shot steady-state NLOS imaging. By adopting a single acquisition method and exploiting uncorrelated speckles on the color axis, this method provides redundant information across multiple spectral channels for the hidden object. Therefore, this method exhibits robustness to system vibration and color perturbations. However, the challenge comes from the mixed illumination from multiple light sources, which introduces instability to the power spectrum of the hidden object on different channels. In addition, multi-spectral modulation brings complexity to the acquisition of color information of the hidden object. Summary of the invention

[0006] The object of the present invention is to provide a highly efficient non-line-of-sight imaging method based on polarization differential correlation to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: an efficient non-line-of-sight imaging method based on polarization differential correlation, comprising:

[0008] The laser source is expanded through the optical path and irradiates the relay wall as a virtual light source VS; the coherent light beam passes through the polarizer and diffuses on the relay wall, generating an initial polarization angle and random phase differences;

[0009] Output light field of virtual light source VS is the polarized laser speckle, Transmitted through free space to the surface of the hidden object, generating an outgoing light field through reflection and diffuse reflection ;

[0010] Continue to propagate to another location on the relay wall, serving as a virtual detector VD; the light field at VD It is expressed as:

[0011] (3)

[0012] is the two-dimensional Fourier transform;

[0013] Intensity image captured by the camera It is expressed as:

[0014] (4)

[0015] In the above two equations, and is a constant term, The noise is caused by the wall background;

[0016] According to the two-point intensity correlation technique and the coherent representation of polarization speckle intensity, the observation angle and The speckle correlation coefficient at can be expressed as:

[0017] (5)

[0018] in , The polarization angles are and The intensity distribution of laser speckle, , is the standard deviation of intensity, let , then formula (5) is simplified to:

[0019] (6)

[0020] The number of periodic patterns of speckle images in deep inverse correlation imaging is replaced by the number of polarization difference speckles. The power spectrum of is expressed as:

[0021] (7)

[0022] and

[0023] (8)

[0024] in represents the number of polarization difference scattered spots, Represents different polarization angles, The polarization angle is The two-dimensional distribution of laser speckle intensity, represents the two-dimensional autocorrelation of the target, is the unit impulse function;

[0025] Polarization speckle intensity collected by the camera and the incident speckle intensity Proportional to:

[0026] (9)

[0027] in , , , , , is the element of the speckle field coherence matrix; Coherence Factor The phase of ; R is the parameter used to simplify the formula, is the initial polarization angle; Obey the polarization angle The collected speckle intensity value is corrected according to the sinusoidal distribution of

[0028] Calculate the autocorrelation estimation value of the hidden object based on the autocorrelation estimation method established by formula (7) and formula (8);

[0029] Hidden objects are reconstructed through phase recovery algorithm.

[0030] An efficient non-line-of-sight imaging system based on polarization differential correlation is used to implement the above method. The system includes an incident laser source, a collimating lens, a deflector, a relay wall, a DMD, a polarization camera and its focusing lens. The incident laser source generates coherent light, which is collimated by the collimating lens and then irradiated to the relay wall through the deflector. The laser speckle generated by the relay wall is incident on a target displayed by the DMD. The target reflects a part of the speckle, which is propagated to the relay wall through the far field and collected by the polarization camera and its focusing lens.

[0031] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the steps of the above method are implemented when the processor executes the program.

[0032] A computer-readable storage medium stores a computer program, which implements the steps of the above method when executed by a processor.

[0033] A computer program product comprises a computer program, which implements the steps of the above method when executed by a processor.

[0034] Compared with the prior art, the present invention has the following beneficial effects: the present invention encodes and reconstructs hidden objects based on the imaging correlation technique of independent polarized speckle illumination; the hidden objects can be accurately restored by utilizing the polarization difference of the speckles; the efficient single-shot polarized speckle illumination strategy avoids mechanical scanning; using a single light source for indirect illumination, the PDC-NLOS imaging system can provide stable imaging quality of hidden objects; for imaging horizontal and vertical striped objects, the PDC-NLOS imaging method is proven to have millimeter-level imaging resolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0036] Figure 1 Schematic diagram of PDC-NLOS, where (a) is the imaging process of PDC-NLOS and (b) is the principle diagram of PDC-NLOS.

[0037] Figure 2 It is the relationship curve between the polarization angle and the correlation coefficient of polarization speckle.

[0038] Figure 3 Figure 2 is the PDC-NLOS experimental setup, where (a) is the system setup diagram, (b) is the polarization camera and lens used, and (c) is the DMD used to display the object.

[0039] Figure 4 The PDC-NLOS experimental results, where (a1)-(a4) are the 4-polarization channel intensity images collected by the polarization camera; (b1)-(b4) are the hidden objects to be detected; (c1)-(c4) are the autocorrelation estimates; (d1)-(d4) are the reconstruction results using the HIO algorithm.

[0040] Figure 5 The results of the PDC-NLOS resolution test experiment, where (a1)-(a2) are horizontal and vertical stripe objects, (b1)-(b2) are autocorrelation estimates, (c1)-(c2) are the results of HIO reconstruction, and (d) is the power spectrum intensity distribution perpendicular to the object direction.

[0041] Figure 6 Practical system setup for the PDC-NLOS imaging approach.

[0042] Figure 7 This is an imaging experiment of the number of polarization difference scattered spots, where (a) is the polarization difference scattered spot used, (b) is the autocorrelation calculated by PDC-NLOS, and (c) is the reconstruction of the hidden object. DETAILED DESCRIPTION

[0043] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Non-line-of-sight (NLOS) imaging can detect and reconstruct hidden objects around corners, and has broad application prospects in autonomous driving, remote sensing, and medical diagnosis. However, existing steady-state NLOS imaging methods face challenges in achieving high efficiency and high accuracy due to the need for multiple diffuse reflections and incomplete Fourier amplitude sampling. This paper proposes a new steady-state NLOS imaging technique based on polarization differential correlation (PDC-NLOS). The method is based on imaging correlography of independent polarized speckle illumination to encode and reconstruct hidden objects. By exploiting the polarization difference of the speckle, the hidden object can be accurately recovered. The efficient single-shot polarized speckle illumination strategy avoids mechanical scanning. Using a single light source for indirect illumination, the PDC-NLOS imaging system can provide stable imaging quality of hidden objects. The PDC-NLOS imaging method is demonstrated to have millimeter-level imaging resolution for imaging horizontal and vertical striped objects.

[0045] The proposed PDC-NLOS imaging process is as follows Figure 1 shown.

[0046] After the laser source is expanded through the optical path, it is used as a virtual light source (VS) to irradiate the relay wall. The coherent light beam passes through the polarizer and diffuses on the relay wall, generating an initial polarization angle and random phase difference. The output light field of VS is polarized laser speckle, which can be described as:

[0047] (1)

[0048] in and Represents the light field in horizontal and vertical unit vectors and The amplitude of the polarization component on , is the random phase factor generated by VS diffusion.

[0049] Transmitted through free space to the surface of the hidden object, generating an outgoing light field through reflection and diffuse reflection ,

[0050] (2)

[0051] in is the light field transmission distance, is the VS axial position, is the albedo function of the hidden object.

[0052] It continues to propagate to another location on the relay wall, which serves as a virtual detector (VD). Since the distance between laser speckle particles is small enough, the propagation process from the hidden object to VD can be regarded as far-field propagation. Therefore, the light field at VD It can be expressed as,

[0053] (3)

[0054] is the two-dimensional Fourier transform;

[0055] Intensity image captured by the camera It can be expressed as,

[0056] (4)

[0057] In the above two equations, and is a constant term, The noise is caused by the wall background.

[0058] According to the two-point intensity correlation technique and the coherent representation of polarization speckle intensity recorded in formula (S1) to formula (S5), the observation angle and The speckle correlation coefficient at can be expressed as:

[0059] (5)

[0060] in , The polarization angles are and The intensity distribution of laser speckle, , is the standard deviation of the intensity. Let , then Formula 5 can be simplified to:

[0061] (6)

[0062] As shown in Formula 6, the correlation coefficient between polarization speckles varies sinusoidally with the difference in polarization angle. Figure 2 As shown, 0.5 is taken as the empirical threshold, and the appropriate polarization angle difference can obtain completely uncorrelated speckles, which can be used to encode hidden objects.

[0063] The number of periodic patterns of speckle images in deep inverse correlation imaging is replaced by the number of polarization difference speckles. The power spectrum of can be expressed as,

[0064] (7)

[0065] and

[0066] (8)

[0067] in represents the number of polarization difference scattered spots, Indicates different polarization angles. The polarization angle is Two-dimensional distribution of laser speckle intensity; represents the two-dimensional autocorrelation of the target, is the unit impulse function.

[0068] By subtracting Formula 8 from Formula 7, the initial autocorrelation of the hidden object can be estimated, which is used to calculate the Fourier amplitude of the target and realize target reconstruction in combination with the phase recovery algorithm.

[0069] Polarization speckle intensity collected by the camera and the incident speckle intensity Proportional,

[0070] (9)

[0071] in , . , , , is the element of the speckle field coherence matrix in formula (S2); Coherence Factor The phase of the light field is used to measure the effective phase difference between the x and y components. R is a parameter used to simplify the formula. is the initial polarization angle.

[0072] Obviously, Obey the polarization angle The acquired speckle intensity value needs to be corrected according to formula 9.

[0073] The present invention constructs Figure 3 The system shown is used to verify the imaging capability of the proposed method.

[0074] The wavelength of the laser was set to 532 nm. The distance between the object and the virtual light source VS was 300 mm. The distance between the camera and the virtual detector VD was 500 mm. The hidden object was displayed on a digital micromirror device (DMD). The size of the DLP area was 14.0 mm × 10.5 mm, and the number of pixels was 1024 × 768. The area of ​​object display was approximately 10 mm × 10 mm. A polarization camera (DYK 33UX250, ImagingSource) was focused onto the virtual detector VD using a 50 mm focal length lens. The relay wall was painted with latex paint.

[0075] Using the above system, we imaged four hidden objects "N", "L", "O", and "S", which are approximately 20 mm × 20 mm in size. Through an effective single-shot polarization speckle illumination strategy, the hidden objects are encoded by four uncorrelated polarization speckles. The power spectrum of the object is obtained by intensity acquisition of the polarization camera. After correcting the collected polarization speckle using Formula 9, the autocorrelation estimation method established based on Formula 7 and Formula 8 calculates the autocorrelation estimate of the hidden object. The hidden object can be reconstructed by adopting a phase recovery algorithm, such as the hybrid input-output (HIO) algorithm, the multiplier-based alternating direction method, or a deep learning network model. The reconstruction process in this paper adopts the HIO algorithm.

[0076] Figure 4 This is the PDC-NLOS experimental result.

[0077] Since polarization encoding uses only a single laser source, the interference caused by the incident light calibration process is reduced. The average object SSIM of the proposed PDC-NLOS reconstruction is about 0.76, which is significantly better than the single NLOS imaging reconstruction result based on the color axis differential correlation method.

[0078] The imaging resolution of the system is measured by calculating the ratio of the DMD resolution to the actual display size and setting the horizontal and vertical resolution test patterns. The pattern is set to two parallel white stripes with a length of 8 mm, a width of 1 mm, and a spacing of 1 mm.

[0079] Figure 5 The experimental results of PDC-NLOS resolution test are as follows: Figure 5 As shown in (c1) and (c2) in Figure 1, it is not difficult to find that the proposed PDC-NLOS imaging method can resolve horizontal stripes at millimeter-level resolution, while the reconstruction results of vertical stripes are aliased. The imaging results show that the PDC-NLOS imaging system has millimeter-level resolution, but the vertical resolution is better than the horizontal resolution. Figure 5 The curves in (d) represent the power spectrum intensity distribution of the horizontal and vertical stripes collected by the system in the vertical and horizontal directions as the object, respectively. The power spectrum information contrast of the horizontal stripes in the vertical direction is greater than the power spectrum information contrast of the vertical stripes in the horizontal direction, which confirms that the vertical resolution of the system is better than the horizontal resolution. From the perspective of system construction, the laser irradiated from VS is incident on the surface of the object obliquely. After emitting from the surface of the object, it is collected obliquely by VD, which compresses the horizontal pixel distribution of the object. The object pixels in the vertical direction will not be compressed, so the horizontal resolution will be lower than the vertical resolution.

[0080] In summary, this paper proposes a PDC-NLOS imaging method that can efficiently image hidden objects around corners. We have demonstrated the imaging capability of the proposed method through a practical experimental setup. Through the designed single-shot polarized speckle illumination strategy, PDC-NLOS can efficiently encode hidden objects without mechanical scanning. Since there is only one laser source, the interference caused by the incident light calibration process is reduced. PDC-NLOS provides valuable insights for the stable and real-time development of steady-state NLOS imaging.

[0081] The PDC-NLOS imaging system provided by the present invention is composed of an incident laser source, a collimating lens, a deflector, a relay wall, a DMD, a polarization camera and a focusing lens thereof. Figure 6 shown.

[0082] The wavelength of the incident laser source is set to 532 nm. The polarization camera contains a microlens array with four polarization angles to achieve polarization imaging.

[0083] The coherent representation of polarization speckle intensity is explained below:

[0084] The coherence matrix It can be expressed as,

[0085] (S1)

[0086] in is the phase difference between the horizontal component and the vertical component. Since the incident light is a quasi-monochromatic plane Wave, . Formula S1 can be written as,

[0087] (S2)

[0088] Consider The speckle field observed at It can be expressed as,

[0089] (S3)

[0090] strength It can be expressed as,

[0091] (S4)

[0092] Therefore, the observation angle and The speckle correlation coefficient at can be expressed as

[0093] (S5)

[0094] The influence of polarization difference speckle number on imaging: According to the depth inverse correlation imaging technology, the represents the number of speckle periodograms. PDC-NLOS replaces this process with polarization-difference speckle encoding. Here we explore the effect of the number of polarization modes of the polarization-difference speckle pattern blocks used on the hidden object autocorrelation estimation and imaging reconstruction.

[0095] like Figure 7 As shown in the figure, when the four polarization difference dispersion spots are used to encode the hidden object at a time, the autocorrelation has good contrast and structural integrity, and the reconstruction result has a clear and complete structure. When the number of polarization difference dispersion spots is reduced to 3, the corresponding formulas 7 and 8 are The number of polarization difference scattered spots is reduced to 2, the contrast of autocorrelation estimation continues to decrease, the structure begins to be defective, the reconstruction result lacks some structure, and the contrast is reduced.

[0096] It is worth noting that the reconstruction results obtained using two polarization difference scattered spots of 0° and 90° and two polarization difference scattered spots of 0° and 180° are not much different. This is because the polarization difference correlation method can obtain stable difference values ​​of polarization speckles outside the correlation range, combined with the designed single polarization speckle illumination strategy and intensity calibration method, to ensure the stability of reconstruction.

[0097] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be considered as exemplary and non-restrictive in all respects.

[0098] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. An efficient non-line-of-sight imaging method based on polarization differential correlation, characterized in that: include: The laser source is expanded through the optical path and irradiates the relay wall as a virtual light source VS; the coherent light beam passes through the polarizer and diffuses on the relay wall, generating an initial polarization angle and random phase differences; Output light field of virtual light source VS is the polarized laser speckle, Transmitted through free space to the surface of the hidden object, generating an outgoing light field through reflection and diffuse reflection ; Continue to propagate to another location on the relay wall, serving as a virtual detector VD; the light field at VD It is expressed as: (3) is the two-dimensional Fourier transform; Intensity image captured by the camera It is expressed as: (4) In the above two equations, and is a constant term, The noise is caused by the wall background; According to the two-point intensity correlation technique and the coherent representation of polarization speckle intensity, the observation angle and The speckle correlation coefficient at can be expressed as: (5) in , The polarization angles are and The intensity distribution of laser speckle, , is the standard deviation of intensity, let , then formula (5) is simplified to: (6) The number of periodic patterns of speckle images in deep inverse correlation imaging is replaced by the number of polarization difference speckles. The power spectrum of is expressed as: (7) and (8) in represents the number of polarization difference scattered spots, Represents different polarization angles, The polarization angle is The two-dimensional distribution of laser speckle intensity, represents the two-dimensional autocorrelation of the target, is the unit impulse function; Polarization speckle intensity collected by the camera and the incident speckle intensity Proportional to: (9) in , , , , , is the element of the speckle field coherence matrix; Coherence Factor The phase of ; R is the parameter used to simplify the formula, is the initial polarization angle; Obey the polarization angle The collected speckle intensity value is corrected according to the sinusoidal distribution of Calculate the autocorrelation estimation value of the hidden object based on the autocorrelation estimation method established by formula (7) and formula (8); Hidden objects are reconstructed through phase recovery algorithm.

2. The method according to claim 1, characterized in that The output light field of the virtual light source VS is polarized laser speckle, which can be described as: (1) in and Represents the light field in horizontal and vertical unit vectors and The amplitude of the polarization component on , is the random phase factor generated by VS diffusion.

3. The method according to claim 2, characterized in that Outgoing light field It is expressed as: (2) in is the light field transmission distance, is the VS axial position, is the albedo function of the hidden object.

4. The method according to claim 1, characterized in that: The coherence matrix It is expressed as: (S1) in is the phase difference between the horizontal component and the vertical component; since the incident light is a quasi-monochromatic plane Wave, , formula (S1) is written as: (S2) Complex Amplitude It is expressed as: (S3) strength It is expressed as: (S4) Therefore, the observation angle and The speckle correlation coefficient at is expressed as (S5)。 5. An efficient non-line-of-sight imaging system based on polarization differential correlation, characterized in that: The system is used to implement the method described in any one of claims 1 to 4, comprising an incident laser source, a collimating lens, a deflector, a relay wall, a DMD, a polarization camera and a focusing lens thereof; the incident laser source generates coherent light, which is collimated by the collimating lens and then irradiated to the relay wall through the deflector; the laser speckle generated by the relay wall is incident on a target displayed by the DMD; the target reflects a portion of the speckle, which is propagated to the relay wall through the far field and collected by the polarization camera and its focusing lens.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Citation Information

Patent Citations

  • Microscopic imaging automatic focusing device based on polarization difference image and method thereof

    CN113433682A

  • On-line calibration-based on-site non-vision field imaging method, device, equipment and medium

    CN115825983A