Multimodal Super-Resolution Quantitative Phase Microscopy Method

Abstract

The present invention discloses a multimodal super-resolution quantitative phase microscopy method. The present invention uses a programmable LED array as an illumination light source, employs an objective lens with a low numerical aperture to obtain a stack of collected light intensity maps in multiple modalities, and uses the calculated light intensity maps as intensity constraints to perform multiple iterative updates, thereby achieving super-resolution quantitative phase microscopy. By virtue of the low numerical aperture objective lens and the programmable LED array, the present invention reduces the requirement for high-coherence illumination of the imaging system, and uses the idea of iterative reconstruction to complete phase recovery, thereby achieving super-resolution quantitative phase microscopy, which has the advantages of high resolution, high signal-to-noise ratio, and large field of view.

Description

Technical Field

[0001] The present invention belongs to the fields of optical microscopic measurement and imaging technology, and specifically relates to a multimodal super-resolution quantitative phase microscopic imaging method. Background Art

[0002] Multimodal microscopes are emerging solutions in recent years that break through the hardware limitations of conventional microscopes in each modality. With the help of a programmable LED array, multimodal microscopes solve the problem that the optical structures of various imaging modes are different and difficult to integrate for imaging. Without adding special optical accessories such as an annular diaphragm in the optical path, integrated imaging of multiple imaging modes such as bright-field imaging, dark-field imaging, and differential phase contrast can be achieved in a single device. "Zuo Chao, Chen Qian, Sun Jiasong, etc. Multimode Microscopic Imaging Method Based on Programmable LED Array Illumination [P]. Jiangsu Province: CN105158887B, 2017-09-22". "Zhang Jialin, Zuo Chao, Sun Jiasong, etc. Multimode Microscopic Imaging System and Method Based on LED Array [P]. Jiangsu Province: CN104765138B, 2017-09-29". At the same time, it ensures the stability and simplicity of the optical path for multimodal imaging, greatly reduces the cost, and can flexibly achieve optical staining by changing the color of the imaging light, and the achievable imaging types are more flexible and rich.

[0003] Higher resolution has always been the goal pursued in the field of microscopic imaging. However, while improving the resolution, the spatial bandwidth of the microscope cannot increase with the increase in resolution. Currently, in order to overcome the contradiction that it is difficult to balance the resolution and the field of view size at the same time, it is usually based on a conventional microscopic imaging system, combined with high-precision mechanical scanning and post-processing spatial domain image stitching technology, to stitch and fuse multiple high-resolution images of small fields of view into a high-resolution image of a large field of view. "Sun Xizhao, Wang Zhen, Li Pan, etc. Device Suitable for Image Stitching of Acid-Fast Stained Images of Mycobacterium Tuberculosis [P]. Shanghai: CN203490416U, 2014-03-19". However, due to the introduction of a mechanical moving device, the imaging stability and imaging speed of the microscopic system have become irreconcilable contradictions.

[0004] In order to break through the contradiction that it is difficult to balance resolution and field of view size without introducing a mechanical moving device, a common method is to use computational imaging methods such as Fourier ptychographic microscopy, "Zheng G, Horstmeyer R, Yang C. Wide-field, high-resolution Fourier ptychographic microscopy. Nature photonics. 2013; 7: 739-45.". Based on the ideas of phase retrieval and synthetic aperture, this method iteratively updates the spectral information within the corresponding sub-apertures in the frequency domain using a series of low-resolution intensity maps captured by a camera. The sub-apertures are superimposed to expand the frequency domain bandwidth, enabling the recovery of high-frequency information beyond the spatial resolution limit of the objective lens, while reconstructing the high-resolution intensity and phase images of the object, thereby achieving large-field-of-view and high-resolution imaging. However, this method requires illuminating each LED light one by one, has high requirements for the spatial coherence of illumination, and due to the low signal-to-noise ratio of the intensity maps corresponding to single-light illumination, its robustness is poor. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-modal super-resolution quantitative phase microscopy imaging method, which can achieve super-resolution quantitative phase microscopy imaging on the basis of multi-modal imaging, and has higher imaging resolution and signal-to-noise ratio and more stable imaging without adding complex external optical structures.

[0006] The technical solution for achieving the purpose of the present invention is as follows: A multi-modal super-resolution quantitative phase microscopy imaging method, and the specific steps are as follows:

[0007] Step 1: Use a programmable LED array as the illumination light source, and at different defocus distances, collect original intensity maps in multiple modalities to establish a bright-field defocus intensity map stack and a dark-field defocus intensity map stack.

[0008] Step 2: Use the bright-field intensity map at the focal distance to initialize the complex amplitude distribution on the focal plane and the corresponding high-resolution complex amplitude distribution in the frequency domain.

[0009] Step 3: Solve illumination multiplexing according to the coherent mode decomposition theory, and calculate the complex amplitudes at different illumination angles and different defocus distances.

[0010] Step 4: Under bright-field and dark-field illuminations respectively, superimpose the intensities of mutually incoherent point light sources with the same defocus distance to obtain a bright-field defocus intensity map stack and a dark-field defocus intensity map stack, and compare them with the captured intensity map stack to constrain and update the complex amplitudes at different angles and different defocus distances.

[0011] Step 5: Based on the theory of coherent mode decomposition under partially coherent illumination, take the Fourier transform of the updated complex amplitude, multiply it by the aperture function to obtain the new sub-aperture spectrum, and synthesize to obtain a high-resolution Fourier frequency domain;

[0012] Step 6: Take the inverse Fourier transform of the Fourier spectrum and then take the modulus squared to obtain the reconstructed high-resolution intensity map;

[0013] Step 7: Repeat Steps 3 to 6, update the entire objective function in each iteration, and after completing the specified number of iterations, obtain the high-resolution reconstructed intensity map.

[0014] Preferably, the implementation process of Step 1 is as follows:

[0015] Place the LED array (1), the sample (2), the microscope objective (3), the tube lens (4), and the camera (5) on the same axis in sequence; among them, the sample (2) is placed on the front focal plane of the microscope objective (3), the tube lens (4) is placed behind the microscope objective (3), the distance from the microscope objective (3) to the tube lens (4) is the sum of their mechanical focal lengths, and the camera (5) is located on the rear focal plane of the tube lens (4);

[0016] Light up all the lamp beads 1 to j corresponding to the bright-field imaging pattern of the LED panel, and capture a stack of bright-field defocus intensity maps at different defocus distances where I cap represents the intensity map captured, b represents the LED illumination angle during bright-field imaging, z represents the serial number of the defocus intensity map. When z = 0, it corresponds to the focused intensity map. When z > 0, it corresponds to the positive defocus intensity map. A total of z num +1 intensity maps are collected, and the defocus distance step size for each group of collections is z step ;

[0017] Light up all the lamp beads j + 1 to N corresponding to the dark-field imaging pattern of the LED panel, where N is the total number of lamp beads on the LED panel, and capture a stack of dark-field defocus intensity maps at different defocus distances where d represents the LED illumination angle during dark-field imaging.

[0018] Preferably, the specific steps for initializing the complex amplitude distribution on the focused plane and the high-resolution complex amplitude distribution in the corresponding frequency domain using the bright-field intensity map at the focused distance are as follows:

[0019] Perform upsampling on the collected focused intensity map under the bright-field illumination mode to obtain the initialized intensity map I h ;

[0020] Determine the initial complex amplitude distribution U h of the sample from the intensity map, and at the same time obtain the high-resolution spectral complex amplitude distribution F h:

[0021]

[0022] F h = F{U h}

[0023] where F{·} represents taking the Fourier transform of the content within the parentheses.

[0024] Preferably, the specific steps of step three for solving illumination multiplexing according to the coherent mode decomposition theory and calculating the complex amplitude at different illumination angles and different defocus distances are as follows:

[0025] According to the coherent decomposition theory, multiply the obtained high-resolution spectral complex amplitude distribution F h by the aperture function P n to obtain the sub-aperture spectral information F sub matched with the illumination mode, and then perform the inverse Fourier transform on it to obtain the focused complex amplitude U n0 :

[0026] F sub = F h · P n

[0027] U n0 = F -1 {F sub}

[0028] where 1 ≤ n ≤ N, n is the illumination angle corresponding to the nth light-emitting diode (LED) in the LED array, P n is the aperture function in the spectrum corresponding to the nth LED unit, and F -1 {·} represents taking the inverse Fourier transform of the content within the parentheses;

[0029] Based on the angular spectrum propagation theory, use the focused complex amplitude U n0 to generate a series of complex amplitudes U nz corresponding to different defocus distances.

[0030] Preferably, step four is specifically as follows:

[0031] From the complex amplitudes U nz at different illumination angles and different defocus distances obtained in step three, take the square of the modulus to obtain the corresponding light intensity Then sum up all the light intensities at the corresponding illumination angles of the bright field and the dark field respectively, so as to obtain the stack of superposed bright-field defocus light intensity maps and the stack of superposed dark-field defocus light intensity maps

[0032]

[0033]

[0034] Among them, I cal represents the light intensity map obtained through calculation;

[0035] Stack the bright and dark field light intensity maps obtained by shooting and the light intensity map obtained by calculation and As conditions, update to obtain the complex amplitude at different illumination angles and different defocus distances Specifically

[0036]

[0037] Among them, U update represents the complex amplitude after iterative reconstruction;

[0038] Based on the angular spectrum propagation theory, the complex amplitudes at different defocus distances and different illumination angles are regenerated into the focused complex amplitudes at different illumination angles and different defocus distances

[0039] Preferably, the specific process of obtaining the high-resolution Fourier frequency domain in step five is as follows:

[0040] The updated Take the Fourier transform and multiply by the aperture function P consistent with the illumination distribution n , thereby obtaining the sub-aperture spectrum information matching the illumination pattern Synthesize the sub-aperture spectrum information at different illumination angles to obtain the high-resolution Fourier spectrum

[0041]

[0042]

[0043] Among them, F update represents the Fourier spectrum after iterative reconstruction;

[0044] Preferably, the specific process of obtaining the reconstructed high-resolution light intensity map by taking the inverse Fourier transform of the Fourier spectrum and then taking the modulus squared is as follows:

[0045] For the high-resolution Fourier spectrum obtained in step five Take the inverse Fourier transform to obtain the reconstructed complex amplitude distribution

[0046]

[0047] Square the modulus of the complex amplitude to obtain the reconstructed high-resolution intensity map.

[0048]

[0049] where I update represents the intensity map after iterative reconstruction;

[0050] Preferably, the implementation process of step seven is as follows:

[0051] The number of iterations is related to the number of defocused intensity maps selected. Generally, the more defocused intensity maps, the fewer the required iterations. For absorptive samples to be measured, the number of iterations is 40 - 100; for phase-type samples to be measured, the number of iterations is 50 - 100.

[0052] Compared with the prior art, the significant advantages of the present invention are: (1) The present invention can achieve large field of view, high resolution, and high throughput imaging. (2) The present invention is combined with a multimodal microscope and is compatible with a microscope illuminated by an LED. (3) The present invention is non-interferometric and can work well under partially coherent illumination, so the working environment requirements are not strict and the application fields are wider. (4) The present invention has a high signal-to-noise ratio, does not require turning on each LED light one by one, converts multi-angle illumination into multi-defocus illumination of illumination multiplexing, and has better robustness.

[0053] The present invention will be further described in detail below with reference to the accompanying drawings. Description of the Drawings

[0054] Figure 1 is a schematic diagram of an image taken by a multimodal super-resolution quantitative phase microscopy method. Figure 1 (a1) represents a stack of defocused intensity maps collected in bright-field illumination mode; Figure 1 (a2) represents a schematic diagram of the microscopy system in bright-field illumination mode; Figure 1 (a3) represents the LED illumination distribution in bright-field illumination mode (turn on all the LED beads in the central area, and the specific number of beads is related to the numerical aperture parameter of the objective lens for image acquisition and the distance from the LED to the sample); Figure 1 (b1) represents a stack of defocused intensity maps collected in dark-field illumination mode; Figure 1 (b2) represents a schematic diagram of the microscopy system in dark-field illumination mode; Figure 1 (b3) represents the LED illumination distribution in dark-field illumination mode (turn on all the LED beads in the non-central area, and the specific number of beads is related to the numerical aperture parameter of the objective lens for image acquisition and the distance from the LED to the sample).

[0055] Figure 2 is a flowchart of the multimodal super-resolution quantitative phase microscopy method.

[0056] Figure 3 It is a simulation result diagram of a multimodal super-resolution quantitative phase microscopy imaging method. Figure 3 In it, (a1) is the initial light intensity diagram; Figure 3 In it, (a2) is the initial frequency spectrum diagram; Figure 3 In it, (b1) is the sub-frequency spectrum under bright-field illumination; Figure 3 In it, (b2) is the focused light intensity diagram corresponding to the sub-frequency spectrum under bright-field illumination; Figure 3 In it, (b3) is the sub-frequency spectrum under dark-field illumination; Figure 3 In it, (b4) is the focused light intensity diagram corresponding to the sub-frequency spectrum under dark-field illumination; Figure 3 In it, (c1) is the superimposed bright-field focused light intensity diagram obtained by coherent mode decomposition; Figure 3 In it, (c2) is the stack of defocused light intensity diagrams collected under bright-field illumination; Figure 3 In it, (d1) is the superimposed dark-field focused light intensity diagram obtained by coherent mode decomposition; Figure 3 In it, (d2) is the stack of defocused light intensity diagrams collected under dark-field illumination; Figure 3 In it, (e1) is the high-resolution frequency spectrum diagram reconstructed by the first iteration; Figure 3 In it, (e2) is the high-resolution light intensity diagram reconstructed by the first iteration; Figure 3 In it, (f1) is the high-resolution light intensity diagram reconstructed by the final iteration;

[0057] Figure 3 In it, (f2) is the high-resolution frequency spectrum diagram reconstructed by the final iteration. Specific implementation manner

[0058] The present invention is based on a microscopy imaging system with an LED illumination array. For the microscopy imaging system: The LED surface array (1), the sample (2), the microscope objective (3), the tube lens (4), and the camera (5) are placed on the same axis. Among them, the LED surface array (1) is arranged in front of the sample (2), the sample (2) is placed on the front focal plane of the microscope objective (3), the tube lens (4) is placed behind the microscope objective (3), the distance from the microscope objective (3) to the tube lens (4) is the sum of their mechanical focal lengths, and the camera (5) is located on the rear focal plane of the tube lens (4). Among them, for the microscope objective (3), generally a low numerical aperture is adopted to ensure large-field-of-view imaging. For example, the microscope objective (3) can select an objective lens with a magnification of 4× and a numerical aperture of 0.1.

[0059] Combined with the process schematic Figure 2 , the multimodal super-resolution quantitative phase microscopy imaging method implemented by the present invention using the above device includes the following seven steps:

[0060] Step 1: Acquire the original images: Use a programmable LED array as the illumination source. At different defocus distances, acquire the original intensity maps in multiple modalities, and establish a stack of bright-field defocus intensity maps and a stack of dark-field defocus intensity maps

[0061] As Figure 1 (a3) shows, first light up all the LED beads corresponding to the bright-field imaging pattern on the LED panel (the LED bead numbers are 1 to j). At different defocus distances, take a set of stacks of bright-field defocus intensity maps When z = 0, it corresponds to the focused intensity map. When 0 < z, it corresponds to the forward defocus intensity map. A total of z num +1 intensity maps are acquired. The defocus distance step size for each set of acquisitions is z step . Then light up all the LED beads corresponding to the dark-field imaging pattern on the LED panel (the LED bead numbers are j + 1 to N). At different defocus distances, take a set of stacks of dark-field defocus intensity maps

[0062] The number of lit LED beads (j) in the bright-field area and the number of lit LED beads (N - j) in the dark-field area are related to the numerical aperture of the microscope objective lens (3) used. For example, as Figure 1 (a3) shows, when using a 9×9 LED array for illumination, for a microscope objective lens (3) with a magnification of 4× and a numerical aperture of 0.1, lighting the circular area of the central 5 LED beads corresponds to bright-field illumination. As Figure 1 (b3) shows, except for the central 5 LED beads not being lit, lighting the remaining beads corresponds to dark-field illumination. For the defocus intensity map, the defocus distance step size z step can generally be selected around 0.5 - 2 μm; and z num The specific value of is related to the total number of effective LED beads for LED illumination. Generally speaking, the value of z num is greater than or equal to the total number of beads.

[0063] Step 2: Upsample the acquired focused intensity map in the bright-field illumination mode to obtain the initial intensity map I h . Determine the initial complex amplitude distribution U h of the sample from the intensity map, and at the same time obtain the high-resolution spectral complex amplitude distribution F h in the corresponding frequency domain:

[0064]

[0065] F h = F{U h}

[0066] where F{·} represents taking the Fourier transform of the content in the parentheses; Fh As Figure 3 (a2) shown;

[0067] Step 3: According to the coherent mode decomposition theory, multiply the obtained high-resolution spectral complex amplitude distribution F h by the aperture function P n to obtain the sub-aperture spectral information F sub matching the illumination pattern, and then perform an inverse Fourier transform on it to obtain the focused complex amplitude U n0 at different illumination angles:

[0068] F sub = F h · P n

[0069] U n0 = F -1 {F sub}

[0070] where 1 ≤ n ≤ N, n is the illumination angle corresponding to the nth light bead in the LED array, P n is the aperture function in the spectrum corresponding to the nth LED unit, and F -1 {·} represents taking the inverse Fourier transform of the content in the brackets.

[0071] Based on the angular spectrum propagation theory, use the focused complex amplitude U n0 to generate a series of complex amplitudes U nz corresponding to different defocus distances. The corresponding spectra under bright-field sub-angle illumination are as Figure 3 (b1) shown, and the corresponding spectra under dark-field sub-angle illumination are as Figure 3 (b3) shown. Figure 3 (b2) is the stack of intensity maps obtained by squaring the corresponding complex amplitude U n under bright-field sub-angle illumination, Figure 3 (b4) is the stack of intensity maps obtained by squaring the corresponding complex amplitude U n under dark-field sub-angle illumination.

[0072] Step 4: From the complex amplitudes U nz at different illumination angles and different defocus distances obtained in Step 3, take the square of the modulus to obtain the corresponding intensity Then sum the intensities of all bright-field and dark-field corresponding illumination angles respectively to obtain the stacked bright-field defocus intensity map stack and the stacked dark-field defocus intensity map stack

[0073]

[0074]

[0075] Among them, I cal represents the light intensity map obtained by calculation;

[0076] Stack the bright and dark field light intensity maps obtained by shooting and the light intensity map calculated above and As conditions, update the complex amplitude at different illumination angles and different defocus distances

[0077]

[0078] Among them, U update represents the complex amplitude after iterative reconstruction;

[0079] Based on the angular spectrum propagation theory, the complex amplitudes at different defocus distances and different illumination angles are regenerated into the focused complex amplitudes at different illumination angles and different defocus distances

[0080] Step Five: After updating Take the Fourier transform and multiply by the aperture function P consistent with the illumination distribution n , thus obtaining the sub-aperture spectrum information matching the illumination pattern Synthesize the sub-aperture spectrum information at different illumination angles to obtain the high-resolution Fourier spectrum

[0081]

[0082]

[0083] Among them, F update represents the Fourier spectrum after iterative reconstruction; the high-resolution spectrum of the first iterative reconstruction Such as Figure 3 (e1) shown.

[0084] Step Six: For the high-resolution Fourier spectrum obtained in Step Five Take the inverse Fourier transform to obtain the reconstructed complex amplitude distribution

[0085]

[0086] Take the square of the modulus of the complex amplitude to obtain the reconstructed high-resolution light intensity map

[0087]

[0088] Among them, I updateIt represents the light intensity map after iterative reconstruction; the high-resolution light intensity map of the first iterative reconstruction Such as Figure 3 shown in (e2).

[0089] Step 7: Repeat Steps 3 to 6. By constraining the known light intensity map and the calculated light intensity values in the Fourier domain, iterative reconstruction of the high-resolution light intensity map is achieved. After completing a sufficient number of iterations, it converges to a more balanced globally consistent solution. Among them, the number of iterations is related to the number of defocused light intensity maps selected. Generally, the more defocused light intensity maps, the fewer iterations required. For absorption-type samples to be measured, the number of iterations is generally 40 - 100 times; for phase-type samples to be measured, the number of iterations is generally 50 - 100 times.

[0090] By leveraging the advantages of stable imaging optical path and flexible imaging modalities in multimodal imaging, the present invention proposes a new quantitative phase microscopy imaging method, breaking through the contradiction that it is difficult to balance resolution and field of view size, and finally realizing super-resolution quantitative phase microscopy imaging.

[0091] By comparing Figure 3 (a1) with Figure 3 (f1), it can be clearly seen that the finally reconstructed high-resolution light intensity map I H has more details than the directly captured bright-field light intensity map. For example, in the reconstruction result, the clothing texture of Cameraman and a clearer facial contour can be seen. It should be noted that this simulation result is based on a low numerical aperture microscope objective, so microscopic results with a large field of view can be obtained. And the problem of low resolution is solved by the high-resolution imaging using the multimodal super-resolution quantitative phase microscopy imaging method proposed by the present invention. Therefore, the present invention can achieve large-field-of-view and high-resolution imaging. Combined with a multimodal microscope and compatible with a microscope with LED illumination, it reduces the requirement for high-coherence illumination of the imaging system. Using the idea of iterative reconstruction, a new quantitative phase imaging method is proposed to complete phase recovery and realize super-resolution quantitative phase microscopy imaging.

Claims

1. A multimodal super-resolution quantitative phase microscopy imaging method, characterized in that The specific steps are as follows: Step 1: Use a programmable LED array as the illumination source. At different defocus distances, collect the original intensity maps in multiple modalities, and establish a stack of bright-field defocus intensity maps and a stack of dark-field defocus intensity maps. Step 2: Use the bright-field intensity map at the focal distance to initialize the complex amplitude distribution of the focal plane and the corresponding high-resolution complex amplitude distribution in the frequency domain. Step 3: According to the coherent mode decomposition theory to solve the illumination multiplexing, calculate the complex amplitudes at different illumination angles and different defocus distances. Step 4: At the same defocus distance, superimpose the intensities at the corresponding illumination angles of mutually incoherent points to obtain the bright-field intensity superposition map and the dark-field intensity superposition map at this defocus distance, thereby obtaining the stack of bright-field defocus intensity maps and the stack of dark-field defocus intensity maps, and compare with the captured intensity map stack to constrain and update the complex amplitudes at different angles and different defocus distances. Step 5: Based on the coherent mode decomposition theory under partial coherent illumination, take the Fourier transform of the updated complex amplitude and multiply it by the aperture function to obtain the new sub-aperture spectrum, and synthesize to obtain the high-resolution Fourier spectrum. Step 6: Take the inverse Fourier transform of the Fourier spectrum and then take the modulus squared to obtain the reconstructed high-resolution intensity map. Step 7: Repeat Steps 3 to 6, update the entire objective function in each iteration. After completing the specified number of iterations, obtain the high-resolution reconstructed intensity map.

2. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, wherein The implementation process of Step 1 is as follows: Place the LED array (1), the sample (2), the microscope objective (3), the tube lens (4), and the camera (5) on the same axis in sequence; among them, the sample (2) is placed on the front focal plane of the microscope objective (3), the tube lens (4) is placed behind the microscope objective (3), the distance from the microscope objective (3) to the tube lens (4) is the sum of their mechanical focal lengths, and the camera (5) is located on the rear focal plane of the tube lens (4). Light up all the light beads 1 to j corresponding to the bright-field imaging pattern of the LED panel, and capture a stack of bright-field defocused light intensity images at different defocus distances. Among them, I cap represents the captured light intensity image, b represents the LED illumination angle during bright-field imaging, z represents the serial number of the defocused light intensity image. When z = 0, it corresponds to the focused light intensity image, and when z > 0, it corresponds to the positive defocused light intensity image. A total of z num + 1 light intensity images are collected, and the defocus distance step size for each group of acquisitions is z step ; Light up all the light beads j + 1 to N corresponding to the dark-field imaging pattern of the LED panel, where N is the total number of light beads of the LED panel, and capture a stack of dark-field defocused light intensity images at different defocus distances. Among them, d represents the dark-field imaging LED illumination angle.

3. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, characterized in that The specific steps for initializing the complex amplitude distribution of the focal plane and the corresponding high-resolution complex amplitude distribution in the frequency domain using the bright-field intensity map at the focal distance are: The acquired acquisition focused light intensity map under bright field illumination mode is upsampled to obtain an initial light intensity map I h ; Determine the initial complex amplitude distribution \(U\) of the sample from the light intensity map h , and simultaneously obtain the high-resolution spectral complex amplitude distribution \(F\) in the corresponding frequency domain h : F h = F{U h} Among them, F{·} represents taking the Fourier transform of the content in the brackets.

4. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, wherein The specific steps for Step 3 to solve the illumination multiplexing according to the coherent mode decomposition theory and calculate the complex amplitudes at different illumination angles and different defocus distances are: According to the coherent mode decomposition theory, the obtained high-resolution spectral complex amplitude distribution F h is multiplied by the aperture function P n to obtain the sub-aperture spectral information F sub matched to the illumination mode. Then, an inverse Fourier transform is performed on it to obtain the focused complex amplitude U n0 at different illumination angles: F sub = F h · P n U n0 = F -1 {F sub} Among them, 1 ≤ n ≤ N, where n is the illumination angle corresponding to the n-th lamp bead in the LED array, and P n is the aperture function in the spectrum corresponding to the n-th LED unit, and F -1 {·} represents the inverse Fourier transform of the content within the brackets; Based on the angular spectrum propagation theory, using the focused complex amplitude U n0 Generate a series of complex amplitudes U corresponding to different defocus distances nz .

5. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, wherein Step 4 is specifically: The complex amplitude U obtained at different illumination angles and different defocus distances in step three nz , after taking the square of the modulus, the corresponding light intensity is obtained Then, the light intensities corresponding to all illumination angles in the bright field and dark field are summed respectively, so as to obtain the stacked bright-field defocus light intensity map and the stacked dark-field defocus light intensity map Among them, I cal represents the light intensity map obtained by calculation; Stack the captured bright and dark field intensity maps and the calculated intensity maps and As conditions, update the complex amplitude at different illumination angles and different defocus distances Specifically Among them, U update represents the complex amplitude after iterative reconstruction; Based on the angular spectrum propagation theory, the complex amplitudes at different defocus distances and different illumination angles are regenerated into the focused complex amplitudes at different illumination angles and different defocus distances 6. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, characterized in that The specific process for Step 5 to obtain the high-resolution Fourier frequency domain is as follows: Updated After taking the Fourier transform, multiply by the aperture function P that is consistent with the illumination distribution n , thereby obtaining the sub-aperture spectrum information that matches the illumination pattern By synthesizing the sub-aperture spectrum information at different illumination angles, a high-resolution Fourier spectrum can be obtained Among them, F update represents the Fourier spectrum after iterative reconstruction.

7. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, characterized in that The specific process for taking the inverse Fourier transform of the Fourier spectrum and then taking the modulus squared to obtain the reconstructed high-resolution intensity map is as follows: For the high-resolution Fourier spectrum obtained in Step Five Taking the inverse Fourier transform, the reconstructed complex amplitude distribution can be obtained Square the modulus of the complex amplitude to obtain the reconstructed high-resolution intensity map Among them, I update represents the light intensity map after iterative reconstruction.

8. The multimodal super-resolution quantitative phase microscopy imaging method according to claim 1, characterized in that, The implementation process of Step 7 is as follows: The number of iterations is related to the number of selected defocus intensity maps. Generally, the more defocus intensity maps, the fewer iterations required. For absorption-type samples to be measured, the number of iterations is 40 - 100 times; for phase-type samples to be measured, the number of iterations is 50 - 100 times.

Images