METHOD FOR SUPER-RESOLUTION EVALUATION OF MICROSCOPE IMAGES ILLUMINATED IN A STRUCTURED MANNER AND MICROSCOPE HAVING STRUCTURED ILLUMINATION

Abstract

Method for super-resolution evaluation of microscope images illuminated in a structured manner and microscope having structured illumination. The resolution can be improved laterally by a factor of up to two using conventional linear structured Illumination (SIM). If a non-linear iterative method is used for the purpose of deconvolution, the achievable resolution can be improved beyond the theoretical limit. However, the known methods only achieve a small amount of increase. The novel method is intended to make improved resolution or improved contrast possible. If a PSF/OTF that is manipulated (individually for each order) in the same (or in a corresponding) way as the relevant order spatial frequency spectrum is used during the re-weighting in the spatial frequency domain (for the deconvolution), the actually achievable resolution can be nearly doubled in comparison with the conventional SIM, both in one-stage and in two-stage variants.

Claims

1. A method for super-resolution evaluation of microscope images of a sample, comprising: providing a plurality of digital raw images of the sample, which are recorded sequentially by means of a microscope by illuminating the sample in different phases with periodically structured illumination light, providing an optical transfer function that represents imaging of the microscope, ascertaining a plurality of order spatial frequency spectra on the basis of the raw images, reconstructing an intermediate result image spatial frequency spectrum WO, weighting the optical transfer function for each ascertained order spatial frequency spectrum, re-weighting the intermediate result image spatial frequency spectrum WO on the basis of the weighted optical transfer function for each ascertained order spatial frequency spectrum, and manipulating at least one of the order spatial frequency spectra before or during the reconstruction and corresponding manipulation of the optical transfer function for the relevant order spatial frequency spectrum, in particular with an identical algorithm or an identical mathematical operation, before or during deconvolution.

2. The method as claimed in claim 1, wherein said manipulating comprises spatial-frequency-dependent weighting, in particular spatial-frequency-dependent filtering, more particularly spatial-frequency-dependent notch filtering, and more particularly around a respective coordinate origin.

3. The method as claimed in claim 1, wherein the illumination light in the sample has at least one repetition frequency and the spatial-frequency-dependent filtering comprises notch filtering that is also dependent on the at least one repetition frequency, in particular notch filtering for the suppression at least of the at least one repetition frequency.

4. The method as claimed in claim 1, wherein the illumination light is also periodically structured along an optical axis of the image recording and different raw images originate from different planes of the sample, wherein the ascertainment of the order spatial frequency spectra on the basis of raw images from different planes of the sample is effected such that three-dimensional order spatial frequency spectra are ascertained, and wherein the optical transfer function is provided three-dimensionally and deconvolution is effected three-dimensionally.

5. The method as claimed in claim 1, wherein no weighting of the order spatial frequency spectra on the basis of the optical transfer function takes place during reconstruction, but in particular with weighting according to the respective illumination intensity.

6. The method as claimed in claim 1, wherein said manipulating differs from an integral transform and from a discrete transform, in particular in a manner such that the optical transfer function for the relevant order spatial frequency spectrum is changed, in particular changed significantly.

7. The method as claimed in claim 1, wherein the deconvolution is iterative, in particular non-linear, more particularly by means of a Richardson-Lucy method or a maximum-likelihood estimation, more particularly in significantly more than five iterations, in particular in more than 40 iterations, or wherein the deconvolution is not iterative.

8. The method as claimed in claim 1, wherein a manipulation of the raw images is performed before the order spatial frequency spectra are ascertained, in particular a spatial-frequency-dependent manipulation, more particularly a spatial-frequency-dependent frequency weighting, more particularly filtering and/or a deconvolution, in particular with the manipulation-free optical transfer function provided, and wherein the ascertainment of the order spatial frequency spectra takes place on the basis of the manipulated raw images.

9. The method as claimed in claim 1, wherein at least one actual parameter of the structured illumination, in particular a repetition frequency and/or an orientation and/or a position, is ascertained on the basis of at least one of the raw images or on the basis of the raw images manipulated as claimed in claim 8.

10. The method as claimed in claim 9, wherein the manipulation of the order spatial frequency spectra and of the OTF for the relevant order spatial frequency spectrum comprises or is a respective shift and/or a respective weighting and/or a weighted summation, on the basis of the at least one ascertained actual parameter.

11. The method as claimed in claim 9, wherein the OTF is provided on the basis of the at least one ascertained actual parameter, in particular by ascertaining simulated individual images under simulated structured illumination with the at least one ascertained actual parameter and an SIM evaluation of the simulated individual images, from which the OTF is ascertained.

12. Method as claimed in claim 1, wherein for each order spatial frequency spectrum a copy of the OTF that is shifted in correspondence with the order spatial frequency spectrum is provided and in all subsequent method steps a respective one of these copies is used as the OTF.

13. Method as claimed in claim 1, wherein different order spatial frequency spectra that are contained in the raw images or in the raw images that were manipulated as claimed in claim 8 are demodulated and separated during the ascertainment of the order spatial frequency spectra and shifted in the spatial frequency domain before or during the reconstruction of the intermediate result image, and wherein the separated and shifted order spatial frequency spectra are combined during the reconstruction of the intermediate result image.

14. The method as claimed in claim 1, wherein the weighted optical transfer functions are combined to form a total transfer function and the deconvolution takes place on the basis of the total transfer function, wherein in particular the manipulation of the optical transfer function for the relevant order spatial frequency spectrum takes place before or during the combination.

15. A method for super-resolution imaging of a sample, wherein, by means of a microscope, in each case with illumination of the sample in different phases with structured illumination light, sample light emitted and/or scattered by the sample is recorded sequentially per phase by means of a detector into a respective raw image in a manner such that it is possible to reconstruct from the raw images an intermediate result image having a resolution that is increased compared to the raw images.

16. The method as claimed in claim 15, wherein a two-dimensionally spatially resolving detector is used to record the raw images, in particular without a stop that optically sections the sample in front of the detector.

17. A microscope having a control unit, and configured for carrying out a method as claimed in claim 1, and having a light source, a two-dimensionally spatially resolving detector for recording raw images of the sample, and means for generating periodically structured illumination light in the sample in different phases, wherein in particular no stop that optically sections the sample is arranged in front of the detector.

18. A computer program, configured for carrying out a method as claimed in claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The invention is explained in more detail below on the basis of exemplary embodiments.

[0055] In the drawings:

[0056] FIG. 1 shows a multimodal microscope,

[0057] FIG. 2 schematically illustrates the known principle of structured illumination in a plurality of phases,

[0058] FIG. 3 schematically illustrates the generation of an illumination pattern with three interfering orders in the pupil,

[0059] FIG. 4 schematically illustrates the generation of an illumination pattern with three interfering orders in the pupil,

[0060] FIG. 5 schematically illustrates the known one-stage SIM evaluation,

[0061] FIG. 6 schematically illustrates the two-stage 3D-SIM evaluation that has been improved according to the invention in comparison with the two-stage 2D-SIM evaluation that has been improved according to the invention,

[0062] FIG. 7 illustrates the two-stage SIM evaluation that has been improved according to the invention in comparison with the known two-stage SIM evaluation, and

[0063] FIG. 8 shows a comparison between differently ascertained images of a sample.

DETAILED DESCRIPTION OF THE DRAWINGS

[0064] In the annexed drawings, corresponding parts are designated by the same reference numerals.

[0065] FIG. 1 shows a microscope 1 having different operating modes. Said microscope is capable of performing both conventional microscopy methods, that is to say microscopy methods the resolution of which is diffraction-limited, and super-resolution microscopy methods, that is to say microscopy methods the resolution of which goes beyond the diffraction limit. This is an inverted microscope. Alternatively, it may be embodied in the form of an upright microscope.

[0066] The microscope 1 captures a sample 2. For this purpose, it has an objective 3 through which the radiation for all microscopy methods described below passes.

[0067] The objective 3 images, via a beam splitter 4, the sample together with a tube lens 5 onto a COD detector 6, which in the example is a two-dimensionally spatially resolving area detector. To this extent, the microscope 1 has a conventional light microscope module 7, and the beam path from the sample 2 through the objective 3 and the tube lens 5 to the COD detector 6 corresponds to a conventional widefield detection beam path 8. As indicated by the double-headed arrow in FIG. 1, the beam splitter 4 is interchangeable so as to be able to switch between beam splitters having different dichroic properties or achromatic beam splitters as per US 2008/0088920.

[0068] Additionally connected in the beam path to the objective 3 is a laser scanning module 9, the LSM illumination and detection beam path of which is coupled into the beam path to the objective 3 via a switching mirror 11, which likewise possesses beam splitter functions. The beam path from the switching mirror 11 to the objective 3 through the beam splitter 4 is thus a beam path in which the illumination beam path and the detection beam path are combined. This is true both with respect to the laser scanning module 9 and also with respect to the widefield detection beam path 8, because, as will be explained below, illumination radiation which, together with the widefield detection beam path 8, i.e. the CCD detector 6, realizes microscopy methods, is also coupled in at the switching mirror 11.

[0069] The switching mirror 11 and the beam splitter 4 are combined to form a beam splitter module 12, as a result of which there is the possibility of interchanging the switching mirror 11 and the beam splitter 4 depending on the application. This is also illustrated by double-headed arrows. Furthermore provided in the beam splitter module 12 is an emission filter 13 that is located in the widefield detection beam path 8 and appropriately filters the spectral components that can propagate through the widefield detection beam path 8. The emission filter 13 in the beam splitter module 12 is of course also interchangeable.

[0070] The laser scanning module 9 receives laser radiation required for the operation from a laser module 15 via an optical fiber 14.

[0071] In the construction illustrated in FIG. 1, a collective illumination beam path 16, through which illumination radiation for various microscopy methods passes, is coupled in at the beam splitter module 12, more specifically at the switching mirror 14. Different illumination beam paths of individual illumination modules are coupled into this collective illumination beam path 16. For example, a widefield module 17 couples widefield illumination radiation into the collective illumination beam path 16 via a switching mirror 18, such that the sample 2 is illuminated in widefield via a tube lens 27 and the objective 3. The widefield illumination module 17 can have an HBO lamp, for example. Provided as a further illumination module is a TIRF illumination module 19, which realizes TIRF illumination with the appropriate positioning of the switching mirror 18. The TIRF illumination module 19 to this end receives radiation from the laser module 15 via an optical fiber 20. The TIRF illumination module 19 has a mirror 21 that is longitudinally displaceable. On account of the longitudinal displacement, the illumination beam emitted by the TIRF illumination module 19 is displaced perpendicularly to the main propagation direction of the emitted illumination beam, as a result of which the TIRF illumination is incident at the objective 3 at an adjustable angle with respect to the optical axis of the objective. In this way, the required angle of total internal reflection at the coverslip can be readily ensured. Other means are of course also suitable for effecting this angle adjustment.

[0072] Furthermore coupled to the collective illumination beam path is the illumination beam path of a manipulator module 22 which likewise receives radiation from the laser module 15 via an optical fiber (not designated further here) and guides a point-shaped or line-shaped beam distribution over the sample 2 in a scanning fashion. The manipulator module 22 thus substantially corresponds to the illumination module of a laser scanning microscope, and, as a consequence, the manipulator module 22 can also be operated in a manner combined with the detector of the laser scanning module 9 or the widefield detection of the COD detector 6.

[0073] A grating 23, having a grating constant below the cut-off frequency that can be transferred with the microscope 1 into the sample 2, is furthermore provided in the collective illumination beam path 16. The grating 23 can be arranged for example in a plane (intermediate image of the sample) of the illumination beam path 16 that is imaged into the sample. The grating 23 is displaceable transversely to the optical axis of the collective illumination beam path 16 in preferably two dimensions transversely to the optical axis. To this end, a corresponding displacement drive 24 is provided.

[0074] An image field rotator 25, which is rotated by a rotator drive 26, is furthermore arranged in the collective illumination beam path 16 downstream of the grating in the illumination direction. The image field rotator can be for example an Abbe-Koenig prism. If the grating 23 is two-dimensionally structured, the image field rotator 25 can be dispensed with because the resulting illumination structure requires no rotation. Instead, it can be displaced in two dimensions, for example.

[0075] The microscope 1 comprises a control unit 28, for example a computer in Von-Neumann architecture, which has in particular a processor as a calculation and control unit, a random access memory as the working memory, and a magnetic hard disk as a mass storage means.

[0076] The modules and the drives and also the detectors of the microscope 1 are all connected to the control unit 28 via lines (not designated further here). The connection can be realized for example via a data and control bus. The control unit 28 controls the microscope 1 in different operating modes. The control device 28 thus permits the performance of conventional microscopy, that is to say widefield microscopy (WF), in particular with structured illumination (SIM), laser scanning microscopy (LSM), and also fluorescence microscopy with total internal reflection (TIRE), on the microscope 1.

[0077] The microscope in FIG. 1 has substantially two modules that are suitable for laser scanner illumination, specifically the laser scanning module 9 and the manipulator module 22. Other combinations are of course also possible. Said modules are coupled via tube lenses with the objective 3 onto the sample 2. The manipulator module 22 includes merely the excitation part of a laser scanning module, without detection. As a result, the sample can be illuminated in point-shaped fashion, and the illumination spot can be scanned over the sample 2.

[0078] Preferably, a switching unit, for example a switching lens or cylindrical lens, with which switching between point-shaped and line-shaped illumination is effected, is also located in the manipulator module 22. Said line-shaped illumination is advantageous in particular when the grating 23 is pivoted in and is located perpendicularly to the line of the line-shaped illumination. Alternatively, the line-shaped illumination could be used for the dynamic (sequential) generation of structured illumination in the sample 2.

[0079] A variably adjustable stripe modulator or a DMD or an SLM can also be used as an alternative to the grating 23 to generate structured illumination in the sample 2. In that case, the displacement drive 24 and the ability to pivot the grating 23 in and out are of course no longer necessary.

[0080] The image field rotator 25 permits the structured illumination that is generated by way of the grating 23 (or by the elements replacing the latter) to be rotated about the optical axis of the collective illumination beam path 16, such that the structured illumination lies at different angles in the sample 2.

[0081] To switch between individual operating types, the switching mirrors 18 and 11 and the beam splitter 4 are adjusted appropriately. In the realization, folding or pivot mirrors can be used to this end, such that switching between the operating types can be effected sequentially. Alternatively, dichroic mirrors that permit simultaneous operation of the various modules are also possible.

[0082] The beam splitter 4 is preferably embodied in the form of a dichroic beam splitter, wherein the spectral properties are adjustable such that spectral components of fluorescence emission from labeling molecules that are to be detected with the aid of the CCD detector 6 pass into the widefield detection beam path 8, and the remaining spectral components are transmitted if possible. To increase the flexibility with respect to the utilizability of labeling molecules with different emission characteristics, a plurality of different beam splitters 4 and emission filters 13 are arranged in the beam splitter module 12 in a manner such that they are interchangeable, for example on a filter wheel.

[0083] The above-described microscope can serve to generate a super-resolved result image. To this end, the control device 28 has an appropriate configuration, for example realized by way of suitable programming.

[0084] FIG. 2 schematically illustrates the concept for producing a super-resolved image in an individual sample plane according to the SIM method. The sample examined under the microscope 1 in FIG. 1 is repeatedly imaged in widefield, wherein different illumination states are set.

[0085] FIG. 2 shows a set of raw images 40 from a single sample plane that all represent the same sample region but differ in terms of a light structure 29 that is transferred during the recording of the raw images 40 into the sample 2 by way of structured illumination by means of the illumination beam path 16. As can be seen, the lateral, periodic light structure 29 generated for example by means of the grating 23 is oriented and positioned differently in the different raw images 40, but has an identical repetition frequency in all the raw images 40 (depending on the grating frequency of the grating 23). In total, the example includes nine raw images 40, which are made up of three different orientations of the structure 29 and three different displacement positions of the structure 29. The various orientations and displacement positions are grouped together under the term phases. Different and larger numbers of different phases are of course also possible, as is known from the publications cited above relating to the principle of SIM.

[0086] However, the structure 29 shown is to be understood to be purely an example. In particular, there is no need for it to be a line structure. It is also possible for the schematically drawn lines to be structured further along the lines. Rather than using the line-type structuring used in the initially cited SIM publications, it is equally possible to use scanned confocal point or line illumination with confocal detection, as is known from the publication “Image scanning microscopy” by C. Müller and J. Enderlein, Physical Review Letters, 104, 198101 (2010). This principle is referred to as ISM. In that case, there are of course not nine orientations of structured illumination but a suitable multiplicity of raw images obtained from scanning a sample 2. Each raw image 40 then corresponds to a specific scanning location, that is to say a specific scanning state during the scanning of the image.

[0087] The control unit 28 calculates a super-resolved result image 50 from the recorded raw images 40. The control unit is here for example configured to be switchable, so that it performs either the known conventional SIM evaluation or the method that has been improved according to the invention.

[0088] FIG. 3 schematically illustrates an example of the realization of the structured illumination by way of the grating 23. By diffracting coherent light, for example coming from the manipulator module 22, at the grating 23, beams that correspond to different orders of diffraction are produced. If, except for the zero order of diffraction and the +/−1 order of diffraction, the remaining orders of diffraction are blocked, three beams that interfere with one another in the sample 2 are present in the pupil of the objective 3, also referred to as back focal plane. The three beams in the objective pupil result in a stripe-shaped SIM illumination pattern 29 being formed in the sample due to interference, which pattern is also structured axially, that is to say longitudinally with respect to the optical axis of the objective 3 through the sample 2, due to the Talbot effect.

[0089] The interfering beams can also be generated in another way, for example by optical fibers ending in (or near) the pupil or by suitably arranged tilt-mirror matrices or by a lightwave-guiding chip as in US 2020/0064609 A1. Nor is the number restricted to three. The generation of the illumination pattern 29 shown in FIG. 3 is exemplary, but it has become the typical approach, in which the various discrete frequency bands present in the illumination pattern 29 are referred to as orders.

[0090] The illumination pattern 29 resulting from three interfering beams can be described by the following equation:

[00001]$\begin{array}{cc}I\left(x,y,z,\phi \right)=a_0^2/2+a^2+a^2\cos \left(2*{\hat{k}}_{x}x+2*{\hat{k}}_{y}y+2\phi \right)+2*{\mathrm{aa}}_0\cos \left({\hat{k}}_{x}x+{\hat{k}}_{y}y+\phi \right)*\cos \left(\left(\hat{k}-{\hat{k}}_z\right)z-{\phi }_0\right)& \left(1\right)\end{array}$

[0091] Here, a.sub.0 is the intensity of the central beam (zero order of diffraction of the grating 23) and a is the intensity of the beams (1 and −1 orders of diffraction of the grating 23) adjacent on both sides. The parameters {circumflex over (k)}.sub.x, {circumflex over (k)}.sub.y, {circumflex over (k)}.sub.z are the x-, y- and z-components of the wave vector of the lateral beams, {circumflex over (k)} is the z-component of the wave vector of the central beam, and ({circumflex over (k)}−{circumflex over (k)}.sub.z) describes the depth dependence of the illumination pattern 29 formed by a cosine centered about a central plane, typically the focal plane of the illumination. Such a depth structure is also referred to as a Talbot pattern. φ.sub.0 is the phase angle of said cosine profile with respect to the z-coordinate. φ.sub.0 refers to the phase angle of the illumination pattern 15 in the x-y-plane, that is to say it is the parameter characterizing the individual SIM illumination patterns. They differ from one another only in terms of the value for φ. By way of example, in the case of five individual SIM illumination pattern phases there are five different values for φ. The phase angles are preferably equally distributed over a region of 180° within specific limits. As easily shown by a Fourier transform, the illumination pattern 29 is characterized by signal components with discrete spatial frequency bands, which correspond to orders.

[0092] Generally, the illumination pattern I(x, y, z) is multiplied on account of the interaction with the sample 2 by the optical properties thereof S(x, y, z) and convolved with the detection PSF H(x, y, z) to the measured light distribution I.sub.em(x, y, z, z, z.sub.0, φ):

I.sub.em(x,y,z,z.sub.0,φ)=∫dx′dy′dz′I(x′,y′,z′,φ)S(x′,y′,z.sub.0−z′)H(x−x′,y−y′,z+z′)  (2),

wherein z.sub.0 indicates the distance of the considered sample plane from the focal plane, and z indicates the position of the focal plane. By setting z=0, a plane of the sample located in the focal plane of the illumination is considered, resulting in:

I.sub.em(x,y,z,φ)=∫dx′dy′dz′I(x′,y′,z′,φ)S(x′,y′,z.sub.0−z′)H(x−x′,y−y′,z′)  (3)

[0093] The optical properties S(x, y, z) of the sample 2 are given for example by the distribution of the concentration of an emitting fluorescent dye in the sample 2 or the reflectivity of the sample 2.

[0094] The intermediate steps of the evaluation are preferably carried out in the spatial frequency domain. A Fourier transform with respect to x, y and z gives:

I.sub.em.sup.f(k.sub.x,k.sub.y,k.sub.z)=∫dk.sub.x.sup.′dk.sub.y.sup.′dk.sub.z.sup.′I.sup.f(k.sub.x.sup.′,k.sub.y.sup.′,k.sub.z.sup.′)S.sup.f(k.sub.x−k.sub.x.sup.′,k.sub.y−k.sub.y.sup.′,k.sub.z)H.sup.f(k.sub.x,k.sub.y,k.sub.z−k.sub.z.sup.′)   (4)

[0095] A system of equations can be established on the basis of the concretely used illumination pattern I(x, y, z), more specifically on the basis of its Fourier transform I.sup.f(k.sub.x, k.sub.y, k.sub.z), the detection PSF H(x, y, z) of the microscope, more specifically on the basis of its Fourier transform H.sup.f(x, y, z) that is to say of the OTF—and the intensity I.sub.em(x, y, z) recorded in the raw images 40, more specifically on the basis of the Fourier transform I.sub.em.sup.f(k.sub.x, k.sub.y, k.sub.z) of the recorded raw images 40.

[0096] For three beams interfering in the sample 2, the result in the first step of inserting equation (1) into equation (4) is:

I.sub.em.sup.f(k.sub.x,k.sub.y,k.sub.zφ)=A.sub.0H.sup.f(k.sub.x,k.sub.y,k.sub.z)S.sup.f(k.sub.x,k.sub.y,k.sub.z)+A.sub.2H.sup.f(k.sub.x,k.sub.y,k.sub.z){S.sup.f(k.sub.x+2{circumflex over (k)}.sub.x,k.sub.y+2{circumflex over (k)}.sub.y,k.sub.z)e.sup.−i2φ+S.sup.f(k.sub.x−2{circumflex over (k)}.sub.x,k.sub.y−2{circumflex over (k)}.sub.y,k.sub.z)e.sup.i2φ}+A.sub.1{H.sup.f(k.sub.x,k.sub.y,k.sub.z+[{circumflex over (k)}.sub.z−{circumflex over (k)}])e.sup.iφ.sup.0+H.sup.f(k.sub.x,k.sub.y,k.sub.z−[{circumflex over (k)}.sub.z−{circumflex over (k)}])e.sup.−φ.sup.0}S.sup.f(k.sub.x,{circumflex over (k)}.sub.x,k.sub.y+{circumflex over (k)}.sub.y,k.sub.z)e.sup.−iφ+A.sub.1{H.sup.f(k.sub.x,k.sub.y,k.sub.z+[{circumflex over (k)}.sub.z−{circumflex over (k)}])e.sup.iφ.sup.0+H.sup.f(k.sub.x,k.sub.yk.sub.z−[{circumflex over (k)}.sub.z−{circumflex over (k)}])e.sup.iφ.sup.0}S.sup.f(k.sub.x−{circumflex over (k)}.sub.x,k.sub.y−{circumflex over (k)}.sub.y,k.sub.z)e.sup.iφ   (5).

[0097] Here, A.sub.j with j=0, 1, 2 denote the intensities of the orders of the illumination pattern, which emerge from the pre-factors in equation (1). Thus, the following applies: A.sub.0=a.sub.0.sup.2/2+a.sup.2; A±.sub.1=a.sup.2 and A±.sub.2=a.sub.0a. Further, k=(k.sub.x, k.sub.y, k.sub.z) is the wave vector which refers to the orders.

[0098] Equation (5) can be expressed in order terms:

I.sub.em.sup.f(k.sub.x,k.sub.y,k.sub.z)=I.sub.em.sup.f(k)=D.sub.0(k)+e.sup.−i2φD.sub.−2(k)+e.sup.i2φD.sub.2(k)+e.sup.i2φD.sub.−1(k)+e.sup.iφD.sub.1(k)   (1),

wherein the orders D.sub.n(k) are defined as

[00002]$\begin{array}{cc}{D}_j\left(k\right)=D_j\left(k_x,k_y,k_z\right)=\{\begin{array}{cc}{A}_0{\mathrm{OTF}}_n\left(k_x,k_y,k_z\right)S\left(k_x,k_y,k_z\right),& j=0\\ A_2{\mathrm{OTF}}_n\left(k_x,k_y,k_z\right)S\left(k_x+2{\hat{k}}_x,k_y+2{\hat{k}}_y,k_z\right),& j=2\\ A_2{\mathrm{OTF}}_n\left(k_x,k_y,k_z\right)S\left(k_x-2{\hat{k}}_x,k_y-2{\hat{k}}_y,k_z\right),& j=-2\\ A_1{\mathrm{OTF}}_n\left(k_x,k_y,k_z\right)S\left(k_x+{\hat{k}}_x,k_y+{\hat{k}}_y,k_z\right),& j=1\\ A_1{\mathrm{OTF}}_n\left(k_x,k_y,k_z\right)S\left(k_x-{\hat{k}}_x,k_y-{\hat{k}}_y,k_z\right),& j=-1\end{array}& \left(2\right)\end{array}$

and the optical transfer functions of the orders are:

OTF.sub.j(k.sub.x,k.sub.y,k.sub.z)=H.sup.f(k.sub.x,k.sub.y,k.sub.z) for j=−2,0,2

OTF.sub.j(k.sub.x,k.sub.y,k.sub.z)=H.sup.f(k.sub.x,k.sub.y,k.sub.z+[{circumflex over (k)}.sub.z−{circumflex over (k)}])e.sup.iφ.sup.0+H.sup.f(k.sub.x,k.sub.y,k.sub.z+[{circumflex over (k)}.sub.z−{circumflex over (k)}])e.sup.−iφ.sup.0 for j=−1,1   (8).

[0099] In order to individually obtain the five orders, n=5 measurements in different phase angles φ=φ.sub.m, m=1, . . . , n of the illumination pattern 29 are necessary. Due to the n different displacement positions of the illumination pattern 29, n equations (6) are obtained therefrom one for each of the phases φ.sub.m. Solving this system of equations brings about a separation of the orders, the result is n order spatial frequency spectra 41, which each contain one of the orders D.sub.j(k).

[0100] These are shifted in each case by {circumflex over (k)}.sub.x, {circumflex over (k)}.sub.y to their original locations in the spatial frequency domain and combined into a spatial frequency spectrum of the intermediate result image:

[00003]$\begin{array}{cc}{S}_{\mathrm{SR}}^f\left(k_x,k_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)\underset{j=-2}{\overset{2}{.Math.}}{\omega }_j\left(k_x,k_y,k_z\right)D_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right),& \left(3\right)\end{array}$

wherein ϕ(k.sub.x, k.sub.y, k.sub.z) is an apodization function and ω.sub.j(k.sub.x, k.sub.y, k.sub.z) are the weights of the j-th order (SR stands for “super resolution”). The apodization function and the weights enable a manipulation of the order spatial frequency spectra 41 or D.sub.j(k), which serves to improve the results.

[0101] There are different possible approaches for the apodization function and the weights. ϕ(k.sub.x, k.sub.y, k.sub.z) can be defined, for example, as Euclidean distance transform of the binary mask that spans the support of the super-resolution OTF:

Σ.sub.j=−2.sup.2|OTF.sub.j(k.sub.x−j{circumflex over (k)}.sub.x,k.sub.y−j{circumflex over (k)}.sub.y,k.sub.z)|.sup.2.

[0102] In conventional one-stage SIM methods (in which the intermediate result image already is the result image), the use of a generalized Wiener filter (as in Gustafsson et al.) gives the following weights:

[00004]$\begin{array}{cc}{\omega }_j\left(k_x,k_y,k_z\right)=\frac{A_j{\mathrm{OTF}}_j^{*}\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)}{w+{.Math.}_{j=-2}^2{A}_j^2{.Math.{\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right).Math.}^2}& \left(4\right)\end{array}$

with the Wiener parameter w and the order intensities A.sub.j. In addition, for example a notch filter g can be used at the center of each order, as described by Bozinovic et al. in: “Fluorescence endomicroscopy with structured illumination,” Optics Express, 2008, vol. 16, page 8016:

[00005]$\begin{array}{cc}{\omega }_j\left(k_x,k_y,k_z\right)=\frac{A_{j}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right){\mathrm{OTF}}_j^{*}\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)}{\begin{array}{c}w+{.Math.}_{j=-2}^2{A}_j^{2}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right)\\ {.Math.{\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right).Math.}^2\end{array}},& \left(5\right)\end{array}$

wherein the notch filter g(x, y) can have, for example, a Gaussian shape:

[00006]$\begin{array}{cc}g\left(x,y\right)=1-a\exp \left(-\frac{x^2}{2{\sigma }^2}-\frac{y^2}{2{\sigma }^2}\right\},& \left(6\right).\end{array}$

[0103] The variables a and a indicate the strength and width of the suppression. By calculating S.sub.SR.sup.f(k.sub.x, k.sub.y, k.sub.z) and an inverse Fourier transform (“iFT”), the result image S.sub.SR(x, y, z) can be directly ascertained.

[0104] In a two-stage method, the following weights can be used for example:

ω.sub.j(k.sub.x,k.sub.y,k.sub.z)=B.sub.jg(k.sub.x−j{circumflex over (k)}.sub.x,k.sub.y−j{circumflex over (k)}.sub.y)   (7),

wherein B.sub.j are order-dependent pre-factors. Other eights are, however, likewise possible.

[0105] The intermediate result image spatial frequency spectrum 44 is then a combination of the orders 41 (D.sub.j) that have been shifted and those that have been manipulated (for example centrally notch-filtered) by the weights ω.sub.j:

[00007]$\begin{array}{cc}{S}_{\mathrm{SR}}^f\left(k_x,k_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)\underset{j=-2}{\overset{2}{.Math.}}{B}_{j}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right)D_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right).& \left(8\right)\end{array}$

[0106] As above, ϕ(k.sub.x, k.sub.y, k.sub.z) is an apodization function and g is for example a Gaussian notch filter. By applying this exemplary notch filter g to higher (non-zero) orders, artifacts can be reduced. In the zero order, by contrast, where the notch filter acts like a high-pass, it serves to block out-of-focus light.

[0107] On the other hand, inserting equation (7) into equation (14) gives for the intermediate result spatial frequency spectrum 44:

[00008]$\begin{array}{cc}{S}_{\mathrm{SR}}^f\left(k_x,k_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)S^f\left(k_x,k_y,k_z\right)\underset{j=-2}{\overset{2}{.Math.}}{B}_j{A}_{j}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right){\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)S^f\left(k_x,k_y,k_z\right)\underset{j=-2}{\overset{2}{.Math.}}{A}_j{\omega }_j\left(k_x,k_y,k_z\right){\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)=,& \left(9\right)\end{array}$

in another representation

S.sub.SR.sup.f(k.sub.x,k.sub.y,k.sub.z)=S.sup.f(k.sub.x,k.sub.y,k.sub.z)OTF.sub.SR(k.sub.x,k.sub.y,k.sub.z)   (10)

with the following total OTF, composed of the OTFs (manipulated, for example centrally notch-filtered, like the orders D.sub.j by the weights ω.sub.j) of the individual orders:

[00009]$\begin{array}{cc}{\mathrm{OTF}}_{\mathrm{SR}}\left(k_x,k_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)\underset{j=-2}{\overset{2}{.Math.}}{B}_j{A}_{j}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right){\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)S^f\left(k_x,k_y,k_z\right)\underset{j=-2}{\overset{2}{.Math.}}{A}_j{\omega }_j\left(k_x,k_y,k_z\right){\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right).& \left(11\right)\end{array}$

[0108] Equation (16) is a convolution represented in the spatial frequency domain, which can be re-weighted on the basis of the combination S.sub.SR.sup.f(k.sub.x, k.sub.y, k.sub.z) of the shifted and manipulated orders that is calculated according to equation (14) and on the basis of the total OTF OTF.sub.SR calculated according to equation (17) and can be solved in many ways using known deconvolution algorithms, for example by non-linear iterative algorithms such as Richardson-Lucy and preferably with the specification of constraints (such as non-negativity and smoothness) and possibly with regularization. Other deconvolution algorithms are also possible. In particular, non-iterative methods can be used.

[0109] For the pre-factors B.sub.i, there are for example two simple possibilities:

[0110] 1. B.sub.i ≡1, so that the strengths A.sub.i of the orders determine the total OTF, or

[0111] 2. B.sub.i=1/A.sub.i, so that the strengths A.sub.i of the orders instead determine the combination of the order spatial frequency spectra 41 (D.sub.j).

[0112] Other pre-factors B.sub.i are also possible.

[0113] It is not necessary to calculate the total OTF as a numerical intermediate result. Rather, for example the individual summation terms of the orders can be used directly in the re-weighting according to equation (16).

[0114] In order to completely capture the sample information, the illumination structure 29 that is one-dimensional in the considered sample plane in this example must be used and evaluated in a plurality of (typically three or five) orientations. The above equations should be expanded accordingly to a plurality of rotation phase angles.

[0115] FIG. 4 schematically illustrates a further example of the realization of the structured illumination. For example five beams lie in the pupil of the objective 3 that interfere with one another in the sample 2 to form a three-dimensionally periodic light structure 29. For example, these five beams are generated using a chip (not depicted) arranged in the pupil. The resulting illumination pattern 29 can be described by the following equation:

I(x,y,z)=α.sub.0.sup.2/2+2*α.sup.2+2*α.sup.2 cos(2*{circumflex over (k)}.sub.yy+φ.sub.−2-1)+2*αα.sub.0 cos({circumflex over (k)}.sub.xx+{circumflex over (k)}.sub.yy+φ.sub.−20)*cos(({circumflex over (k)}−{circumflex over (k)}.sub.z)z−φ.sub.0)+2*α.sup.2 cos(2*{circumflex over (k)}.sub.xx+φ.sub.−21)+α.sup.2 cos(2*{circumflex over (k)}.sub.xx+2*{circumflex over (k)}.sub.yy+φ.sub.−22)+2*αα.sub.0 cos({circumflex over (k)}.sub.xx−{circumflex over (k)}.sub.yy+φ.sub.−10)*cos(({circumflex over (k)}−{circumflex over (k)}.sub.z)z−φ.sub.0)+α.sup.2 cos(2*{circumflex over (k)}.sub.xx−2*{circumflex over (k)}.sub.yy+φ.sub.−11)

[0116] This illumination pattern is two-dimensionally periodic (grid pattern) in the sample plane under consideration. It results in n=13 different orders D.sub.j, that is to say order spatial frequency spectra 41, which must be extracted in the spatial frequency domain and separated. Therefore, at least 13 raw images 40 of every sample plane to be considered are needed. As for the remainder, the evaluation corresponds to the procedure described above for three interfering beams with corresponding equations for 13 orders.

[0117] The intermediate result image spatial frequency spectrum is then composed of the thirteen manipulated order spatial frequency spectra 41 (that is to say D) that have been shifted to their original locations,

[00010]$\begin{array}{cc}{S}_{\mathrm{SR}}^f\left(k_x,k_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)\underset{j=-6}{\overset{6}{.Math.}}{B}_{j}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right)D_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)& \left(12\right)\end{array}$

with the total OTF

[00011]$\begin{array}{cc}{\mathrm{OTF}}_{\mathrm{SR}}\left(k_x,k_y,k_z\right)=\Phi \left(k_x,k_y,k_z\right)\underset{j=-6}{\overset{6}{.Math.}}{B}_j{A}_{j}g\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y\right){\mathrm{OTF}}_j\left(k_x-j{\hat{k}}_x,k_y-j{\hat{k}}_y,k_z\right)& \left(13\right)\\ \mathrm{and}& \\ D_j\left(k\right)=D_j\left(k_x,k_y,k_z\right)=\{\begin{array}{cc}{A}_0{\mathrm{OTF}}_0\left(k_x,k_y,k_z\right)S\left(k_x,k_y,k_z\right),& j=0\\ A_1{\mathrm{OTF}}_1\left(k_x,k_y,k_z\right)S\left(k_x,k_y+2{\hat{k}}_y,k_z\right),& j=1\\ A_1{\mathrm{OTF}}_1\left(k_x,k_y,k_z\right)S\left(k_x,k_y-2{\hat{k}}_y,k_z\right),& j=-1\\ A_2{\mathrm{OTF}}_2\left(k_x,k_y,k_z\right)S\left(k_x+{\hat{k}}_x,k_y+{\hat{k}}_y,k_z\right),& j=2\\ A_2{\mathrm{OTF}}_2\left(k_x,k_y,k_z\right)S\left(k_x-{\hat{k}}_x,k_y-{\hat{k}}_y,k_z\right),& j=-2\\ A_3{\mathrm{OTF}}_3\left(k_x,k_y,k_z\right)S\left(k_x+2{\hat{k}}_x,k_y,k_z\right),& j=3\\ A_3{\mathrm{OTF}}_3\left(k_x,k_y,k_z\right)S\left(k_x-2{\hat{k}}_x,k_y,k_z\right),& j=-3\\ A_4{\mathrm{OTF}}_4\left(k_x,k_y,k_z\right)S\left(k_x+2{\hat{k}}_x,k_y+2{\hat{k}}_y,k_z\right),& j=4\\ A_4{\mathrm{OTF}}_4\left(k_x,k_y,k_z\right)S\left(k_x-2{\hat{k}}_x,k_y-2{\hat{k}}_y,k_z\right),& j=-4\\ A_5{\mathrm{OTF}}_5\left(k_x,k_y,k_z\right)S\left(k_x+{\hat{k}}_x,k_y-{\hat{k}}_y,k_z\right),& j=5\\ A_5{\mathrm{OTF}}_5\left(k_x,k_y,k_z\right)S\left(k_x-{\hat{k}}_x,k_y+{\hat{k}}_y,k_z\right),& j=-5\\ A_5{\mathrm{OTF}}_6\left(k_x,k_y,k_z\right)S\left(k_x+2{\hat{k}}_x,k_y-2{\hat{k}}_y,k_z\right),& j=6\\ A_5{\mathrm{OTF}}_6\left(k_x,k_y,k_z\right)S\left(k_x-2{\hat{k}}_x,k_y+2{\hat{k}}_y,k_z\right),& j=-6\end{array}& \left(14\right)\\ {\mathrm{OTF}}_j\left(k_x,k_y,k_z\right)=H^f\left(k_x,k_y,k_z\right)\mathrm{for}j=0,1,3,4,6\mathrm{and}{\mathrm{OTF}}_j\left(k_x,k_y,k_z\right)=H^f\left(k_x,k_y,k_z+\left[{\hat{k}}_z-\hat{k}\right]\right)e^{+{i\phi }_0}+H^f\left(k_x,k_y,k_z-\left[{\hat{k}}_z-\hat{k}\right]\right)e^{-{i\phi }_0}\mathrm{for}j=2\mathrm{and}j=5& \left(15\right)\end{array}$

[0118] For the pre-factors B.sub.i, the possibilities shown in relation to FIG. 3 exist for example, but others are also possible.

[0119] In all cases (as also in FIG. 3), the system of equations to be solved (and thus also the further evaluation steps) can be expanded to a plurality of sample planes (and thus three spatial dimensions), for example by introducing into the equation (9) a Dirac comb that links the equations for different sample planes. Both the order spatial frequency spectra 40 and the OTF 30 are then three-dimensional.

[0120] FIG. 5 schematically illustrates the procedure of a conventional one-stage SIM evaluation method.

[0121] As the starting point, thirteen raw images 40 and an OTF 30 of the microscope 1 are provided in step S0. The raw images were recorded for example in the microscope 1 at thirteen different illumination phase angles.

[0122] In step S1, as described above, a linear equation system is established and solved on the basis of the raw images 40 and on the basis of the OTF 30. The result is thirteen order spatial frequency spectra 41. The zero-order spectrum is illustrated by an emboldened border so as to stand out. In addition, in step 2, the respective actual phase angle (position) of the illumination structure 20 is ascertained from the raw images, and the actual grid frequencies of the illumination structure (generally the repetition frequencies thereof) are ascertained from the order spatial frequency spectra 41. These are used in the further evaluation.

[0123] In step S3, the order spatial frequency spectra 41 are manipulated, for example resulting in filtered order spatial frequency spectra 42. For example, a notch filter is applied centrally in each order 41 (illustrated by a circle at the center of each spectrum 41). In the zero order 41, this is used to suppress out-of-focus light, and in the remaining orders 41 to suppress grid frequencies. It is possible to apply a plurality of filters in succession or in combination.

[0124] In step S4, the filtered (or otherwise manipulated) order spatial frequency spectra 42 are shifted to their original locations in the spatial frequency domain. In step S5, these shifted spectra are weighted on the basis of their illumination intensity and on the basis of the OTF 30 and are put together by means of a generalized Wiener filter and an apodization function to form the super-resolved result image spatial frequency spectrum 44. The weighting on the basis of the OTF 30 is illustrated by various gray levels and a different degree of interruption in the lines.

[0125] The result image spatial frequency spectrum 44 is ultimately transferred from the spatial frequency domain into the spatial domain using an inverse Fourier transform in order to obtain the super-resolved result image 50.

[0126] Arrows labeled “A” are used in FIG. 5 to illustrate that the data from the raw images 40 are transferred from one evaluation step to the next one and are processed further. An arrow labeled “B” signifies that the OTF 30 is used only in the generalized Wiener filter in step S5, but not during the manipulation of the order spatial frequency spectra 41 in step S3.

[0127] FIG. 6 schematically illustrates the procedure of a two-stage SIM evaluation method that has been improved according to the invention as compared to a conventional two-stage SIM method.

[0128] As the starting point, thirteen raw images 40 are provided in step S0 and transformed, for example by means of a Fourier transform, into the spatial frequency domain. In addition, the respective actual phase angle (position and orientation) of the illumination structure 29 is ascertained from the raw images, and the actual repetition frequencies (grid frequencies) of the illumination structure are ascertained from the local frequency spectra of the raw images. The actual repetition frequencies are used in the entire subsequent method as an actual parameter of the illumination structure.

[0129] The raw images were recorded for example using the microscope 1 at thirteen different illumination phase angles and are loaded from a mass storage means for the purposes of making them available, for example. In step S1, a PSF of the microscope 1 is ascertained by loading it for example from a mass storage means, and the OTF 30 is calculated therefrom by a Fourier transform and copied to obtain thirteen copies. The pre-stored PSF was ascertained for example by a measurement of a calibration sample with structured illumination.

[0130] In step S2, as described above, a linear equation system is established and solved on the basis of the raw images 40. The result is thirteen order spatial frequency spectra 41. The zero-order spectrum is illustrated by an emboldened border so as to stand out. Exactly one of the OTF copies 30 is assigned to each of the order spatial frequency spectra 41.

[0131] In step S3, the order spatial frequency spectra 41 (D.sub.j) are manipulated, for example resulting in filtered order spatial frequency spectra 42 (ω.sub.jD.sub.j). For example, a notch filter (g) is applied centrally in each order 41 (illustrated by a circle at the center of each spectrum 41). It is possible to apply a plurality of different filters successively or in combination, in particular depending on the relevant order. Depending on the type of manipulation (for example depending on the filter), per filter only one of the order spatial frequency spectra 41 or more than one or even all of them are manipulated. In the method that has been improved according to the invention, in an expanded step S3, the OTF copies 30 are manipulated accordingly with for example identical parameters, in particular filter parameters, as the respectively associated order spatial frequency spectrum 41.

[0132] In the conventional step S4, the filtered (or otherwise manipulated) order spatial frequency spectra 42 are shifted to their original locations in the spatial frequency domain, and in the conventional step S5 they are weighted on the basis of their illumination intensity (and not on the basis of the OTF 30 and without Wiener filter). Both in the conventional step S5 and in the improved version S5′, a total OTF is ascertained in that the OTF copies 30 for each order spatial frequency spectrum 42 are shifted like the respectively associated order spatial frequency spectrum 42 and are added in a weighted manner to form a total OTF (not shown). The step S5′ of the method that has been improved according to the invention differs from the conventional step 5 in that the OTF copies 30 manipulated in step S3 are used rather than the non-manipulated starting OTF used in the conventional step S5.

[0133] In the conventional step S6, the intermediate result image spatial frequency spectrum 44 is re-weighted on the basis of the (unchanged) OTF 30 and is deconvolved for example by an iterative method according to Richardson-Lucy, as a result of which the result image spatial frequency spectrum (not shown) is obtained. The result image 50 can be ascertained therefrom using an inverse Fourier transform.

[0134] The step S6′ of the method that has been improved according to the invention differs from the conventional step S6 in that the re-weighting (and thus the deconvolution) is effected on the basis of the total OTF, which is composed of the OTF copies 30 that have been manipulated like the order spatial frequency spectra 42. Alternatively, the manipulated OTF copies 30 can be used without prior combination directly in the re-weighting (deconvolution).

[0135] The fact that the copies of the OTF are manipulated (in contrast to the conventional method) in correspondence with the order spatial frequency spectra 41 is illustrated by the arrows labeled “B.”

[0136] FIG. 7 illustrates a method that has been improved according to the invention for three spatial dimensions as compared to the previously described 2D method. The OTFs 30 provided for the 3D case are three-dimensional from the start. The system of equations for ascertaining the three-dimensional order spatial frequency spectra 41 is also established and solved for three dimensions. The manipulation (filtering) of the order spatial frequency spectra 41 and of the OTF copies 30 also takes place in three dimensions, as do the weighting and combining to form the intermediate result spatial frequency spectrum and the re-weighting (finally deconvolution) on the basis of the manipulated OTF 30, for example by forming a total OTF (not shown).

[0137] FIG. 8 shows, one next to the other, a diffraction-limited image (A), a super-resolved result image (B) ascertained using conventional one-stage SIM, and a super-resolved image (C) ascertained according to the invention by means of two-stage SIM. The images all show the same region of a sample 2 to allow direct comparison. Beneath, enlarged sections from the conventional result image (D) and the result image (E) that has been ascertained according to the invention are illustrated. It is clear that the resolution in the images C and E is significantly better than in the images B and C.

[0138] The imaged sample 2 contained DNA origamis of GATTAquant (https://www.gattaquant.com) at a size of 60 nm with the fluorescent dye Alexa 488. Image A and the SIM raw images (not shown) were recorded using a 63×1.4 plan apochromat oil objective. The illumination pattern was generated as five-beam interference according to FIG. 4 and projected in 13 different phases into the sample 2, A 3D stack with seven planes was recorded, wherein 13 raw images 40 were recorded in each plane. Only one plane is shown.

[0139] It can be seen on the basis of the separately identifiable ends of the origamis that the SIM evaluation that has been improved according to the invention achieves, by contrast to the conventional SIM, a resolution of 60 nm.

[0140] While the invention has been illustrated and described in connection with currently preferred embodiments shown and described in detail, it is not intended to be limited to the details shown since various modifications and structural changes may be made without departing in any way from the spirit of the present invention. The embodiments were chosen and described in order to best explain the principles of the invention and practical application to thereby enable a person skilled in the art to best utilize the invention and various embodiments with various modifications as are suited to the particular use contemplated.

LIST OF REFERENCE SIGNS

[0141] 1 Microscope [0142] 2 Sample [0143] 3 Objective [0144] 4 Beam splitter [0145] 5 Tube lens [0146] 6 CCD detector [0147] 7 Light microscope module [0148] 8 Widefield detection beam path [0149] 9 Laser scanning module [0150] 11 Switching mirror [0151] 12 Beam splitter module [0152] 13 Emission filter [0153] 14 Optical fiber [0154] 15 Laser module [0155] 16 Collective illumination beam path [0156] 17 Widefield illumination module [0157] 18 Switching mirror [0158] 19 TIRF illumination module [0159] 20 Optical fiber [0160] 21 Mirror [0161] 22 Manipulator module [0162] 23 Grating [0163] 24 Displacement drive [0164] 25 Image field rotator [0165] 26 Rotator drive [0166] 27 Tube lens [0167] 28 Control device [0168] 29 Illumination structure [0169] 30 PSF/OTF [0170] 40 Raw image [0171] 41 Order spatial frequency spectrum [0172] 42 Filtered order spatial frequency spectrum [0173] 43 Weighted filtered order spatial frequency spectrum [0174] 44 Intermediate result image spatial frequency spectrum [0175] 50 Result image [0176] A Sample measurement data (raw images) [0177] B PSF/OTF data