Leveling system and leveling method based on double-spot common-mode decoupling

Abstract

The present application belongs to the technical field of microscopic imaging automatic focusing, and particularly relates to a leveling system and method based on double-spot common-mode decoupling. The present application is based on an oblique incidence laser automatic focusing framework, adopts double laser beams to form common-mode and differential-mode motion decoupling, perceives chip tilting through overall translation of double spots, detects Z-axis defocusing through spot spacing variation, and obtains three-dimensional to-be-adjusted poses in one measurement, filling the blank of existing products that only support single-dimensional defocusing detection. Polarization light splitting and a quarter-wave plate are used to realize spatial separation of transmitting and receiving light paths, a dichroic mirror and a fluorescence imaging light path share an objective lens, the structure is compact and does not interfere with fluorescence imaging. A closed-loop control strategy clearly defines the leveling first and focusing later process, avoiding repeated iteration. In summary, the present application has significant advancement in multi-dimensional perception, system integration, anti-interference capability and the like.

Description

Technical Field

[0001] This invention belongs to the field of automatic focusing technology for microscopic imaging, and particularly relates to a leveling system and leveling method based on dual-spot common-differential mode decoupling. Background Technology

[0002] Autofocus technology in microscopic imaging is a core technology that determines image clarity and ensures long-term detection stability. In precision microscopic detection scenarios such as high-throughput gene sequencing and semiconductor wafer inspection, as the numerical aperture of microscope objectives continues to increase and the depth of focus of the system continues to decrease (e.g., when NA=1.4, the depth of focus is only about 0.1μm), traditional manual focusing methods can no longer meet the requirements of high-precision and high-efficiency detection. Therefore, autofocus technology has become an indispensable foundation for precision microscopic applications.

[0003] Currently, microscopic autofocus technology is mainly divided into two categories: passive and active. Passive autofocus works based on the image sharpness evaluation function (contrast method) or the phase detection principle, relying on the texture features of the sample itself to complete the focus, and is suitable for samples with obvious structural features; however, this technology has poor adaptability to weakly textured biological samples and fluid dynamic pattern samples, and requires multiple image acquisitions to complete the calculation, resulting in a relatively slow response speed.

[0004] Active autofocus acquires defocus information by emitting a probe signal to the sample, without relying on the sample surface texture. Mainstream technologies include astigmatism, confocal autofocus, and image-based autofocus. Astigmatism uses cylindrical lenses to convert spot deformation signals into focusing error signals, achieving sub-micron focusing accuracy, but it has weak adaptability to non-planar sample surfaces and a small effective working range. Confocal autofocus filters out defocus stray light through a pinhole structure, achieving a working range of ±200μm and a wide adaptability, but its focusing accuracy is relatively limited to approximately ±1.5μm. Image-based autofocus uses a CCD / CMOS image sensor as its core, extracting focal plane position information by analyzing the shape and spacing of reflected light spots. It offers high flexibility but is susceptible to ambient light interference.

[0005] In terms of commercial applications, international companies such as Prior Scientific have launched mature laser autofocus modules. Among them, the PureFocus850 is equipped with an 850nm infrared laser, which can be adapted to mainstream commercial microscopes and is widely used in life sciences and industrial inspection fields; the LF210 laser autofocus system uses a 670nm laser, integrates an intelligent microprocessor and closed-loop servo control architecture, and is compatible with microscopes from multiple brands such as Nikon and Olympus.

[0006] Domestically, the PMS-LFS series laser autofocus sensors from Promis can achieve sub-micron level focusing accuracy, and the Focus X laser autofocus module has a dynamic response time of less than 0.4 seconds and a focusing frequency of up to 5000Hz, demonstrating excellent performance. However, current commercial products, both domestically and internationally, still have significant limitations: First, they all use single-point measurement schemes, which can only detect the defocus amount in a limited area at the center of the field of view and cannot identify the overall tilt posture of the sample; second, they only support unidirectional Z-axis focusing adjustment and lack closed-loop control capabilities for active adjustment of sample position in the XY plane; third, there are few systematic integration application cases in fully automated gene sequencers, and most modules require additional adaptation and debugging to meet the specific application requirements of gene sequencing scenarios.

[0007] Overall, current microscopic autofocus technology exhibits four main development trends: First, the detection dimension is being upgraded, moving from single Z-axis defocus detection to simultaneous sensing of three-dimensional pose quantities to be adjusted, including Z-axis displacement, Rx and Ry tilt angles; second, hardware algorithm optimization is being implemented, replacing traditional four-quadrant detectors with CMOS area array detectors, and combining sub-pixel algorithms to significantly improve the system's anti-interference capabilities; third, the optical path structure is being integrated, with the focusing optical path and imaging optical path sharing the same microscope objective, achieving miniaturization and compact integration of the equipment; and fourth, the control mode is iterating, gradually shifting from traditional passive error compensation to full-process active precision calibration. Summary of the Invention

[0008] In view of this, the present invention aims to provide a leveling system and method based on dual-spot common-difference mode decoupling, to solve the problems of insufficient detection dimension, separation of focusing and imaging systems, complex optical paths, and difficulty in simultaneously achieving focusing accuracy and stability in existing laser focusing technologies. The present invention constructs an optical structure in which the focusing optical path and the fluorescence imaging optical path share the same objective lens, and adopts a design method that separates the focusing laser emission path from the detection path to achieve highly stable and sensitive focusing signal acquisition. Simultaneously, by introducing an axial displacement calculation model based on the lateral offset of the spot, nanometer-level focusing accuracy is achieved. Furthermore, by designing a multi-stage leveling control process, automated high-precision control of sample focusing in the Z direction and XY plane leveling is achieved, thereby meeting sub-micron-level axial focusing accuracy and XY plane adjustment accuracy.

[0009] To achieve the above objectives, the technical solution created by this invention is implemented as follows: A leveling system based on dual-spot common-difference mode decoupling includes an infrared light source module, a beam-splitting and collimating module, a polarization beam-splitting module, a laser detection and imaging module, a fluorescence light source module, a fluorescence detection module, a fluorescence imaging module, a dichroic mirror, a microscope objective, a sample to be tested, a leveling stage, and a focusing control module, wherein: The sample to be tested is placed on a leveling stage. Infrared laser light emitted from the infrared light source module is sequentially incident on the sample surface through a beam-splitting collimation module, a polarization beam-splitting module, a dichroic mirror, and a microscope objective, forming a double infrared spot. The infrared laser light reflected from the sample surface is incident on the laser detection imaging module through the microscope objective, a dichroic mirror, and the polarization beam-splitting module to obtain an infrared image. Similarly, fluorescence emitted from the fluorescence light source module is incident on the sample surface through a fluorescence detection module, a dichroic mirror, and a microscope objective. The fluorescence light reflected from the sample surface is incident on the fluorescence imaging module through the microscope objective, a dichroic mirror, and the fluorescence detection module to obtain a fluorescence image. The focusing control module acquires the measured centroid coordinates of the two infrared spots contained in the infrared image, and calculates the three-dimensional pose of the sample under test by combining the calibrated centroid coordinates of the two infrared spots. Based on the calculation result of the three-dimensional pose of the sample under test, the module drives the microscope objective and the leveling stage to move, thereby achieving parallelism leveling between the plane of the sample under test and the imaging focal plane of the microscope objective.

[0010] Furthermore, the process of obtaining the calibration centroid coordinates of each of the two infrared spots is as follows: Select a flat, uniformly reflective standard sample with a standard pattern, and place the standard sample on the leveling motion table. The fluorescence images acquired by the fluorescence imaging module are input into the sharpness evaluation function to obtain the normalized gradient sum; If the sum of the normalized gradients is less than the preset threshold, the leveling stage is driven to move along the Z-axis, so that the fluorescence imaging module re-acquires the fluorescence image and compares the sum of the normalized gradients of the current fluorescence image with the preset threshold until the sum of the normalized gradients is greater than the preset threshold, thus obtaining the optimal focal plane position of the microscope objective; otherwise, the optimal focal plane position of the microscope objective is obtained directly. The current position of the microscope objective on the Z-axis is taken as the zero reference position. The centroid coordinates of each of the current dual infrared spots are taken as the calibration centroid coordinates. The reference spacing and reference centroid coordinates are calculated based on the calibration centroid coordinates of each of the dual infrared spots. The leveling stage is driven to move along the Z-axis, and the differences between the microscope objective at different positions on the Z-axis and the zero-point reference position are collected, as well as the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis. The differences between the microscope objective at different positions on the Z-axis and the zero-point reference position, as well as the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis, are fitted to obtain the fitting function.

[0011] Furthermore, the three-dimensional pose of the sample to be adjusted is calculated. , , The specific process is as follows: The measured centroid coordinates of the first infrared spot are ( X 1,Y 1), The measured centroid coordinates of the second infrared spot are ( X 2, Y 2), then the measured spacing between the current dual infrared spots D for: ; Based on the fitting function, the difference between the current position of the microscope objective on the Z-axis and the zero-point reference position is calculated. ; Calculate the X-direction offset of the sample under test. and Y-direction offset :

[0012] in,( , () is the reference centroid coordinate, , () represents the measured coordinates of the centroid.

[0013] Furthermore, the beam splitting and collimation module includes a diffraction beam splitter and a multi-beam collimating lens arranged sequentially along the optical path. The diffraction beam splitter is used to split the single-path incident infrared laser into two independent sub-beams, and the multi-beam collimating lens is used to collimate and shape the two independent sub-beams after beam splitting, so that each independent sub-beam becomes parallel light.

[0014] Furthermore, the polarization beam splitting module includes a polarization beam splitter and a quarter-wave plate arranged sequentially along the optical path, which separate the two parallel beams emitted from the beam splitting collimation module into P-polarized light and S-polarized light, and the quarter-wave plate converts the P-polarized light from linearly polarized light into circularly polarized light.

[0015] Furthermore, the laser detection imaging module includes a bandpass filter and a CMOS focusing spot detector arranged sequentially along the optical path. The infrared laser reflected from the sample surface is filtered by the bandpass filter after passing through a microscope objective, a dichroic mirror, and a polarization beam splitter to remove background noise, and then incident on the CMOS focusing spot detector to obtain an infrared image.

[0016] A leveling method based on dual-spot common-difference mode decoupling is implemented using a leveling system based on dual-spot common-difference mode decoupling, and specifically includes the following steps: S1: The centroid coordinates of the two infrared spots are calibrated using standard samples, and the difference between the microscope objective at different Z-axis positions and the zero-point reference position, as well as the spacing between the two infrared spots at different Z-axis positions, are fitted to obtain the fitting function. S2: Place the sample to be tested on the leveling motion stage, and use the laser detection imaging module to image the dual infrared light spots on the surface of the sample to obtain the measured distance between the current dual infrared light spots. S3: Input the measured spacing of the current dual infrared spots into the fitting function for calculation, obtain the difference between the current position of the microscope objective on the Z axis and the zero reference position, and calculate the X-direction offset and Y-direction offset of the sample to be tested, and obtain the three-dimensional pose of the sample to be tested. S4: If the three-dimensional pose of the sample to be tested is not zero, the microscope objective and the leveling stage are moved based on the three-dimensional pose of the sample to be tested. The laser detection imaging module is used to re-acquire the dual infrared spot image of the sample surface. The measured spacing of the newly acquired dual infrared spot is used to replace the measured spacing of the current dual infrared spot. Step S3 is executed. Otherwise, the parallelism leveling of the test plane of the sample to be tested and the imaging focal plane of the microscope objective is directly completed.

[0017] Furthermore, step S1 specifically includes the following steps: S11: Select a flat, uniformly reflective standard sample with a standard pattern, and place the standard sample on the leveling motion table. S12: Input the fluorescence image acquired by the fluorescence imaging module into the sharpness evaluation function to obtain the normalized gradient sum; S13: If the sum of normalized gradients is less than the preset threshold, drive the leveling stage to move along the Z-axis, so that the fluorescence imaging module re-acquires the fluorescence image and compares the sum of normalized gradients of the current fluorescence image with the preset threshold until the sum of normalized gradients is greater than the preset threshold, and obtain the optimal focal plane position of the microscope objective; otherwise, directly obtain the optimal focal plane position of the microscope objective. S14: Take the current position of the microscope objective on the Z-axis as the zero reference position, take the centroid coordinates of each of the current dual infrared spots as the calibration centroid coordinates, and calculate the reference spacing and reference centroid coordinates based on the calibration centroid coordinates of each of the dual infrared spots. S15: Drive the leveling stage to move along the Z-axis, collect the difference between the microscope objective at different positions on the Z-axis and the zero reference position, as well as the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis, and fit the difference between the microscope objective at different positions on the Z-axis and the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis to obtain the fitting function.

[0018] Furthermore, in step S12, the sharpness evaluation function adopts the image edge gradient method.

[0019] Furthermore, in step S3, the three-dimensional pose of the sample to be tested is calculated ( , , Specifically, it includes: S31: The measured centroid coordinates of the first infrared spot are ( X 1, Y1), The measured centroid coordinates of the second infrared spot are ( X 2, Y 2), then the measured spacing between the current dual infrared spots D for: ; S32: Input the measured spacing of the current dual infrared spots into the fitting function to calculate the difference between the current position of the microscope objective on the Z-axis and the zero-point reference position. ; S33: Calculate the X-direction offset of the sample to be tested. and Y-direction offset :

[0020] in,( , () is the reference centroid coordinate, , () represents the measured coordinates of the centroid.

[0021] Compared with the prior art, the present invention can achieve the following beneficial effects: (1) The leveling system and method based on dual-spot common-difference mode decoupling described in this invention uses two laser beams parallel to the objective lens optical axis but symmetrically distributed at a certain distance from the optical axis, incident from the objective lens side. At this time, on the object side, the laser beam is obliquely incident onto the sample and returns to the objective lens, hence the name oblique incidence. The oblique incidence autofocus frame is adopted, relying on the dual laser beams to achieve decoupling of common-mode and differential-mode motion: the tilt state of the chip is sensed by the overall translation of the dual spots, and the Z-axis defocus is detected by the change of the spot spacing. A single measurement can obtain the three-dimensional pose information of the sample to be adjusted, filling the technical gap of existing products (such as Prior PureFocus850, etc.) that only support single-dimensional defocus detection. By using polarization beam splitting and a quarter-wave plate, the spatial separation of the transmitting and receiving optical paths is achieved. At the same time, the objective lens is shared by the dichroic mirror and the fluorescence imaging optical path, resulting in a compact structure that does not interfere with fluorescence imaging. The closed-loop control strategy clearly defines the process of "leveling first, then focusing" to avoid repeated iterations. In summary, this invention has significant advantages in terms of multi-dimensional pose perception capability, system integration, and environmental interference resistance.

[0022] (2) The leveling system and leveling method based on dual-spot common-differential mode decoupling created by the present invention adopts a common optical path structure in which the focusing optical path and the fluorescence imaging optical path share the same microscope objective. The efficient beam combining and separation of infrared laser and visible fluorescence are achieved through a dichroic mirror. There is no need to configure a separate objective or independent optical path for the focusing system, which reduces the number of optical components and the volume of the optomechanical structure, improves the system integration and compactness, and is easy to embed into various microscope imaging devices.

[0023] (3) The leveling system and method based on dual-spot common-mode decoupling described in this invention, combined with a polarization beam splitter (PBS) and a quarter-wave plate, achieves physical spatial isolation between the focusing laser transmitter and the reflecting receiver. This polarization isolation mechanism effectively suppresses the interference of stray light and laser echo on the light source and detector, and significantly improves the signal-to-noise ratio of the detection signal and the stability of the system during long-term operation.

[0024] (4) The leveling system and method based on dual-spot covariance mode decoupling described in this invention are based on a displacement calculation model of dual-spot lateral offset. The spot displacement is extracted through a sub-pixel centroid positioning algorithm, and the defocus amount is inverted using a pre-calibrated displacement-defocus mapping curve. This method avoids the dependence of traditional astigmatism methods on the precision of cylindrical lenses, and can achieve sub-micron level focusing accuracy, meeting the high-precision imaging requirements such as microscopic gene sequencing.

[0025] (5) The leveling system and method based on dual-spot common-differential mode decoupling described in this invention proposes a multi-stage closed-loop control strategy. By decoupling the common-mode displacement and differential-mode displacement of the spot in real time, the leveling motion stage is driven to perform tilt correction and precise focusing. This process avoids repeated iterations between defocusing and tilt adjustment, significantly improving the system convergence speed.

[0026] (6) The leveling system and method based on dual-spot common-differential mode decoupling described in this invention adopts a single light source combined with a beam splitter to obtain the three-dimensional pose information of the sample to be adjusted through high-precision spot positioning. Compared with the existing technical solutions that use multiple lasers to realize defocus detection and tilt sensing respectively, it significantly reduces the number and cost of optical components, while simplifying the optical path assembly and calibration process, and has good engineering economy and promotion value. Attached Figure Description

[0027] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a schematic diagram of the leveling system based on dual-spot common-differential mode decoupling as described in an embodiment of the present invention; Figure 2 A schematic diagram of the polarization beam splitting optical path described in the embodiment of the present invention; Figure 3 The following are schematic diagrams of the optical path changes and spot shifts corresponding to Z-axis defocusing in the embodiments of the present invention: (a) is a schematic diagram of the optical path changes corresponding to Z-axis defocusing, and (b) is a schematic diagram of the spot shifts corresponding to Z-axis defocusing. Figure 4The following are schematic diagrams of the optical path changes and spot shifts when the XY plane of the sample is clear, as described in the embodiments of the present invention: (a) is a schematic diagram of the optical path changes when the XY plane of the sample is clear, and (b) is a schematic diagram of the spot shifts when the XY plane of the sample is clear. Figure 5 The following is a mapping diagram of the light spot characteristics and defocusing situation described in the embodiment of the present invention: (a) is a diagram showing the relationship between the Z-axis displacement and the change of the Euclidean distance between the two light spots; (b) is a diagram showing the relationship between the connection vector of the zero-position light spot centroid and the measured light spot centroid and the change of the dihedral angle size and orientation.

[0028] Explanation of reference numerals in the attached figures: 1. Infrared laser; 2. Laser collimator; 3. Beam splitter; 4. Adjustable collimation assembly; 5. Polarization beam splitter; 6. Quarter-wave plate; 7. Bandpass filter; 8. CMOS focusing spot detector; 9. Dichroic mirror; 10. Microscope objective; 11. Sample to be tested; 12. Leveling stage; 13. Tube mirror; 14. Beam splitter prism; 15. CMOS imaging detector; 16. Fluorescent laser collimator; 17. Fluorescent excitation source. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof.

[0030] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0031] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0032] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0033] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0034] like Figure 1 As shown, this invention proposes a leveling system based on dual-spot common-difference mode decoupling, comprising an infrared light source module, a beam splitting and collimation module, a polarization beam splitting module, a laser detection and imaging module, a fluorescence light source module, a fluorescence detection module, a fluorescence imaging module, a dichroic mirror 9, a microscope objective 10, a sample to be tested 11, a leveling motion stage 12, and a focusing control module, wherein: The sample 11 to be tested is placed on the leveling stage 12. The infrared laser emitted by the infrared light source module is sequentially incident on the surface of the sample 11 through the beam-splitting collimation module, the polarization beam-splitting module, the dichroic mirror 9, and the microscope objective 10, forming a double infrared spot. The infrared laser reflected from the surface of the sample 11 is incident on the laser detection imaging module through the microscope objective 10, the dichroic mirror 9, and the polarization beam-splitting module to obtain an infrared image. The fluorescence emitted from the fluorescence light source module is incident on the surface of the sample 11 through the fluorescence detection module, the dichroic mirror 9, and the microscope objective 10. The fluorescence reflected from the surface of the sample 11 is incident on the fluorescence imaging module through the microscope objective 10, the dichroic mirror 9, and the fluorescence detection module to obtain a fluorescence image. The focusing control module acquires the measured centroid coordinates of the two infrared spots contained in the infrared image, and calculates the three-dimensional pose of the sample 11 to be adjusted by combining the calibrated centroid coordinates of the two infrared spots. Based on the calculation result of the three-dimensional pose of the sample 11 to be adjusted, the microscope objective 10 and the leveling stage 12 are driven to move, so as to achieve parallelism leveling between the plane of the sample 11 to be measured and the imaging focal plane of the microscope objective 10.

[0035] It should be noted that this invention discloses a method and system for dynamically determining the position and orientation of a sample in a microscopic system, and for achieving automatic focusing and leveling of the sample. The key feature of this invention is that the focusing and imaging systems share a common optical path. By utilizing the projection of multiple focusing lasers and the precise positioning of the detection spot, the three-dimensional pose of the sample is calculated, solving the problems of insufficient detection dimension, separation of the focusing and imaging systems, and difficulty in simultaneously achieving focusing accuracy and stability in existing laser focusing technologies. Specifically, this invention mainly includes an infrared light source module, a beam splitting and collimation module, a polarization beam splitting module, a laser detection and imaging module, and a focusing control module. By constructing an optical structure where the focusing optical path and the fluorescence imaging optical path share the same objective lens, and by adopting a design that separates the focusing laser emission path and the detection path, high-stability and high-sensitivity focusing signal acquisition is achieved. By introducing a three-dimensional pose calculation method based on the lateral offset of the common mode and differential mode of the spot, sub-micron-level three-dimensional pose of the sample is acquired. Furthermore, by designing a multi-stage leveling control process, automated control of sample focusing in the Z-direction and leveling in the XY plane is achieved.

[0036] Furthermore, this invention utilizes an autofocus framework based on obliquely incident lasers, employing dual laser beams to decouple common-mode and differential-mode motion. It senses chip tilt through the overall translation of the dual-spot system and detects Z-axis defocus by monitoring changes in spot spacing. A single measurement reveals the three-dimensional pose to be adjusted, filling the gap in existing products (such as Prior PureFocus850) that only support single-dimensional defocus detection. Polarization beam splitting and a quarter-wave plate 6 achieve spatial separation of the transmitting and receiving optical paths. Simultaneously, a dichroic mirror 9 shares an objective lens with the fluorescence imaging optical path, resulting in a compact structure that does not interfere with fluorescence imaging. The closed-loop control strategy clearly defines the "leveling first, then focusing" process, avoiding repeated iterations. In summary, this invention demonstrates significant advancements in multi-dimensional sensing, system integration, and anti-interference capabilities.

[0037] The optical system structure is briefly described below: (1) Infrared light source module, including an infrared laser 1 and a laser collimator 2 arranged sequentially along the optical path, is used to emit a stable, high-quality infrared laser beam and provide beam collimation. The divergence angle of the collimated laser beam should not exceed 1 mrad. The infrared band is selected as the light source to significantly separate it from commonly used fluorescence excitation wavelengths (470 nm, 532 nm) and imaging bands. It can be efficiently combined with the imaging optical path through a dichroic mirror 9, and has minimal impact on the phototoxicity and photobleaching effect of biological samples.

[0038] (2) The beam splitting and collimation module consists of a beam splitting element 3 and an adjustable collimation component 4. Specifically, it includes a diffraction beam splitter and a multi-beam collimating lens arranged sequentially along the optical path. The beam splitting element 3 (diffraction beam splitter) splits the incident collimated laser into two or more beams of equal energy. The adjustable collimation component 4 (multi-beam collimating lens) adjusts the beam propagation direction to a parallel beam through a focusing lens group or a steering prism group, ensuring that the beams are tilted onto the sample surface at the same incident angle. Any method of splitting a laser beam into two parallel laser beams is part of the protection of this invention. It should be emphasized that the two beams output by the beam splitting element 3 may have a fixed angle. The function of the adjustable collimation component 4 is to guide the two beams to exit in parallel, while realizing flexible adjustment of the parallel beam spacing.

[0039] (3) A polarization beam splitter module, including a polarization beam splitter 5 (PBS) and a quarter-wave plate 6, is used to achieve spatial separation of the focusing laser emission path and the receiving path. The laser light is converted into circularly polarized light by the PBS and the quarter-wave plate 6. The circularly polarized light is combined with the imaging path by a dichroic mirror 9, and then illuminates the sample surface through the microscope objective 10. After being reflected by the sample, its polarization direction changes, and after passing through the quarter-wave plate 6 again, its polarization direction is orthogonal to the incident light. This orthogonally polarized light is reflected by the PBS and thus received by the subsequent laser detection imaging module, without being transmitted back to the infrared laser 1.

[0040] (4) Laser detection and imaging module, including a high-sensitivity CMOS area array detector (CMOS focusing spot detector 8) and a bandpass filter 7, is used to receive and image the signal light reflected by the sample, and extract the centroid coordinates of the spot through image processing to provide raw data for subsequent pose calculation.

[0041] (5) Focusing control module, used for decision-making and execution of the automatic focusing system, receives the centroid coordinates output by the laser detection imaging module, calculates the current three-dimensional pose to be adjusted of the sample, drives the leveling motion stage 12 to perform leveling and focusing actions, and forms a closed-loop control. This module consists of two parts: control algorithm software and hardware interface.

[0042] The system selects an infrared laser 1 in the infrared band as the focusing light source. This infrared wavelength is significantly separated from the fluorescence excitation band (470nm, 532nm) and the fluorescence imaging band, and can be efficiently combined with the imaging optical path by a dichroic mirror 9 without crosstalk. After the laser is coupled out through a single-mode fiber, a collimated beam is obtained through a fiber collimator, and its divergence angle is controlled within 1mrad.

[0043] The collimated beam first enters the beam-splitting collimation module, whose core component is a diffractive optical element (DOE) or a beam-splitting prism 14. The DOE splits the single collimated laser beam into diffracted beams of equal energy with an angle of less than 10 degrees between each pair of beams. The multiple beams are then converted into parallel light by the adjustable collimation assembly 4 (focusing lens group or steering prism group, i.e., focusing lens group), and the center distance can be adjusted within the range of 1 mm to 8 mm according to the aperture of the microscope objective 10. The parallel light is symmetrically distributed within the entrance pupil plane of the microscope objective 10.

[0044] The parallel light then enters the polarization beam splitter. The beam first enters the PBS, which allows the polarized laser to be projected into the quarter-wave plate 6 and the dichroic mirror 9. The fast axis of the quarter-wave plate 6 forms a 45° angle with the polarization direction of the incident light, and the linearly polarized light becomes circularly polarized light after passing through the quarter-wave plate 6 once. The dichroic mirror 9 is designed to have high reflectivity (>95%) for infrared lasers and high transmittance (>90%) for fluorescence excitation and fluorescence imaging bands. Therefore, the circularly polarized infrared laser is reflected by the dichroic mirror 9, deflected, and enters the rear pupil plane of the microscope objective 10. The microscope objective 10 focuses the two parallel beams onto the sample surface. Since the two beams are incident at a certain distance and tilt angle, two independent illumination spots are formed on the sample surface. The positional distribution of the two spots is related to the distance between the two beams and the focal length of the microscope objective 10, with a typical distance ranging from tens to hundreds of micrometers, depending on the magnification of the microscope objective 10 and the setting of the distance between the two beams.

[0045] After passing through the sample surface, the reflected light changes direction, carrying sample surface information—including the sample's height (Z-axis) and surface tilt angle (rotation around the X and Y axes)—back along the original optical path. The reflected light then passes through the microscope objective 10 and the dichroic mirror 9 before entering the quarter-wave plate 6. At this point, the direction-changed reflected light is converted into linearly polarized light orthogonal to the polarization state of the incident laser. This orthogonal polarization conversion is crucial for achieving physical separation of the transmission and reception paths.

[0046] The returned light, after being filtered by the bandpass filter 7, is focused onto the CMOS focusing spot detector 8 to form dual infrared spots. The relative positions of these spots have a definite mapping relationship with the spatial pose of the sample. By extracting the centroid coordinates of the two infrared spots in real time and calculating their spacing change and average position change, the Z-axis defocusing amount and tilt angle of the sample can be solved, realizing real-time monitoring of the three-dimensional pose to be adjusted.

[0047] In some embodiments, the process of obtaining the fitting function is as follows: Select a flat, uniformly reflective standard sample with a standard pattern, and place the standard sample on the leveling motion table 12. The fluorescence images acquired by the fluorescence imaging module are input into the sharpness evaluation function to obtain the normalized gradient sum; If the sum of normalized gradients is less than the preset threshold, the leveling stage 12 is driven to move along the Z-axis, so that the fluorescence imaging module re-acquires the fluorescence image and compares the sum of normalized gradients of the current fluorescence image with the preset threshold until the sum of normalized gradients is greater than the preset threshold, thus obtaining the optimal focal plane position of the microscope objective 10; otherwise, the optimal focal plane position of the microscope objective 10 is obtained. The current position of the microscope objective 10 on the Z-axis is taken as the zero reference position. The centroid coordinates of each of the current dual infrared spots are taken as the calibration centroid coordinates. The reference spacing and reference centroid coordinates are calculated based on the calibration centroid coordinates of each of the dual infrared spots. The leveling stage 12 is driven to move along the Z-axis, and the differences between the microscope objective 10 at different positions on the Z-axis and the zero-point reference position, as well as the spacing between the two infrared spots corresponding to the microscope objective 10 at different positions on the Z-axis, are collected. The differences between the microscope objective 10 at different positions on the Z-axis and the zero-point reference position, as well as the spacing between the two infrared spots corresponding to the two positions on the Z-axis, are fitted to obtain the fitting function.

[0048] The preset threshold is 0.95. The sharpness evaluation function uses the image edge gradient method, where the sum of the edge gradients of the microscopic image is maximized at the ideal focus position; when the sample is out of focus, the image is blurred, and the sum of the gradients decreases. Specifically, the Sobel operator is used to calculate the gradient values ​​in the horizontal and vertical directions of the image, and the gradients in these two directions are added together as a metric and converted into a normalized gradient. The preset threshold for the sum of the normalized gradients is set to 0.95. When the actual sum is greater than this threshold, the current position is considered to have reached the sharp imaging range; otherwise, it is in a defocused state.

[0049] In some embodiments, the three-dimensional pose of the sample 11 to be adjusted is calculated. , , The specific process is as follows: The measured centroid coordinates of the first infrared spot are ( X 1, Y 1), The measured centroid coordinates of the second infrared spot are ( X 2, Y 2), then the measured spacing between the current dual infrared spots D for: ; Based on the fitting function, the difference between the current position of the microscope objective 10 on the Z-axis and the zero-point reference position is calculated. ; Calculate the X-direction offset of the sample 11 to be tested. and Y-direction offset :

[0050] in,( , () is the reference centroid coordinate, , () represents the measured coordinates of the centroid.

[0051] In some embodiments, the beam splitting and collimation module includes a diffraction beam splitter and a multi-beam collimating lens arranged sequentially along the optical path. The diffraction beam splitter is used to split a single-path incident infrared laser beam into two independent sub-beams, and the multi-beam collimating lens is used to collimate and shape the two independent sub-beams after beam splitting, so that each independent sub-beam becomes parallel light.

[0052] In some embodiments, the polarization beam splitting module includes a polarization beam splitter 5 and a quarter-wave plate 6 arranged sequentially along the optical path, which separate the two parallel beams emitted from the beam splitting collimation module into P-polarized light and S-polarized light, and the quarter-wave plate 6 converts the P-polarized light from linearly polarized light into circularly polarized light.

[0053] In some embodiments, the laser detection imaging module includes a bandpass filter and a CMOS focusing spot detector 8 arranged sequentially along the optical path. The infrared laser reflected from the sample surface is filtered by the bandpass filter after passing through the microscope objective 10, the dichroic mirror 9 and the polarization beam splitter to remove background noise, and then incident on the CMOS focusing spot detector 8 to obtain an infrared image.

[0054] The key to this invention lies in: (1) The adjustable center distance beam-splitting structure is compatible with a variety of NA microscope objectives 10; (2) The focusing optical path and the fluorescence imaging optical path are designed to share a common optical path through the same microscope objective 10; (3) The emission path of the focusing laser and the echo detection path are physically decoupled through polarization separation.

[0055] This structure not only reduces the complexity of the optical system, but also effectively reduces crosstalk and improves system stability and repeatability.

[0056] To achieve high-precision axial focusing measurement, this invention utilizes the off-axis imaging characteristics of the optical system to convert the minute displacement of the sample along the Z-axis direction and the dihedral angle of the sample surface into the lateral displacement of the light spot on the detection plane, thereby achieving high-sensitivity detection of focusing errors.

[0057] This invention proposes a leveling method based on dual-spot common-difference mode decoupling, which is implemented using a leveling system based on dual-spot common-difference mode decoupling, and specifically includes the following steps: S1: The centroid coordinates of the two infrared spots are calibrated using standard samples, and the difference between the microscope objective 10 at different Z-axis positions and the zero-point reference position, as well as the spacing between the two infrared spots at different Z-axis positions, are fitted to obtain the fitting function. S2: Place the sample 11 to be tested on the leveling motion stage 12, and use the laser detection imaging module to image the dual infrared light spots on the surface of the sample 11 to obtain the measured distance between the current dual infrared light spots. S3: Input the measured spacing of the current dual infrared spots into the fitting function for calculation, obtain the difference between the current position of the microscope objective 10 on the Z axis and the zero reference position, and calculate the X-direction offset and Y-direction offset of the sample 11 to obtain the three-dimensional pose of the sample 11 to be adjusted. S4: If the three-dimensional pose to be adjusted of the sample 11 is not 0, then drive the microscope objective 10 and the leveling stage 12 to move based on the three-dimensional pose to be adjusted of the sample 11. Use the laser detection imaging module to re-acquire the dual infrared spot imaging on the surface of the sample 11. Use the newly acquired measured spacing of the dual infrared spots to replace the current measured spacing of the dual infrared spots and execute step S3. Otherwise, directly complete the parallelism leveling of the plane to be tested of the sample 11 and the imaging focal plane of the microscope objective 10.

[0058] In some embodiments, step S1 specifically includes the following steps: S11: Select a flat, uniformly reflective standard sample with a standard pattern and place the standard sample on the leveling motion table 12. S12: Input the fluorescence image acquired by the fluorescence imaging module into the sharpness evaluation function to obtain the normalized gradient sum; S13: If the sum of normalized gradients is less than the preset threshold, drive the leveling stage 12 to move along the Z-axis so that the fluorescence imaging module can re-acquire the fluorescence image and compare the sum of normalized gradients of the current fluorescence image with the preset threshold until the sum of normalized gradients is greater than the preset threshold to obtain the optimal focal plane position of the microscope objective 10; otherwise, obtain the optimal focal plane position of the microscope objective 10. S14: Take the current position of the microscope objective 10 on the Z-axis as the zero reference position, take the centroid coordinates of each of the current dual infrared spots as the calibration centroid coordinates, and calculate the reference spacing and reference centroid coordinates based on the calibration centroid coordinates of each of the dual infrared spots. S15: Drive the leveling stage 12 to move along the Z-axis, collect the difference between the microscope objective 10 at different positions on the Z-axis and the zero-point reference position, as well as the spacing between the two infrared spots corresponding to the microscope objective 10 at different positions on the Z-axis, and fit the difference between the microscope objective 10 at different positions on the Z-axis and the spacing between the two infrared spots corresponding to the two positions on the Z-axis to obtain the fitting function.

[0059] In some embodiments, in step S12, the sharpness evaluation function may employ the image edge gradient method, using the Sobel operator to calculate the gradient values ​​in the horizontal and vertical directions of the image respectively, and convert them into normalized gradients, and determine the focus position through threshold determination.

[0060] In some embodiments, in step S3, the three-dimensional pose to be adjusted of the sample 11 to be tested is calculated. , , Specifically, it includes: S31: The measured centroid coordinates of the first infrared spot are ( X 1, Y 1), The measured centroid coordinates of the second infrared spot are ( X 2, Y 2), then the measured spacing between the current dual infrared spots D for: ; S32: Input the measured spacing of the current dual infrared spots into the fitting function to calculate the difference between the current position of the microscope objective 10 on the Z-axis and the zero-point reference position. ; S33: Calculate the X-direction offset of the sample 11 to be tested. and Y-direction offset :

[0061] in,( , () is the reference centroid coordinate, , () represents the measured coordinates of the centroid.

[0062] Specifically, after the focusing laser is focused onto the sample surface by the microscope objective 10, the reflected light from the sample surface returns along the original optical path. Under ideal focusing conditions, the propagation path of the reflected beam is symmetrical with the system's optical axis, and its imaging position on the CMOS focusing spot detector 8 is at a preset reference position. When the sample defocuses or tilts along the Z-axis, the beam's incident point shifts, causing a lateral shift in its imaging position on the CMOS focusing spot detector 8.

[0063] The distance between the two infrared spots corresponding to different positions of the microscope objective 10 on the Z-axis, i.e., the lateral offset Δx, and the difference between the microscope objective 10 at different positions on the Z-axis and the zero-point reference position, i.e., the sample axial displacement Δz, have a monotonic functional relationship. The specific functional form is determined by the optical system parameters, including the focal length f of the microscope objective 10, the system magnification M, and the initial offset angle of the incident beam.

[0064] By establishing an optical geometric model, the Z-axis displacement of the sample is mapped to the lateral displacement of the light spot on the detector, thus achieving indirect measurement of axial displacement.

[0065] By calibrating the system, a one-to-one mapping relationship between Δx and Δz can be established, thereby enabling the axial displacement to be deduced by measuring the lateral offset.

[0066] In practical implementation, the lateral offset is detected by an image sensor. Preferably, a CMOS focusing spot detector 8 is used to acquire the spot, and the centroid position of the spot is accurately extracted by combining a sub-pixel-level positioning algorithm (such as centroid method, Gaussian fitting method, or correlation matching method). By weighting the gray-scale distribution of the spot, a positioning accuracy better than that of a single pixel can be achieved. Taking an image sensor with a pixel size of approximately 3μm as an example, under the condition of 1 / 10 pixel positioning accuracy, a position resolution on the order of approximately 300nm can be achieved. Combined with the system sensitivity amplification effect, an axial focusing accuracy better than 200nm can be further achieved.

[0067] Furthermore, to improve measurement stability, this invention can preprocess the light spot, including background noise suppression, bandpass filtering, and threshold segmentation, to improve the signal-to-noise ratio. Simultaneously, the influence of random noise can be reduced through multi-frame sampling and time averaging, thereby improving measurement repeatability.

[0068] During the calibration phase, the sample can be driven to move axially using a known Z-axis displacement stepping device (such as a piezoelectric displacement stage). The lateral positions of the corresponding light spots are recorded at multiple known displacement points, and a mapping model is then established through linear or nonlinear fitting. In some implementations, this mapping relationship can be approximated as a linear relationship, thereby simplifying the real-time calculation process. In high-precision scenarios, polynomial fitting or lookup table methods can be used to achieve higher precision mapping.

[0069] Compared to traditional focusing methods based on image sharpness evaluation functions (such as gradient and contrast), this invention establishes a geometric mapping relationship between axial displacement and lateral offset of the sample, transforming the focusing signal from "blur evaluation" to "position measurement." This makes the focusing process independent of image content, significantly improving the system's adaptability to different sample types. Furthermore, this method is insensitive to changes in ambient light and fluctuations in fluorescence intensity, exhibiting higher robustness.

[0070] Furthermore, this method is particularly suitable for high numerical aperture (NA≥0.9) microscopy systems. In such systems, the depth of focus is typically only on the order of hundreds of nanometers, resulting in extremely small focusing tolerances, making stable control difficult with traditional methods. This invention, however, introduces a lateral offset amplification mechanism to convert nanometer-level Z-displacement into micrometer-level lateral displacement, thereby significantly improving measurement sensitivity and enabling the system to achieve fine resolution within the depth of focus range.

[0071] In summary, this invention achieves highly sensitive and stable measurement of axial displacement by constructing a clear physical mapping model and a high-precision image positioning algorithm, providing a reliable measurement basis for subsequent automatic focusing and leveling control.

[0072] Based on the above-mentioned focusing measurement method, the present invention can achieve precise focusing of the sample in the Z-axis direction and tilt correction in the XY plane direction, thereby ensuring the alignment accuracy of the overall system.

[0073] This method employs a phased closed-loop control strategy, decoupling the focusing process from the leveling process and introducing a state determination mechanism to achieve rapid convergence and stable control of the system.

[0074] Compared with the prior art, the multi-stage leveling control method of the present invention has the following advantages: First, by performing focusing and leveling in stages and introducing a clear state determination mechanism, the control instability problem caused by multivariable coupling is avoided, and the system convergence speed and robustness are improved. Secondly, by introducing a real-time solution algorithm for multi-spot common-mode and differential-mode signals, the system can achieve parallel sensing and coordinated adjustment between Z-axis focusing and XY-plane leveling, keeping the control process continuous and smooth. Finally, this method makes full use of the common-mode and differential-mode information provided by the multi-beam probe spot, and can acquire the defocus and tilt amounts simultaneously in a single image acquisition without the need for an additional independent spatial measurement module, which significantly reduces system complexity and hardware cost.

[0075] In summary, the multi-stage automatic leveling control method of the present invention, through process design and closed-loop control strategy, achieves high-precision and high-stability automatic focusing and leveling functions, providing reliable technical support for high-precision chip alignment and microscopic operations.

[0076] The overall optical path of this invention consists of an infrared optical path on the left and a fluorescence optical path on the right. The infrared optical path uses a 1×2 diffraction beam splitter to split the infrared laser into two beams. The infrared optical path and the fluorescence optical path are coupled through a dichroic mirror 9.

[0077] The infrared optical path includes an infrared light source module, a beam splitting and collimation module, a polarization beam splitting module, and a laser detection and imaging module. The collimated laser emitted by the infrared light source module first enters a 1×2 diffraction beam splitter, which forms a beam with a certain angle. After passing through a multi-beam collimating lens, it enters the polarization beam splitter 5.

[0078] Polarizing beam splitter 5 separates the incident light according to its polarization state and guides the beam into the main optical path. A quarter-wave plate 6 is placed after polarizing beam splitter 5 to convert linearly polarized light into circularly polarized light, causing the beam's polarization state to change after reflection. This allows the beam to be guided by polarizing beam splitter 5 into the laser detection imaging module in the return path, achieving separation of the emitted and detected optical paths. Figure 2 As shown.

[0079] The polarization-modulated light beam continues to propagate to the dichroic mirror 9. The dichroic mirror 9 is used to achieve beam splitting control of light of different wavelengths. The infrared laser used for focusing is reflected into the microscope objective 10, while the fluorescence excitation light and imaging light can pass through the dichroic mirror 9.

[0080] The infrared laser, reflected by the dichroic mirror 9, is focused onto the sample surface by the microscope objective 10. The sample reflects the incident light, which returns along the original optical path. After passing through the microscope objective 10 and the dichroic mirror 9, it reaches the quarter-wave plate 6. Due to the change in polarization direction, the polarization direction of the probe light passing through the quarter-wave plate 6 is orthogonal to the original infrared laser. Therefore, it is guided to the detection path at the polarization beam splitter 5 and finally received by the image sensor to form an infrared image with dual infrared spots.

[0081] A bandpass filter assembly, or bandpass filter, is placed in front of the image sensor to filter out non-target wavelength light and improve the signal-to-noise ratio of the infrared laser used for focusing.

[0082] The fluorescence optical path includes a fluorescence excitation source 17, a fluorescence laser collimator 16, a beam splitter 14, a tube mirror 13, and a CMOS imaging detector 15. The excitation light emitted by the fluorescence excitation source 17 (e.g., in the 470nm or 532nm band) is coupled into the system through the optical path, transmitted through the dichroic mirror 9, and enters the microscope objective 10, irradiating the sample and thereby exciting the sample to generate a fluorescence signal.

[0083] The fluorescence signal generated by the sample is collected by the microscope objective 10 and returned. It is then transmitted through the dichroic mirror 9 into the imaging optical path and finally received by the CMOS imaging detector 15, thus realizing the acquisition of the fluorescence image.

[0084] Through the above design, the infrared light path and the fluorescence light path achieve a common light path at the microscope objective 10, thereby ensuring the consistency between the focusing position and the imaging position.

[0085] Before focusing and adjusting, the system needs to be calibrated to determine the reference zero point under standard operating conditions. The standard state is defined as follows: the sample is in a clear image and the sample surface is not effectively tilted relative to the focal plane of the microscope objective.

[0086] The specific calibration steps are as follows: Prepare a flat, reflective standard sample with standard patterns or markings, and place it on the leveling motion table 12.

[0087] The actuator performs a coarse scan, while the sharpness of the standard image is observed through the CMOS focusing spot detector 8. The optimal focal plane position is found using a sharpness evaluation function (such as image variance, Brenner gradient, etc.) and used as the zero-point reference position.

[0088] Under the aforementioned standard conditions, an infrared image containing two infrared spots is acquired, the centroid coordinates of the two infrared spots are calculated, and the reference spacing is recorded. (Euclidean distance between the two light spots) and reference centroid (Average coordinates of the centroids of the two light spots). These reference values ​​will be used for error calculation in subsequent closed-loop control.

[0089] Once calibrated, the reference value can be directly used in daily work without recalibrating each time, unless the objective lens, sample type, or drift exceeding the limit is detected after the system has been running for a long time.

[0090] like Figure 3 (a) and Figure 3 As shown in (b), when the sample undergoes defocusing displacement along the Z-axis, the distance between the two infrared spots undergoes linear scaling (differential mode displacement, i.e., "breathing scaling effect"); as Figure 4 (a) and Figure 4 As shown in (b), when the sample surface is tilted, the two infrared spots shift in the same direction (common-mode displacement). By separating the common mode and the differential mode, the Z-axis defocusing and tilting of the sample can be calculated independently.

[0091] During focus detection and adjustment, the system acquires infrared images containing two infrared spots in real time, extracts the centroid coordinates of the two infrared spots, and calculates the current measured distance. and current focus .

[0092] like Figure 5 As shown in (a), the Z-axis defocusing amount: ,when When the sample is close to the microscope objective 10, the sample is moved away from the microscope objective 10; like Figure 5 As shown in (b), the tilt vector of the XY plane: The direction of this vector represents the tilt direction of the sample normal, for example, when When parallel to the unit vector (0,1), it indicates that the sample rotates along the x-direction and the normal deflects in the positive y-axis direction.

[0093] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0094] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A leveling system based on dual-spot common-differential mode decoupling, characterized in that: It includes an infrared light source module, a beam splitting and collimation module, a polarization beam splitting module, a laser detection and imaging module, a fluorescence light source module, a fluorescence detection module, a fluorescence imaging module, a dichroic mirror, a microscope objective, a sample to be tested, a leveling stage, and a focusing control module, among which: The sample to be tested is placed on a leveling stage. Infrared laser light emitted from the infrared light source module is sequentially incident on the sample surface through a beam-splitting collimation module, a polarization beam-splitting module, a dichroic mirror, and a microscope objective, forming a double infrared spot. The infrared laser light reflected from the sample surface is incident on the laser detection imaging module through the microscope objective, a dichroic mirror, and the polarization beam-splitting module to obtain an infrared image. Similarly, fluorescence emitted from the fluorescence light source module is incident on the sample surface through a fluorescence detection module, a dichroic mirror, and a microscope objective. The fluorescence light reflected from the sample surface is incident on the fluorescence imaging module through the microscope objective, a dichroic mirror, and the fluorescence detection module to obtain a fluorescence image. The focusing control module acquires the measured centroid coordinates of the two infrared spots contained in the infrared image, and calculates the three-dimensional pose of the sample under test by combining the calibrated centroid coordinates of the two infrared spots. Based on the calculation result of the three-dimensional pose of the sample under test, the module drives the microscope objective and the leveling stage to move, thereby achieving parallelism leveling between the plane of the sample under test and the imaging focal plane of the microscope objective.

2. The leveling system based on dual-spot common-difference mode decoupling according to claim 1, characterized in that: The process of obtaining the calibration centroid coordinates of each of the two infrared spots is as follows: Select a flat, uniformly reflective standard sample with a standard pattern, and place the standard sample on the leveling motion table. The fluorescence images acquired by the fluorescence imaging module are input into the sharpness evaluation function to obtain the normalized gradient sum; If the sum of the normalized gradients is less than the preset threshold, the leveling stage is driven to move along the Z-axis, so that the fluorescence imaging module re-acquires the fluorescence image and compares the sum of the normalized gradients of the current fluorescence image with the preset threshold until the sum of the normalized gradients is greater than the preset threshold, thus obtaining the optimal focal plane position of the microscope objective; otherwise, the optimal focal plane position of the microscope objective is obtained directly. The current position of the microscope objective on the Z-axis is taken as the zero reference position. The centroid coordinates of each of the current dual infrared spots are taken as the calibration centroid coordinates. The reference spacing and reference centroid coordinates are calculated based on the calibration centroid coordinates of each of the dual infrared spots. The leveling stage is driven to move along the Z-axis, and the differences between the microscope objective at different positions on the Z-axis and the zero-point reference position are collected, as well as the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis. The differences between the microscope objective at different positions on the Z-axis and the zero-point reference position, as well as the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis, are fitted to obtain the fitting function.

3. The leveling system based on dual-spot common-difference mode decoupling according to claim 2, characterized in that: Calculate the three-dimensional pose of the sample to be adjusted. , , The specific process is as follows: The measured centroid coordinates of the first infrared spot are ( X 1, Y 1), the measured centroid coordinates of the second infrared spot are ( X 2, Y 2), then the measured spacing between the current dual infrared spots D for: ； Based on the fitting function, the difference between the current position of the microscope objective on the Z-axis and the zero-point reference position is calculated. ; Calculate the X-direction offset of the sample under test. and Y-direction offset : in,( , () is the reference centroid coordinate, ( , () represents the measured coordinates of the centroid.

4. The leveling system based on dual-spot common-difference mode decoupling according to claim 1, characterized in that: The beam splitting and collimation module includes a diffraction beam splitter and a multi-beam collimating lens arranged sequentially along the optical path. The diffraction beam splitter is used to split the single-path incident infrared laser into two independent sub-beams, and the multi-beam collimating lens is used to collimate and shape the two independent sub-beams after beam splitting, so that each independent sub-beam becomes parallel light.

5. The leveling system based on dual-spot common-difference mode decoupling according to claim 1, characterized in that: The polarization beam splitter module includes a polarization beam splitter and a quarter-wave plate arranged sequentially along the optical path. It separates the two parallel beams emitted from the beam splitter collimator module into P-polarized light and S-polarized light. The quarter-wave plate converts the P-polarized light from linearly polarized light into circularly polarized light.

6. The leveling system based on dual-spot common-difference mode decoupling according to claim 1, characterized in that: The laser detection imaging module includes a bandpass filter and a CMOS focusing spot detector arranged sequentially along the optical path. The infrared laser reflected from the sample surface is filtered by the bandpass filter after passing through a microscope objective, a dichroic mirror and a polarization beam splitter to remove background noise, and then incident on the CMOS focusing spot detector to obtain an infrared image.

7. A leveling method based on dual-spot common-differential mode decoupling, implemented using the leveling system based on dual-spot common-differential mode decoupling as described in any one of claims 1-6, characterized in that: Specifically, the steps include the following: S1: The centroid coordinates of the two infrared spots are calibrated using standard samples, and the difference between the microscope objective at different Z-axis positions and the zero-point reference position, as well as the spacing between the two infrared spots at different Z-axis positions, are fitted to obtain the fitting function. S2: Place the sample to be tested on the leveling motion stage, and use the laser detection imaging module to image the dual infrared light spots on the surface of the sample to obtain the measured distance between the current dual infrared light spots. S3: Input the measured spacing of the current dual infrared spots into the fitting function for calculation, obtain the difference between the current position of the microscope objective on the Z axis and the zero reference position, and calculate the X-direction offset and Y-direction offset of the sample to be tested, and obtain the three-dimensional pose of the sample to be tested. S4: If the three-dimensional pose of the sample to be tested is not zero, the microscope objective and the leveling stage are moved based on the three-dimensional pose of the sample to be tested. The laser detection imaging module is used to re-acquire the dual infrared spot image of the sample surface. The measured spacing of the newly acquired dual infrared spot is used to replace the measured spacing of the current dual infrared spot. Step S3 is executed. Otherwise, the parallelism leveling of the test plane of the sample to be tested and the imaging focal plane of the microscope objective is directly completed.

8. The leveling method based on dual-spot common-difference mode decoupling according to claim 7, characterized in that: Step S1 specifically includes the following steps: S11: Select a flat, uniformly reflective standard sample with a standard pattern, and place the standard sample on the leveling motion table. S12: Input the fluorescence image acquired by the fluorescence imaging module into the sharpness evaluation function to obtain the normalized gradient sum; S13: If the sum of normalized gradients is less than the preset threshold, drive the leveling stage to move along the Z-axis, so that the fluorescence imaging module re-acquires the fluorescence image and compares the sum of normalized gradients of the current fluorescence image with the preset threshold until the sum of normalized gradients is greater than the preset threshold, and obtain the optimal focal plane position of the microscope objective; otherwise, directly obtain the optimal focal plane position of the microscope objective. S14: Take the current position of the microscope objective on the Z-axis as the zero reference position, take the centroid coordinates of each of the current dual infrared spots as the calibration centroid coordinates, and calculate the reference spacing and reference centroid coordinates based on the calibration centroid coordinates of each of the dual infrared spots. S15: Drive the leveling stage to move along the Z-axis, collect the difference between the microscope objective at different positions on the Z-axis and the zero reference position, as well as the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis, and fit the difference between the microscope objective at different positions on the Z-axis and the spacing between the two infrared spots corresponding to the microscope objective at different positions on the Z-axis to obtain the fitting function.

9. The leveling method based on dual-spot common-difference mode decoupling according to claim 8, characterized in that: In step S12, the sharpness evaluation function adopts the image edge gradient method.

10. The leveling method based on dual-spot common-difference mode decoupling according to claim 7, characterized in that: In step S3, the three-dimensional pose of the sample to be tested is calculated. , , Specifically, it includes: S31: The measured centroid coordinates of the first infrared spot are ( X 1, Y 1), the measured centroid coordinates of the second infrared spot are ( X 2, Y 2), then the measured spacing between the current dual infrared spots D for: ； S32: Input the measured spacing of the current dual infrared spots into the fitting function to calculate the difference between the current position of the microscope objective on the Z-axis and the zero-point reference position. ; S33: Calculate the X-direction offset of the sample to be tested. and Y-direction offset : in,( , () is the reference centroid coordinate, ( , () represents the measured coordinates of the centroid.

Images