CN116183568B - High-fidelity reconstruction method and device for three-dimensional structured light illumination super-resolution microscopic imaging - Google Patents

Abstract

A method for high fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging, comprising: dividing a beam of parallel light into three beams of parallel light beams with equal intensity and consistent polarization directions, carrying out interference on a sample to form a three-dimensional non-uniform illumination light field, and carrying out frequency shift on a frequency spectrum of the sample after the sample is modulated by the non-uniform illumination light field; receiving fluorescent signals sent by a sample through an objective lens, converging the fluorescent signals to an imaging image surface through a field lens, and receiving the fluorescent signals through a detector to obtain a low-resolution image mixed with high-low frequency information of the sample; the space displacement and the direction of the illumination light field are changed for a plurality of times, and fluorescent signals modulated by the light field are shot again to obtain a series of low-resolution images mixed with high-low frequency information of the sample, and the low-resolution images are used as original images. And performing subsequent image processing on the original image, firstly performing parameter estimation on the initial phase and the spatial frequency of the illumination light field, then separating each frequency band of the sample, and finally combining each frequency spectrum to reconstruct a high-fidelity super-resolution image of the sample.

Description

A method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging and device

Technical field

The present invention relates to the field of optical super-resolution microscopy imaging, and specifically to a method and device for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging.

Background technique

Scientific microscopy has the advantages of small light damage, strong specificity, fast speed, large field of view, and the ability to perform three-dimensional imaging. It is the only technology in the existing microscopy technology that is possible to perform three-dimensional rapid imaging of living cells. However, due to the existence of the optical diffraction limit, the lateral resolution of conventional optical microscopy imaging is generally limited to the half-wavelength level, and the axial resolution is even lower, which can only reach about one-third of the lateral resolution, which cannot satisfy the needs of cells. The need for observation of living organelles at the sub-hundred nanometer scale. In the past two or three decades, various super-resolution imaging technologies have been proposed, including stimulated radiation depletion microscopy (STED), stochastic light reconstruction microscopy (PALM/STORM) and structured light illumination super-resolution microscopy. Technology (SIM), the introduction of these technologies has greatly promoted the development of life sciences.

Among many super-resolution microscopy imaging methods, SIM has many natural advantages for live cell imaging: for example, compared with STED, which achieves super-resolution through high-power loss light (approximately GW/cm ² ), SIM requires excitation light intensity It is lower (in the order of 10W/ ^cm2 ), so it has the inherent advantage of low photobleaching and photodamage; it does not need to acquire thousands of original images to reconstruct a single molecule like single molecule localization microscopy (SMLM). For super-resolution images, SIM requires only 9 frames (2D-SIM) or 15 frames (3D-SIM single layer) to reconstruct a super-resolution image, so it has the inherent advantage of fast imaging and is especially suitable for biological living cells. Imaging.

However, the acquisition of SIM super-resolution images also depends on the subsequent complex image reconstruction process. During the imaging process, SIM modulates the sample by illuminating the sample with structured light. Through the moiré effect, the high-frequency information of the sample that the system cannot detect originally is modulated into the low-pass band, thus improving the resolution. However, since the detected image is an aliasing of various frequency bands, in order to accurately separate the components of each frequency band, it is necessary to adjust the modulation degree a _m of the illumination light field, the spatial frequency k _xy , and the phase Make accurate estimates. The value of the modulation degree a _m affects the ratio of high-frequency components in the entire reconstructed spectrum, and ultimately only affects the contrast of the image, so the impact is relatively small; the spatial frequency p can be obtained by spectral component comparison and other methods, and the accuracy of the estimate can meet the needs ;When solving for each frequency band component, it is necessary to know the precise phase/> value, otherwise each spectrum cannot be fully realized when separated, and each frequency band will contain spectral components of other frequency bands. These spectral residuals will cause errors when they are subsequently shifted back to their own positions, leading to the appearance of artifacts. Traditional methods have estimation errors through spectrum peak positions or autocorrelation when performing phase estimation. Especially when the image contains noise, phase estimation errors will occur, thus affecting the quality of the final image reconstruction.

Contents of the invention

The present invention aims to overcome the shortcomings of the existing technology and provide a method and device for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopic imaging.

In view of the above problems, the present invention proposes a new phase estimation method in the image reconstruction process. By constructing an intermediate function, compared with the traditional phase estimation method, this method can accurately estimate the illumination light field and perform five-step phase shifting. The initial phase corresponding to the time; in addition, this method can also be used to eliminate the impact of inaccurate phase estimation due to the influence of noise during the phase calculation process. Using the precise phase calculation method proposed by the present invention, each frequency band component can be accurately separated, eliminating the influence of reconstruction artifacts caused by inaccurate phase estimation, and finally obtaining high-fidelity reconstruction of the sample.

The present invention provides a method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging, which includes the following steps:

(1) Divide a parallel light beam into three parallel light beams with equal intensity and consistent polarization direction, converge to the entrance pupil surface of the objective lens, and then become three parallel light beams after passing through the objective lens. The three beams of light interfere on the sample to form The fluorescent sample is illuminated by an illumination light field containing a periodic structure in both the lateral and axial directions. The spectrum of the fluorescent sample is modulated by a non-uniform illumination light field and causes a frequency shift. After the fluorescent signal emitted by the fluorescent sample is received by the objective lens, it is converged into the imaging image through the field lens. On the surface, a detector is used to receive the fluorescence signal, and a low-resolution image mixed with high and low frequency information of the fluorescent sample is obtained;

(2) Change the spatial displacement and direction of the illumination light field multiple times, and shoot the fluorescence signal modulated by the stripe intensity again until the corresponding image of a layer of sample is shot; then change the axial position of the sample, and repeatedly shoot the spatial position of different illumination light fields. and the fluorescence signal of the sample under directional illumination to obtain a series of three-dimensional low-resolution images mixed with high and low-frequency information of the fluorescent sample as the original image;

(3) Perform subsequent image processing on the original image obtained in step (2). First, perform parameter estimation, including the spatial frequency p of the illuminating light field, and the phase. Secondly, each frequency band of the fluorescent sample is separated; finally, each spectrum is combined to reconstruct a high-fidelity super-resolution image of the sample.

Further, the objective lens used in step (1) is an oil-immersion objective lens with a numerical aperture NA greater than 1.33.

Further, in the step (1), among the three beams of light, one beam converges to the center of the entrance pupil of the objective lens, and then passes through the objective lens and is vertically incident on the sample, and the other two parallel beams converge to the entrance pupil of the objective lens. At the edge, the straight line connecting the two focusing points passes through the center of the entrance pupil circle. The distance between the two focusing points is close to the diameter of the entrance pupil to make full use of the numerical aperture of the objective lens. After passing through the objective lens, the two focused beams of light are incident on the fluorescent sample at an angle exceeding the critical angle. , the three beams of light finally interfere on the sample to form a non-uniform illumination light field to illuminate the sample.

Further, the steps for generating the original image in step (2) are as follows:

(2.1) The illumination light field in each direction changes the optical path of the central light path and simultaneously changes the optical path of one of the edge beams, so that the latter optical path is twice the former optical path, so that the illumination light field is in Each time it moves one-fifth of its period laterally, while the axial light field does not move relative to the objective lens, achieving a five-step phase shift of the illumination, as shown in Figure 2;

(2.2) Sequentially change the position of the two focused light spots at the entrance pupil of the objective lens where the two edge lights converge, thereby interfering to form corresponding illumination light fields in the transverse direction until illumination light fields in three directions are uniformly generated within a π azimuth angle. , providing isotropic resolution improvement in the lateral direction;

(2.3) Each time the spatial displacement or direction of the illumination light field is changed, as well as the axial position of the sample, the fluorescent sample is modulated and emits a mixed signal that is received by the detector to form a low-resolution image, forming a total of 15* n (n is the number of sample layers) original images are used as the original images for subsequent image reconstruction;

Further, in step (3), the subsequent image processing process includes the following steps:

(3.1) First perform parameter estimation, including the spatial frequency k _xy of the illuminating light field, and the phase As a prerequisite for restoring a high-quality super-resolution image in the subsequent reconstruction process, it specifically includes the following sub-steps:

a. Establish an imaging model for five-step phase-shift illumination in a certain direction of illumination light field:

D(r)=[S(r)·I(r)]*H(r)+D _b (r), (1) Among them, D(r) is the original image captured by 3D-SIM, H(r) is the three-dimensional PSF of the system, S(r) is the distribution function of the sample, is the illumination light field, k _xy is the transverse modulation frequency of the illumination light field, r is the spatial coordinate,/> is the phase of the illumination light field, and D _b (r) is the background noise. When the axis of the illumination light field does not move relative to the objective lens, formula (1) can be rewritten as the following form:

b. Transform the above equation into Fourier space, and obtain the Fourier spectrum corresponding to the original image taken at the nth phase shift:

In the formula, is the OTF corresponding to each transverse component m, and k is the coordinate in Fourier space. It can be seen from the above formula that the spectrum of the original image obtained is the aliasing of each frequency band;

c. Express the original image spectrum obtained by five-step phase shift into matrix form:

Among them, vector and/> respectively defined as/> as well as And vector/> Elements in /> for And the element M _nm in the matrix M is/>

d. By performing autocorrelation on the original image spectrum expressed in formula (3), the transverse modulation frequency k _xy of the illumination light field can be obtained:

Indicates correlation, superscript * indicates complex conjugation, /> The intensity maximum value will be obtained at k′=mk _xy , and the value of mk _xy can be obtained by finding its corresponding coordinates;

e. Construct an auxiliary function to restore the phase of the illumination light field. When it is considered that the phase of each five-step phase shift is 2π/5, construct the following function:

Therefore, the initial phase can be accurately calculated according to the following formula:

Among them, for Real numbers, it can be seen from equations (6) and (7) that the background noise appearing in the original image is eliminated during the phase solution process.

f. Perform step ae on the original image obtained under illumination of all square spots to obtain the modulation frequency k _xy and initial phase of all illumination light fields.

(3.2) After completing parameter estimation, image reconstruction is performed, which specifically includes the following steps:

a. According to formula (4) and the obtained initial phase parameters of the illumination light field, perform each frequency band in the original image spectrum. The separation of can be expressed as:

Among them, M ^-1 is the inverse matrix of M;

b. Apply the frequency band separation process in step a to the original images taken under illumination light fields in the other two directions to obtain the separated frequency bands in the corresponding directions.

c. Finally, the obtained frequency bands are spliced using type deconvolution, which can be expressed as:

in, is the spectrum corresponding to the reconstructed super-resolution image, ω ² is the parameter, A(k) is the apodization function, and the final estimated value of the distribution function of the object/> You can do this by pairing/> Obtained by doing the inverse Fourier transform.

The invention also includes a device for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopic imaging, including a laser 1, a polarization-maintaining single-mode optical fiber 2, a collimator 3, a first reflector 4, a first half One wave plate 5, first polarizing beam splitter 6, second half wave plate 7, second polarizing beam splitter 8, second reflecting mirror 9, third half wave plate 14, first scan Galvanometer 15, first scanning lens 16, third reflecting mirror 12, first piezoelectric ceramic 13, second scanning galvanometer 10, first scanning lens 11, first beam combiner 17, first polarization rotator 18, The first lens 19, the fourth mirror 20, the second piezoelectric ceramic 21, the second lens 22, the second polarization rotator 23, the third lens 24, the second beam combiner 25, the first field mirror 26, two-way Color mirror 27, objective lens 28, fluorescence sample 29, second field lens 30, EMCCD 31, computer 32.

The laser 1 emits linearly polarized light, which is collimated by the collimator 3 after being transmitted through the polarization-maintaining fiber 2; after being reflected by the first mirror 4, it enters the first half-wave plate 5 and the first polarizing beam splitter 6, and adjusts the first polarization beam splitter 6. The one-half wave plate 5 makes the intensity ratio of reflected and transmitted light 1:2. The reflected light constitutes the central beam of the illumination light field formed by interference; the transmitted light then passes through the second one-half wave plate. 7. The first polarizing beam splitter 8, rotate the second half-wave plate 7 so that the intensity ratio of reflected and transmitted light is 1:1; the transmitted light and reflected light constitute the illumination light field formed by interference. Edge beam; the transmitted light is reflected by the second mirror 9 and enters the third half-wave plate 14 and the first scanning galvanometer 15. The third half-wave plate 14 is rotated to make the polarization direction of the beam become vertical. Polarization; the light beam scanned by the first scanning galvanometer 15 enters the first scanning lens 16 to converge the light beam; the light beam reflected by the second polarizing beam splitter passes through the third piezoelectric ceramic 13 on which the first piezoelectric ceramic 13 is installed. The reflection mirror 12 reflects and enters the second scanning galvanometer 10 for scanning; the scanning beam is converged through the second scanning lens 11; the reflected and transmitted beams of the second polarizing beam splitter 8 are combined by the first beam combiner 17; Then it passes through the first polarization rotator to change its polarization direction at different scanning positions; the combined beam is collimated through the first lens; the beam reflected by the first polarizing beam splitter 6 passes through the first polarization beam splitter mounted on it. The third mirror 21 of the piezoelectric ceramic 20 reflects, and then the second lens 22 converges it, and then changes its polarization direction through the second polarization rotator 23 and then collimates it into parallel light through the third lens; the first polarization splitter The three beams of light transmitted and reflected by the beam mirror 6, and transmitted and reflected by the second polarizing beam splitter, are then combined by the second beam combiner; the combined beams are converged to the entrance pupil surface of the objective lens through the field lens, two of which The edge beam convergence point is located at the edge of the entrance pupil of the objective lens, and its connection line passes through the center of the objective lens, and the central beam converges to the center of the entrance pupil; the edge beam and the central beam are collimated by the objective lens and then incident on the sample to interfere to form an illumination light field, where the center beam The beam is vertically incident on the sample, and the edge beam is incident at an angle greater than the critical angle of total reflection to form an interference field with the smallest period to illuminate the sample; by controlling the piezoelectric ceramics 21 and 13, the fourth reflector 20 and the second reflection are changed. The position of the mirror changes the optical path of the central beam relative to the first edge beam, and the optical path of the second edge beam relative to the first edge beam. The optical path changed by the latter is twice that of the former, thereby achieving a phase shift of the interference fringes. ; By controlling the scanning angles of the first scanning galvanometer 15 and the second scanning galvanometer 10, the position of the three focused lights on the back focal plane of the objective lens is changed, thereby changing the direction of the interference fringes;

The fluorescent sample 29 emits fluorescence after being illuminated by the illumination light field. The fluorescence is received by the objective lens 28, reflected by the dichroic mirror 27, and then imaged onto the camera EMCCD 31 through the second field lens 30; the computer 32 controls the first piezoelectric ceramic 15 and the movement of the second piezoelectric ceramic 21, the scanning of the first scanning galvanometer 15 and the second scanning galvanometer 10, and the collection of original images by the EMCCD 31;

Image processing is performed to recover a super-resolution image of the sample.

Analyzing the technical solution of the present invention, it can be concluded that:

(1) It can be seen from formulas (6) and (7) that there is no approximate process in the estimation process of the illumination light field, and it is an accurate numerical solution;

(2) It can be seen from formulas (6) and (7) that during the phase estimation process, the background noise is offset by the constructed function, thus eliminating the impact of noise on phase estimation;

(3) Accurate phase solution can achieve accurate frequency band separation and avoid artifacts in reconstruction caused by moving frequency bands to wrong positions.

Compared with the existing technology, the present invention has the following beneficial technical effects:

The present invention constructs an intermediate function for estimating the phase of the lighting light field. Compared with the traditional phase estimation method, the present invention can accurately estimate the initial phase corresponding to the five-step phase shift of the lighting light field; the present invention can also eliminate the need for The impact of inaccurate phase estimation due to the influence of noise during the phase calculation process; using the accurate phase calculation method proposed by the present invention, accurate separation of each frequency band component can be achieved, eliminating reconstruction caused by inaccurate phase estimation The influence of artifacts can be obtained to obtain high-fidelity reconstruction of the sample; the present invention is suitable for interference-type illumination light fields formed by three beams generated by a spatial light modulator or three beams generated by polarizing beam splitter. It is also suitable for use with digital The illumination light field formed by the projection of micromirrors.

Description of the drawings

Figure 1 is a schematic diagram of a super-resolution microscopy imaging device according to an embodiment of the present invention;

Figure 2(a) is a schematic diagram of the position of the three beams of light at the entrance pupil surface of the objective lens. The three black dots indicate the position where the three beams of light converge. Figure 2(b) shows the angle at which the three beams of light interfere after passing through the objective lens.

Figure 3 shows the axial distribution of the corresponding illumination light field during the five-step phase shift proposed by the present invention. The transverse direction produces one-fifth periodic movement each time, and the axis does not move relative to the objective lens;

Figure 4 is a flow chart of the method of the present invention.

Detailed ways

The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

Example 1

This embodiment provides a method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging, including the following steps:

(1) Divide a parallel light beam into three parallel light beams with equal intensity and consistent polarization direction, converge to the entrance pupil surface of the objective lens, and then become three parallel light beams after passing through the objective lens. The three beams of light interfere on the sample to form The fluorescent sample is illuminated by an illumination light field containing a periodic structure in both the lateral and axial directions. The spectrum of the fluorescent sample is modulated by a non-uniform illumination light field and causes a frequency shift. After the fluorescent signal emitted by the fluorescent sample is received by the objective lens, it is converged into the imaging image through the field lens. On the surface, a detector is used to receive the fluorescence signal, and a low-resolution image mixed with high and low frequency information of the fluorescent sample is obtained.

The objective lens used is an oil-immersion objective lens with a numerical aperture NA greater than 1.33. Among the three beams of light, one beam converges to the center of the entrance pupil of the objective lens, and then passes through the objective lens and is vertically incident on the sample. The other two parallel beams converge to the objective lens. At the edge of the entrance pupil, the straight line connecting the two focusing points passes through the center of the entrance pupil circle. The distance between the two focusing points is close to the diameter of the entrance pupil to make full use of the numerical aperture of the objective lens. After passing through the objective lens, the two focused beams of light emerge at an angle exceeding the critical angle. After being incident on the fluorescent sample, the three beams of light finally interfere on the sample to form a non-uniform illumination light field to illuminate the sample.

(2) Change the spatial displacement and direction of the illumination light field multiple times, and shoot the fluorescence signal modulated by the stripe intensity again until the corresponding image of a layer of sample is shot; then change the axial position of the sample, and repeatedly shoot the spatial position of different illumination light fields. and the fluorescence signal of the sample under directional illumination to obtain a series of three-dimensional low-resolution images mixed with high and low-frequency information of the fluorescent sample as the original image;

The steps to generate the original image are as follows:

(2.1) The illumination light field in each direction changes the optical path of the central light path and simultaneously changes the optical path of one of the edge beams, so that the latter optical path is twice the former optical path, so that the illumination light field is in Each time it moves one-fifth of its period laterally, while the axial light field does not move relative to the objective lens, achieving a five-step phase shift of the illumination, as shown in Figure 2;

(2.2) Sequentially change the position of the two focused light spots at the entrance pupil of the objective lens where the two edge lights converge, thereby interfering to form corresponding illumination light fields in the transverse direction until illumination light fields in three directions are uniformly generated within a π azimuth angle. , providing isotropic resolution improvement in the lateral direction;

(2.3) Each time the spatial displacement or direction of the illumination light field is changed, the fluorescent sample is modulated and emits a mixed signal that is received by the detector to form a low-resolution image, forming a total of 15*n (n is the number of sample layers ) original image as the original image for subsequent image reconstruction.

(3) Perform subsequent image processing on the original image obtained in step (2). First, perform parameter estimation, including the spatial frequency p of the illuminating light field, and the phase. Secondly, each frequency band of the fluorescent sample is separated; finally, each spectrum is combined to reconstruct a high-fidelity super-resolution image of the sample.

The image processing process is shown in Figure 4, including the following steps:

(3.1) First perform parameter estimation, including the spatial frequency k _xy of the illuminating light field, and the phase As a prerequisite for restoring a high-quality super-resolution image in the subsequent reconstruction process, it specifically includes the following sub-steps:

a. Establish an imaging model for five-step phase-shift illumination in a certain direction of illumination light field:

D(r)＝[S(r)·I(r)]*H(r)+D _b (r), (1)

Among them, D(r) is the original image taken by 3D-SIM, H(r) is the three-dimensional PSF of the system, S(r) is the distribution function of the sample, is the illumination light field, k _xy is the transverse modulation frequency of the illumination light field, r is the spatial coordinate,/> is the phase of the illumination light field, and D _b (r) is the background noise. When the axis of the illumination light field does not move relative to the objective lens, formula (1) can be rewritten as the following form:

b. Transform the above equation into Fourier space, and obtain the Fourier spectrum corresponding to the original image taken at the nth phase shift:

In the formula, is the OTF corresponding to each transverse component m, and k is the coordinate in Fourier space. It can be seen from the above formula that the spectrum of the original image obtained is the aliasing of each frequency band;

c. Express the original image spectrum obtained by five-step phase shift into matrix form:

Among them, vector and/> respectively defined as/> as well as And vector/> Elements in /> for And the element M _nm in the matrix M is/>

d. By performing autocorrelation on the original image spectrum expressed in formula (3), the transverse modulation frequency k _xy of the illumination light field can be obtained:

Indicates correlation, superscript * indicates complex conjugation, /> The intensity maximum value will be obtained at k′=mk _xy , and the value of mk _xy can be obtained by finding its corresponding coordinates;

e. Construct an auxiliary function to restore the phase of the illumination light field. When it is considered that the phase of each five-step phase shift is 2π/5, construct the following function:

Therefore, the initial phase can be accurately calculated according to the following formula:

Among them, for Real numbers, it can be seen from equations (6) and (7) that the background noise appearing in the original image is eliminated during the phase solution process.

f. Perform step ae on the original image obtained under illumination of all square spots to obtain the modulation frequency k _xy and initial phase of all illumination light fields.

(3.2) After completing the lighting light field parameter estimation, image reconstruction is performed, which specifically includes the following sub-steps:

a. According to formula (4) and the obtained initial phase parameters of the illumination light field, perform each frequency band in the original image spectrum. The separation of can be expressed as:

Among them, M ^-1 is the inverse matrix of M;

b. Apply the frequency band separation process in step (3.1) and step (3.2) b to the original images taken under illumination light fields in the other two directions to obtain the separated frequency bands in the corresponding directions.

c. Finally, the obtained frequency bands are spliced using type deconvolution, which can be expressed as:

in, is the spectrum corresponding to the reconstructed super-resolution image, ω ² is the parameter, A(k) is the apodization function, and the final estimated value of the distribution function of the object/> You can do this by pairing/> Obtained by doing the inverse Fourier transform.

Example 2

Figure 1 shows a device for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopic imaging, but is not limited to the device shown in Figure 1.

The device of this embodiment includes a laser 1, a polarization-maintaining single-mode optical fiber 2, a collimator 3, a first reflecting mirror 4, a first half-wave plate 5, a first polarizing beam splitter 6, a second half-wave plate 5 A wave plate 7, a second polarizing beam splitter 8, a second reflecting mirror 9, a third half-wave plate 14, a first scanning galvanometer 15, a first scanning lens 16, a third reflecting mirror 12, a first Piezoelectric ceramic 13, second scanning galvanometer 10, first scanning lens 11, first beam combiner 17, first polarization rotator 18, first lens 19, fourth mirror 20, second piezoelectric ceramic 21, Second lens 22, second polarization rotator 23, third lens 24, second beam combiner 25, first field lens 26, dichroic mirror 27, objective lens 28, fluorescent sample 29, second field lens 30, EMCCD 31, computer 32.

The process of realizing three-dimensional structured light illumination super-resolution microscopy imaging using the device shown in Figure 1 is as follows:

(1) The laser 1 emits linearly polarized light, which is collimated by the collimator 3 after being transmitted through the polarization-maintaining fiber 2; reflected by the first mirror 4, it enters the first half-wave plate 5 and the first polarizing beam splitter 6 , adjust the first half-wave plate 5 so that the intensity ratio of reflected and transmitted light is 1:2. The reflected light constitutes the central beam of the illumination light field formed by interference; the transmitted light then passes through the second half-wave plate. A wave plate 7 and a first polarizing beam splitter 8 rotate the second half-wave plate 7 so that the intensity ratio of reflected and transmitted light is 1:1; the transmitted light and reflected light constitute the illumination used for interference formation The edge beam of the light field; the transmitted light is reflected by the second reflector 9 and enters the third half-wave plate 14 and the first scanning galvanometer 15. The third half-wave plate 14 is rotated to change the polarization direction of the beam. It becomes vertically polarized; the light beam scanned by the first scanning galvanometer 15 enters the first scanning lens 16 to converge the light beam; the light beam reflected by the second polarizing beam splitter passes through the first piezoelectric ceramic 13 installed on it The third mirror 12 reflects and enters the second scanning galvanometer 10 for scanning; the scanning beam is converged through the second scanning lens 11; the second polarizing beam splitter 8 reflects and transmits the beam through the first beam combiner 17 Combine the beam; then change its polarization direction at different scanning positions through the first polarization rotator; the combined beam is collimated by the first lens; the beam reflected by the first polarizing beam splitter 6 is mounted on it It is reflected by the third mirror 21 of the first piezoelectric ceramic 20, and then converged by the second lens 22, and then changes its polarization direction through the second polarization rotator 23, and then is collimated into parallel light by the third lens; A polarizing beam splitter 6 transmits and reflects, and the three beams of light transmitted and reflected by the second polarizing beam splitter are then combined by the second beam combiner; the combined light passes through the field lens and converges to the entrance pupil surface of the objective lens, The convergence point of the two edge beams is located at the edge of the entrance pupil of the objective lens, and their connection line passes through the center of the objective lens, and the central beam converges to the center of the entrance pupil, as shown in Figure 2(a); the edge beam and the center beam are collimated by the objective lens The interference incident on the sample forms an illumination light field, in which the central beam is vertically incident on the sample, and the edge beam is incident greater than the critical angle of total reflection, as shown in Figure 2(b) to form an interference field with the smallest period to illuminate the sample; by controlling the pressure The electric ceramic 21 and the piezoelectric ceramic 13 change the positions of the fourth reflector 20 and the second reflector, thereby changing the optical path of the central beam relative to the first edge beam, and the light path of the second edge beam relative to the first edge beam. The optical path changed by the latter is twice that of the former, thereby realizing the phase shift of the interference fringes, as shown in Figure 3; by controlling the scanning angles of the first scanning galvanometer 15 and the second scanning galvanometer 10, the focus of the three beams is changed. The position of the light on the back focal plane of the objective lens changes the direction of the interference fringes, as shown in Figure 2(a);

(2) The fluorescent sample 29 emits fluorescence after being illuminated by the illumination light field. The fluorescence is received by the objective lens 28, reflected by the dichroic mirror 27, and then imaged onto the camera EMCCD 31 through the second field lens 30; the computer 32 controls the first The movement of the piezoelectric ceramic 15 and the second piezoelectric ceramic 21, the scanning of the first scanning galvanometer 15 and the second scanning galvanometer 10, and the acquisition of original images by the EMCCD 31;

(3) Perform image processing according to the steps shown in Figure 4 to restore a three-dimensional super-resolution image of the sample:

1) Parameter estimation

a. Establish an imaging model for five-step phase-shift illumination in a certain direction of illumination light field:

D(r)=[S(r)·I(r)]*H(r)+D _b (r), (1) Among them, D(r) is the original image captured by 3D-SIM, H(r) is the three-dimensional PSF of the system, S(r) is the distribution function of the sample, is the illumination light field, k _xy is the transverse modulation frequency of the illumination light field, r is the spatial coordinate,/> is the phase of the illumination light field, and D _b (r) is the background noise. When the axis of the illumination light field does not move relative to the objective lens, formula (1) can be rewritten as the following form:

b. Transform the above equation into Fourier space, and obtain the Fourier spectrum corresponding to the original image taken at the nth phase shift:

In the formula, is the OTF corresponding to each transverse component m, and k is the coordinate in Fourier space. It can be seen from the above formula that the spectrum of the original image obtained is the aliasing of each frequency band;

c. Express the original image spectrum obtained by five-step phase shift into matrix form:

Among them, vector and/> respectively defined as/> as well as And vector/> Elements in /> for And the element M _nm in the matrix M is/>

d. By performing autocorrelation on the original image spectrum expressed in formula (3), the transverse modulation frequency k _xy of the illumination light field can be obtained:

Indicates correlation, superscript * indicates complex conjugation, /> The intensity maximum value will be obtained at k′=mk _xy , and the value of mk _xy can be obtained by finding its corresponding coordinates;

e. Construct an auxiliary function to restore the phase of the illumination light field. When it is considered that the phase of each five-step phase shift is 2π/5, construct the following function:

Therefore, the initial phase can be accurately calculated according to the following formula:

Among them, for Real numbers, it can be seen from equations (6) and (7) that the background noise appearing in the original image is eliminated during the phase solution process.

f. Perform step ae on the original image obtained under illumination of all square spots to obtain the modulation frequency k _xy and initial phase of all illumination light fields.

2)Image reconstruction

a. According to formula (4) and the obtained initial phase parameters of the illumination light field, perform each frequency band in the original image spectrum. The separation of can be expressed as:

Among them, M ^-1 is the inverse matrix of M;

b. Apply the frequency band separation process in steps 1) and 2)a to the original images taken under illumination light fields in the other two directions to obtain the separated frequency bands in the corresponding directions.

c. Finally, the obtained frequency bands are spliced using Wiener-type deconvolution, which can be expressed as:

in, is the spectrum corresponding to the reconstructed super-resolution image, ω ² is the parameter, A(k) is the apodization function, and the final estimated value of the distribution function of the object/> You can do this by pairing/> Obtained by doing the inverse Fourier transform.

The above are only examples of the preferred implementation of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention. within.

Claims (7)

1. A method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging, which is characterized by including the following steps: (1) Divide a parallel light beam into three parallel light beams with equal intensity and consistent polarization direction, converge to the entrance pupil surface of the objective lens, and then become three parallel light beams after passing through the objective lens. The three beams of light interfere on the sample to form The fluorescent sample is illuminated by an illumination light field containing periodic structures in both the lateral and axial directions. The spectrum of the fluorescent sample is modulated by the non-uniform illumination light field and causes a frequency shift. After the fluorescent signal emitted by the fluorescent sample is received by the objective lens, it is converged to the imaging image plane through the field lens. , use a detector to receive the fluorescence signal, and obtain a low-resolution image mixed with high and low-frequency information of the fluorescent sample; (2) Change the spatial displacement and direction of the illumination light field multiple times, and shoot the fluorescence signal modulated by the stripe intensity again until the corresponding image of a layer of sample is shot; then change the axial position of the sample, and repeatedly shoot the spatial position of different illumination light fields. and the fluorescence signal of the sample under directional illumination, a series of three-dimensional low-resolution images mixed with high and low-frequency information of the fluorescent sample are obtained as the original image; the steps to generate the original image are as follows: (2.1) The illumination light field in each direction changes the optical path of the central light path and simultaneously changes the optical path of one of the edge beams, so that the latter optical path is twice the former optical path, so that the illumination light field is in Each time it moves one-fifth of its period laterally, while the axial light field does not move relative to the objective lens, achieving a five-step phase shift of the illumination; (2.2) Sequentially change the position of the two focused light spots at the entrance pupil of the objective lens where the two edge lights converge, thereby interfering to form corresponding illumination light fields in the transverse direction until illumination light fields in three directions are uniformly generated within a π azimuth angle. , providing isotropic resolution improvement in the lateral direction; (2.3) Each time the spatial displacement or direction of the illumination light field is changed, the fluorescent sample is modulated and emits a mixed signal that is received by the detector to form a low-resolution image; (2.4) After taking one layer of the sample, move the sample to the next layer, repeat steps (2.1)-(2.3), and take low-resolution images corresponding to the samples in different layers; (3) Perform subsequent image processing on the original image obtained in step (2). First, perform parameter estimation, including the spatial frequency p of the illuminating light field, and the phase. Secondly, each frequency band of the fluorescent sample is separated; finally, each spectrum is combined to reconstruct a high-fidelity super-resolution image of the sample; subsequent image processing includes: (3.1) First perform parameter estimation, including the spatial frequency k _xy of the illuminating light field, and the phase As a prerequisite for restoring a high-quality super-resolution image in the subsequent reconstruction process, it specifically includes: a. Establish an imaging model for five-step phase-shift illumination in a certain direction of illumination light field: D(r)＝[S(r)·I(r)]*H(r)+D _b (r), (1) Among them, D(r) is the original image taken by 3D-SIM, H(r) is the three-dimensional PSF of the system, S(r) is the distribution function of the sample, M=2 is the illumination light field, k _xy is the transverse modulation frequency of the illumination light field, r is the spatial coordinate,/> is the phase of the illumination light field, and D _b (r) is the background noise. When the axis of the illumination light field does not move relative to the objective lens, formula (1) can be rewritten as the following form: b. Transform the above equation into Fourier space, and obtain the Fourier spectrum corresponding to the original image taken at the nth phase shift: In the formula, is the OTF corresponding to each transverse component m, and k is the coordinate in Fourier space; from the above formula, it can be seen that the spectrum of the original image obtained is the aliasing of each frequency band; c. Express the original image spectrum obtained by five-step phase shift into matrix form: Among them, vector and/> respectively defined as/> as well as And vector/> Elements in /> for And the element M _nm in the matrix M is/> d. By performing autocorrelation on the original image spectrum expressed in formula (3), the transverse modulation frequency k _xy of the illumination light field can be obtained: Indicates correlation, superscript * indicates complex conjugation, /> The intensity maximum value will be obtained at k′=mk _xy , and the value of mk _xy can be obtained by finding its corresponding coordinates; e. Construct an auxiliary function to restore the phase of the illumination light field. When it is considered that the phase of each five-step phase shift is 2π/5, construct the following function: Therefore, the initial phase can be accurately calculated according to the following formula: Among them, for Real numbers, it can be seen from equations (6) and (7) that the background noise appearing in the original image is eliminated during the phase solution process; f. Perform step ae on the original image obtained under illumination of all square spots to obtain the modulation frequency k _xy and initial phase of all illumination light fields. (3.2) Use the parameters estimated in step (3.1) for subsequent image reconstruction. 2. A method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopic imaging according to claim 1, characterized in that the objective lens used in step (1) is an oil with a numerical aperture NA greater than 1.33. Immersion type objective lens. 3. A method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging according to claim 1, characterized in that, in the step (1), among the three beams of light, one beam is converged to The center of the entrance pupil of the objective lens is located, and then passes through the objective lens and is vertically incident on the sample. The other two parallel beams converge at the edge of the entrance pupil of the objective lens. The straight line connecting the two focusing points passes through the center of the entrance pupil circle, and the distance between the two focusing points is close to the entrance pupil. The pupil diameter is designed to make full use of the numerical aperture of the objective lens. After the two focused beams of light emerge from the objective lens, they are incident on the fluorescent sample at an angle exceeding the critical angle. The three beams of light finally interfere on the sample to form a non-uniform illumination light field to illuminate the sample. 4. A method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopic imaging according to claim 1, characterized in that in the step (2.4), a total of 15*n original images are formed as The original image of subsequent image reconstruction, n is the number of sample layers. 5. A method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging according to claim 1, characterized in that the image reconstruction described in step (3.2) specifically includes: a. According to formula (4) and the obtained initial phase parameters of the illumination light field, perform each frequency band in the original image spectrum. The separation of can be expressed as: Among them, M ^-1 is the inverse matrix of M; b. Apply the frequency band separation process in step a to the original images taken under illumination light fields in the other two directions to obtain the separated frequency bands in the corresponding directions. c. Finally, the obtained frequency bands are spliced using Wiener-type deconvolution, which can be expressed as: in, is the spectrum corresponding to the reconstructed super-resolution image, ω ² is the Wiener parameter, A(k) is the apodization function, and finally the estimated value of the distribution function of the object/> You can do this by pairing/> Obtained by doing the inverse Fourier transform. 6. A device for implementing a method for high-fidelity reconstruction of three-dimensional structured light illumination super-resolution microscopy imaging according to one of claims 1 to 5, characterized by: comprising a laser (1), a polarization-maintaining single-mode optical fiber ( 2), collimator (3), first reflector (4), first half-wave plate (5), first polarizing beam splitter (6), second half-wave plate (7 ), the second polarizing beam splitter (8), the second reflecting mirror (9), the third half-wave plate (14), the first scanning galvanometer (15), the first scanning lens (16), the Three reflecting mirrors (12), first piezoelectric ceramic (13), second scanning galvanometer (10), first scanning lens (11), first beam combiner (17), first polarization rotator (18) , the first lens (19), the fourth mirror (20), the second piezoelectric ceramic (21), the second lens (22), the second polarization rotator (23), the third lens (24), the second Beam combiner (25), first field lens (26), dichroic mirror (27), objective lens (28), fluorescence sample (29), second field lens (30), EMCCD (31), computer (32 ). 7. The device of claim 6, characterized in that: The laser (1) emits linearly polarized light, which is collimated by the collimator (3) after being transmitted through the polarization-maintaining fiber (2); it is reflected by the first mirror (4) and then enters the first half-wave plate (5) The first polarizing beam splitter (6) adjusts the first half-wave plate (5) so that the intensity ratio of reflected and transmitted light is 1:2. The reflected light forms the center of the illumination light field used for interference formation. Light beam; the transmitted light then passes through the second half-wave plate (7) and the first polarizing beam splitter (8), and the second half-wave plate (7) is rotated so that the intensity ratio of the reflected and transmitted light is 1: 1; The transmitted light and reflected light constitute an edge beam for the illumination light field formed by interference; the transmitted light is reflected by the second reflector (9) and enters the third half-wave plate (14) and the first The scanning galvanometer (15) rotates the third half-wave plate (14) to change the polarization direction of the light beam into vertical polarization; the light beam scanned by the first scanning galvanometer (15) exits and enters the first scanning lens (16) Converge the light beam; the light beam reflected by the second polarizing beam splitter is reflected by the third mirror (12) on which the first piezoelectric ceramic (13) is installed, and enters the second scanning galvanometer (10) ) scanning; the scanning beam is converged by the second scanning lens (11); the reflected and transmitted beams of the second polarizing beam splitter (8) are combined by the first beam combiner (17); and then rotated by the first polarization The device changes its polarization direction at different scanning positions; the combined beam is collimated by the first lens; the beam reflected by the first polarizing beam splitter (6) passes through a first piezoelectric ceramic ( 20) is reflected by the third mirror (21), and then converged by the second lens 22), and then changes its polarization direction through the second polarization rotator (23) and then is collimated into parallel light through the third lens; the third lens 22) A polarizing beam splitter (6) transmits and reflects, and the three beams of light transmitted and reflected by the second polarizing beam splitter are combined by a second beam combiner; the combined light passes through the field lens and converges to the entrance pupil of the objective lens surface, where the convergence point of the two edge beams is located at the edge of the entrance pupil of the objective lens, their connection line passes through the center of the objective lens, and the central beam converges to the center of the entrance pupil; the edge beam and the central beam are collimated by the objective lens and then incident on the sample to interfere and form illumination Light field, in which the central beam is vertically incident on the sample, and the edge beam is incident at a angle greater than the critical angle of total reflection to form an interference field with the smallest period to illuminate the sample; by controlling the piezoelectric ceramics (21) and piezoelectric ceramics (13), the The positions of the fourth reflector (20) and the second reflector further change the optical path of the central beam relative to the first edge beam, and the optical path of the second edge beam relative to the first edge beam. The optical path changed by the latter is Twice the former, thereby achieving a phase shift of the interference fringes; by controlling the scanning angles of the first scanning galvanometer (15) and the second scanning galvanometer (10), the positions of the three focused lights on the back focal plane of the objective lens are changed, thereby changing direction of interference fringes; The fluorescent sample (29) emits fluorescence after being illuminated by the illumination light field. The fluorescence is received by the objective lens (28), reflected by the dichroic mirror (27), and then imaged to the camera EMCCD (31) through the second field lens (30). on; the computer (32) controls the movement of the first piezoelectric ceramic (15) and the second piezoelectric ceramic (21), the scanning of the first scanning galvanometer (15) and the second scanning galvanometer (10), and the EMCCD (31 ) collection of original images; Image processing is performed to recover a super-resolution image of the sample.