CN111580271A - Self-adaptive aberration correction method and light sheet microscopic imaging device based on same - Google Patents
Self-adaptive aberration correction method and light sheet microscopic imaging device based on same Download PDFInfo
- Publication number
- CN111580271A CN111580271A CN202010587864.3A CN202010587864A CN111580271A CN 111580271 A CN111580271 A CN 111580271A CN 202010587864 A CN202010587864 A CN 202010587864A CN 111580271 A CN111580271 A CN 111580271A
- Authority
- CN
- China
- Prior art keywords
- aberration correction
- zernike
- optimal solution
- correction method
- adaptive
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G02—OPTICS
- G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G02B27/00—Optical systems or apparatus not provided for by any of the groups G02B1/00 - G02B26/00, G02B30/00
- G02B27/0025—Optical systems or apparatus not provided for by any of the groups G02B1/00 - G02B26/00, G02B30/00 for optical correction, e.g. distorsion, aberration
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/64—Fluorescence; Phosphorescence
- G01N21/645—Specially adapted constructive features of fluorimeters
- G01N21/6456—Spatial resolved fluorescence measurements; Imaging
- G01N21/6458—Fluorescence microscopy
-
- G—PHYSICS
- G02—OPTICS
- G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G02B27/00—Optical systems or apparatus not provided for by any of the groups G02B1/00 - G02B26/00, G02B30/00
- G02B27/0012—Optical design, e.g. procedures, algorithms, optimisation routines
Abstract
The invention provides a self-adaptive aberration correction method and an optical sheet microscopic imaging device based on the self-adaptive aberration correction method, wherein the self-adaptive aberration correction method comprises the following steps: establishing a Zernike polynomial characterization model of aberration; solving an approximately optimal solution of the Zernike polynomial representation model by using a genetic algorithm; further solving the global optimal solution of the Zernike polynomial representation model by using a random parallel gradient descent algorithm by taking the approximate optimal solution as a starting point; and calculating a phase to be corrected based on the global optimal solution, wherein the phase to be corrected is an aberration correction parameter. The self-adaptive aberration correction method has the advantages of large correctable aberration range and higher aberration correction speed.
Description
Technical Field
The invention relates to the technical field of optical imaging, in particular to a self-adaptive aberration correction method and an optical sheet microimaging device based on the self-adaptive aberration correction method.
Background
The light sheet microscope has the advantages of low bleaching property and high axial resolution for biological tissues, but the imaging depth of the light sheet microscope is limited, and the development of researches in the fields of biomedicine and the like puts forward higher requirements on the imaging depth of a modern microscope system. With the increase of the imaging depth, the aberration introduced by refractive index mismatch can seriously affect the microscopic imaging quality, and the type of aberration can change along with the change of the observation environment and the sample, so that the real-time dynamic correction of the aberration needs to be performed by introducing an adaptive optical technology.
The aberration range corrected by the existing adaptive optics technology is too small, and the aberration correction speed is too slow.
Disclosure of Invention
The invention solves the problems that the aberration range corrected by the existing adaptive optics technology is too small, and the aberration correction speed is too slow.
To solve the above problem, the present invention provides a method for adaptive aberration correction, comprising:
establishing a Zernike polynomial characterization model of aberration;
solving an approximately optimal solution of the Zernike polynomial representation model by using a genetic algorithm;
further solving the global optimal solution of the Zernike polynomial representation model by using a random parallel gradient descent algorithm by taking the approximate optimal solution as a starting point;
and calculating a phase to be corrected based on the global optimal solution, wherein the phase to be corrected is an aberration correction parameter.
Optionally, the zernike polynomial characterization model is a linear combination of zernike polynomials of order 5-36.
Optionally, said solving the near-optimal solution of the zernike polynomial representation model using a genetic algorithm comprises:
coding the Zernike coefficients in the Zernike polynomial characterization model, and taking the coded Zernike coefficients as gene values;
setting the number of groups to be M, wherein each group comprises N Zernike coefficients, and selecting N random numbers in an aberration correction range as each Zernike coefficient, wherein N is equal to the polynomial term number of the Zernike polynomial expression model;
based on Zernike coefficients of M populations, phase distribution corresponding to each population is sequentially calculated and sequentially applied to a wavefront corrector, and image evaluation functions under the phase distribution corresponding to each population are respectively calculated, wherein the image evaluation functions are fitness functions of the genetic algorithm;
reserving the individuals with the highest fitness in the M populations to directly enter the next generation, selecting individuals from the M populations based on a preset cross rule to perform cross operation to obtain new individuals, and selecting individuals from the M populations based on a preset variation rule to perform variation operation;
judging whether a preset iteration termination condition is met or not;
if the preset iteration termination condition is not met, returning to execute the step of sequentially calculating the phase distribution corresponding to each population based on the Zernike coefficients of the M populations;
and if the preset iteration termination condition is met, taking the individual with the highest fitness in the current population as the approximate optimal solution.
Optionally, the encoding the zernike coefficients in the zernike polynomial representation model comprises: and floating-point coding is carried out on the Zernike coefficients in the Zernike polynomial characterization model.
Optionally, the intersection operation is performed by using a uniform intersection operator, and the mutation operation is a gaussian mutation operation.
Optionally, the further solving the global optimal solution of the zernike polynomial representation model by using a stochastic parallel gradient descent algorithm with the approximately optimal solution as a starting point includes:
setting the near-optimal solution to a ═ a5,a6,a7,...a36]Generating a random vector [ a ] following a Bernoulli distribution5,a6,a7,...a36]；
Calculating a Zernike coefficient vector disturbed by the two parties according to a preset formula, wherein the preset formula is as follows:
a+＝a+＝[a5+a5,a6+a6,a7+a7,...a36+a36]，
a-＝a-＝[a5-a5,a6-a6,a7-a7,...a36-a36]，
by vector a+And a-Generating corresponding phase distributions psi + and psi-according to the Zernike polynomial characterization model, applying psi + and psi-to the wavefront corrector respectively, and calculating image evaluation functions J under psi + and psi-+And J-Wherein the image evaluation function J is the fitness function;
calculating the difference between the values of the two image evaluation functions, Δ J ═ J+-J-；
Updating the Zernike coefficient vector based on a difference between values of the two image merit functions;
judging whether a preset iteration termination condition is met or not;
if the preset iteration termination condition is not met, returning to execute the random vector which is generated and obeys the Bernoulli distribution;
and if the preset iteration termination condition is met, calculating corresponding phase distribution by the updated Zernike coefficient vector according to the Zernike polynomial representation model, wherein the phase distribution is the global optimal solution.
Optionally, the image evaluation function comprises an image gray scale variance function and/or an image gradient sum function.
The invention also provides an optical sheet microimaging device based on the self-adaptive aberration correction method, which comprises the following steps: a correction controller that calculates aberration correction parameters based on the adaptive aberration correction method as described in any one of the above, a spatial light modulator as a wavefront corrector for correcting aberrations based on the aberration correction parameters, and a camera for imaging a sample.
Optionally, the light sheet microscopic imaging device further includes an acousto-optic deflector, the spatial light modulator is further configured to generate defocused mode phase distributions with different coefficients to focus light at different depths of the sample, so as to implement longitudinal scanning of the sample, and the acousto-optic deflector is configured to deflect light at different angles, so as to implement transverse scanning of the sample.
Optionally, the light sheet micro-imaging device further comprises: the optical path of the second micro objective, the imaging lens and the sCMOS camera is orthogonal and vertical to the optical path of the acousto-optic deflector, the scanning lens, the collecting lens and the first micro objective.
The invention combines the genetic algorithm and the random parallel gradient descent algorithm, firstly solves the approximate optimal solution by utilizing the genetic algorithm, provides a better search starting point for the random parallel gradient descent algorithm, reduces the probability of local optimal problem of the random parallel gradient descent algorithm, i.e. the invention integrates the advantages of the genetic algorithm and the random parallel gradient descent algorithm, avoids the defects of the random parallel gradient descent algorithm and the random parallel gradient descent algorithm, can correct the complex aberration existing in a light sheet microscope system, has simple aberration correction operation and low cost, and achieves higher aberration correction speed and higher aberration correction precision.
Drawings
FIG. 1 is a schematic diagram of an adaptive aberration correction method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of another embodiment of an adaptive aberration correction method according to the present invention;
FIG. 3 is a schematic view of an embodiment of an optical sheet micro-imaging device according to the present invention;
FIG. 4 is a schematic view of another embodiment of the light sheet micro-imaging device of the present invention.
Description of reference numerals:
1-a laser; 2-a first shaping lens; 3-a second shaping lens; 4-a polarizing plate; 5-two-dimensional scanning system; 51-a spatial light modulator; 52-an acousto-optic deflector; 6-a scanning lens; 7-a collecting lens; 8-a first microscope objective; 9-sample; 10-a second microscope objective; 11-an imaging lens; 12-a camera; 13-correction controller.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
To facilitate an understanding of the present invention, a brief description of adaptive optics techniques will first be provided.
The adaptive optics technology mainly includes a direct wavefront detection technology that relies on a wavefront sensor and an indirect wavefront detection technology that does not rely on a wavefront sensor. The wavefront controller is used for calculating control parameters of the wavefront corrector according to a wavefront control algorithm and feeding the control parameters back to the wavefront corrector, the wavefront corrector is an executive device for aberration correction, and optical path differences of different positions of an incident beam are changed to carry out phase compensation of wavefront distortion based on the control parameters calculated by the wavefront controller so as to realize correction of the aberration.
The aberration and the fluorescence image have indirect corresponding relation, when the aberration is zero, the image evaluation index of the fluorescence image is optimal, therefore, the control parameter of the wave-front corrector is determined based on the change of the image evaluation index of the fluorescence image, the image quality is optimal, and the purpose of aberration correction is indirectly achieved.
The invention provides a self-adaptive aberration correction method.
Fig. 1 is a schematic diagram of an adaptive aberration correction method according to an embodiment of the present invention, as shown in fig. 1, the adaptive aberration correction method includes:
step S10, establishing a Zernike polynomial characterization model of the aberration;
the Zernike polynomial characterization model can characterize a large amount of aberration, the aberration is characterized by the sum of the Zernike polynomials, the problem of solving the optimal control voltage of the wavefront corrector can be converted into the problem of solving the Zernike coefficient of the aberration, and the difficulty of self-adaptive aberration correction is simplified.
Wherein, the Zernike polynomial characterization model of the aberration is as follows:
where r, θ are normalized polar coordinates of the pupil plane, i is the Zernike polynomial order, ψ (r, θ) is the wavefront phase distribution function, Zi(r, θ) is the ith zernike basis function, aiAre coefficients of the ith order zernike aberration.
Optionally, the zernike polynomial characterization model is a linear combination of zernike polynomials of order 5-36.
The 5-36 order Zernike polynomial has abundant information quantity which can represent a large amount of aberration, thereby ensuring the accuracy of aberration representation and improving the accuracy and the correction effect of aberration correction.
Step S20, solving the approximate optimal solution of the Zernike polynomial characterization model by using a genetic algorithm;
alternatively, referring to fig. 2, step S20 includes:
step S21, coding the Zernike coefficients in the Zernike polynomial characterization model, and taking the coded Zernike coefficients as gene values;
optionally, the zernike coefficients are floating point encoded. In one embodiment, the aberration is characterized by the sum of 5-36 order Zernike polynomials and its corresponding Zernike coefficient a5To a36As gene value [ a5,a6,a7,...a36]。
Step S22, setting the number of populations to be M, wherein each population comprises N Zernike coefficients, and selecting N random numbers in an aberration correction range as each Zernike coefficient, wherein N is equal to the polynomial term number of the Zernike polynomial expression model;
referring to fig. 2, step S22 is population initialization in fig. 2. Optionally, in one embodiment, the aberration is characterized by a sum of 5-36 order zernike polynomials, and a population comprises 32 zernike coefficients, and 32 random numbers are selected for each population as the zernike coefficients.
Alternatively, the aberration correction range is [ -2 π, +2 π ], from which a random number is taken.
Step S23, based on Zernike coefficients of M populations, sequentially calculating phase distribution corresponding to each population, sequentially applying the phase distribution to a wavefront corrector, and respectively calculating an image evaluation function under the phase distribution corresponding to each population, wherein the image evaluation function is a fitness function of the genetic algorithm;
step S23 is a fitness calculation operation in fig. 2. Based on the zernike polynomial representation model formula ψ (r, θ) of the aberration in step S10, the phase distribution corresponding to each population is calculated.
Optionally, the image evaluation function comprises an image gray scale variance function. The image gray variance describes the size of the gray distribution dispersion degree in the image, the larger the image gray variance is, the larger the representing gray distribution dispersion degree is, the richer the image gray level is, the clearer the image is, the smaller the image gray variance is, the smaller the gray dispersion degree is, and the more fuzzy the image is.
Optionally, the image evaluation function comprises an image gradient sum function. The image gradient information describes the definition degree of the image edge, the larger the gray gradient is, the clearer the edge is, the clearer the details of the whole image are, and the better the image quality is.
Optionally, the image evaluation function comprises an image gray variance function and an image gradient sum function.
And taking the image evaluation function as a fitness function of the genetic algorithm, wherein the higher the image quality determined based on the image evaluation function is, the higher the fitness of the image is.
Step S24, reserving the individuals with the highest fitness in the M populations to directly enter the next generation, selecting individuals from the M populations to perform cross operation based on a preset cross rule to obtain new individuals, and selecting individuals from the M populations to perform variation operation based on a preset variation rule;
in the previous generation population, the probability of the individual to be selected to enter the next generation is in direct proportion to the adaptation, and the higher the individual fitness is, the higher the probability of the individual to enter the next generation is. Wherein, the individual with the highest fitness can directly enter the next generation and exist as the individual of the next generation.
Referring to fig. 2, several individuals are selected from the previous generation, and crossover and mutation operations are performed based on the selected individuals. Alternatively, a preset number of individuals with the highest fitness, such as the first 20 individuals with the highest fitness, such as the first 5% of individuals with the highest fitness, may be selected from the M populations.
And performing cross operation on the selected individuals of the previous generation: setting a cross rate Pc, and selecting two individuals to be crossed from the selected individuals according to the cross rate Pc
Can generate new individuals according to the idea of uniform crossover operators
The formula is as follows:
wherein 0< a <1, 0< B <1, and a + B ═ 1;
optionally, the mutation operation is a gaussian mutation operation. Specifically, an individual to be mutated is selected from a plurality of selected individuals according to the mutation probability Pm, and a random number conforming to the mean value mu and the variance sigma 2 is used for replacing a gene value corresponding to the selected variant individual. The specific operation is as follows: first, 20 are generated in [0,1 ]]Random numbers r uniformly distributed in the rangei(i ═ 1,2, … 20), a random number R that conforms to a normal distribution of N (μ, σ 2) can be determined by the following equation:
in one embodiment, the quiltSelecting variant individuals participating in operation as 5-36 order Zernike coefficients [ a ]5,a6,a7,...ak...a36]Is provided with akIs a variation point and has a value range of [ Amin,Amax]At this point, the subject is subjected to Gaussian mutation by using [ mu ] - (A)min+Amax)/2，σ＝(Amax-Amin) 10, then individual [ a5,a6,a7,...ak'...a36]New gene value of (a)k' is:
step S25, judging whether a preset iteration termination condition is met;
optionally, the preset iteration termination condition is that the iteration number is greater than or equal to a first preset number.
Optionally, the preset iteration termination condition is to achieve a fitness function optimization goal, and at this time, the fitness function of each individual of the current population is calculated to obtain the fitness of each individual. For example, the fitness function is an image gray variance function, and the image gray variance is greater than or equal to a preset variance as an optimization target.
Step S26, if the preset iteration termination condition is not satisfied, returning to execute step S23, that is, returning to the fitness calculation step as shown in fig. 2, calculating the fitness of the new generation individual population, and continuing the iteration process;
step S27, if the preset iteration termination condition is satisfied, taking the individual with the highest fitness in the current population as the approximately optimal solution, and taking the individual with the highest fitness as the optimization starting point of the subsequent random parallel gradient descent algorithm (i.e., the SPGD algorithm in fig. 2).
Step S30, further solving the global optimal solution of the Zernike polynomial representation model by using a stochastic parallel gradient descent algorithm (stochastics parallel gradient device algorithm-SPGD) with the approximate optimal solution as a starting point.
Alternatively, referring to fig. 2, step S30 includes:
step S31, setting the near-optimal solution to a ═ a5,a6,a7,...a36]Generating a random vector [ a ] following a Bernoulli distribution5,a6,a7,...a36]The random vector is the perturbation vector in fig. 2;
and (4) taking the approximate optimal solution as a starting point of a random parallel gradient descent algorithm to solve the global optimal solution.
Step S32, calculating the Zernike coefficient vector disturbed by the two parties according to a preset formula, wherein the preset formula is as follows:
a+＝a+＝[a5+a5,a6+a6,a7+a7,...a36+a36]，
a-＝a-＝[a5-a5,a6-a6,a7-a7,...a36-a36]，
step S33, calculating the vector a+And a-Generating corresponding phase distributions psi + and psi-according to the Zernike polynomial characterization model, applying psi + and psi-to the wavefront corrector respectively, and calculating image evaluation functions J under psi + and psi-+And J-Wherein the image evaluation function J is the fitness function;
in step S34, the difference Δ J between the values of the two image evaluation functions is calculated as J+-J-；
A step S35 of updating the zernike coefficient vector based on a difference between values of the two image evaluation functions;
specifically, the zernike coefficient vector is updated based on the formula a- μ Δ J, where μ is the iteration step size to control the convergence speed of the algorithm.
Step S36, judging whether a preset iteration termination condition is met;
optionally, the preset iteration termination condition here is that the number of iterations is greater than or equal to a second preset number.
Optionally, the preset iteration termination condition is that a preset definition requirement is satisfied, specifically, an image evaluation function value requirement may be used as a condition, at this time, an image evaluation function under the latest zernike coefficient vector is calculated to obtain a current image evaluation value, and when the image evaluation value is greater than the preset value, it is determined that the preset iteration termination condition is satisfied. For example, the image evaluation function is an image gradient sum function, and the image gradient sum is greater than or equal to a preset gradient as an iteration termination condition.
And step S37, if the preset iteration termination condition is not met, returning to execute the step of generating the random vector obeying the Bernoulli distribution, namely continuing the iteration.
And step S38, if the preset iteration termination condition is met, calculating corresponding phase distribution by the updated Zernike coefficient vector according to the Zernike polynomial representation model, wherein the phase distribution is the global optimal solution.
Step S40, calculating a phase to be corrected based on the global optimal solution, wherein the phase to be corrected is an aberration correction parameter.
And the global optimal solution is a group of optimal Zernike coefficients, the phase to be corrected is calculated based on the optimal Zernike coefficients and a formula psi (r, theta) of the Zernike polynomial representation model, the phase to be corrected is fed back to the wavefront corrector, and aberration correction is carried out by the wavefront corrector.
The random parallel gradient descent algorithm is easy to fall into local optimization and has poor stability, the genetic algorithm is used as a global optimization algorithm and has poor local optimization capability and is easy to generate the phenomenon of 'precocity', the genetic algorithm and the random parallel gradient descent algorithm are combined, the genetic algorithm is firstly used for solving an approximate optimal solution, a better search starting point is provided for the random parallel gradient descent algorithm, the probability of the local optimization problem of the random parallel gradient descent algorithm is reduced, namely the embodiment of the invention integrates the advantages of the genetic algorithm and the random parallel gradient descent algorithm, avoids the defects of the genetic algorithm and the random parallel gradient descent algorithm, can correct the complex aberration in a light sheet microscope system, is simple in aberration correction operation and low in cost, and achieves higher aberration correction speed and higher aberration correction precision.
The invention further provides an optical sheet micro-imaging device based on the adaptive aberration correction method according to any of the above embodiments, and referring to fig. 3, the optical sheet micro-imaging device includes: a correction controller 13, a spatial light modulator 51 and a camera 12, wherein the correction controller 13 calculates aberration correction parameters based on the adaptive aberration correction method as described in the above embodiments, the spatial light modulator 51 functions as a wavefront corrector for correcting aberrations based on the correction parameters, and the camera 12 is used for imaging a sample.
Wherein the camera 12 may be a sCMOS camera or a CCD camera. sCMOS (complementary metal oxide semiconductor) is a new CMOS device obtained by improving the traditional CMOS chip manufacturing process and processing capacity.
The sample is imaged by the camera 12, the image evaluation index of the image obtained by imaging thereof is calculated by the correction controller 13, and the aberration correction parameter is calculated based on the image evaluation index and the adaptive aberration correction method as described in the above respective embodiments, and aberration correction is performed by the spatial light modulator 51.
Optionally, as shown in fig. 4, in another embodiment of the light sheet micro-imaging device provided by the present invention, the light sheet micro-imaging device further includes a two-dimensional scanning system 5, the two-dimensional scanning system 5 includes the spatial light modulator 51 and the acoustic light deflector 52, the spatial light modulator 51 is further configured to generate defocused mode phase distributions with different coefficients to focus light to different depths of the sample, so as to implement longitudinal scanning of the sample, and the acoustic light deflector 52 deflects light at different angles, so as to implement transverse scanning of the sample.
The spatial light modulator and the acousto-optic deflector jointly form a scanning device of the light sheet microscopic imaging device, the spatial light modulator can generate phase distribution in a defocusing mode to focus light at different depths of a sample, the spatial light modulator generates phase distribution with different focusing depths according to a certain time sequence in the scanning process to realize longitudinal scanning of the sample, the acousto-optic deflector can change the light propagation direction, and the acousto-optic deflector generates different deflection angles according to a certain time sequence in the scanning process to realize transverse scanning of the sample.
Optionally, as shown in fig. 4, the light sheet micro-imaging device further includes: the laser device comprises a laser device 1, a first shaping lens 2, a second shaping lens 3, a polaroid 4, a scanning lens 6, a collecting lens 7, a first microscope objective 8, a second microscope objective 10 and an imaging lens 11, wherein the light path of the second microscope objective 10, the light path of the imaging lens 11 and the light path of the camera 12 are in orthogonal and perpendicular relation with the light path of the acousto-optic deflector 52, the scanning lens 6, the collecting lens 7 and the light path of the first microscope objective 8.
Wherein, the laser 1 is used for illumination, the laser emitted by the laser 1 is expanded by the first shaping lens 2 and the second shaping lens 3 and then passes through the polaroid 4, the emergent light only contains the light in the incident polarization direction allowed by the spatial light modulator 51, the light incident to the spatial light modulator 51 is modulated by the SLM and then deflected by the acousto-optic deflector 52, then is incident to the first microscope objective 8 through the scanning lens 6 and the collecting lens 7, is focused to the sample 9 by the first microscope objective 8, the light reflected by the sample 9 is incident to the second microscope objective 10 and is focused to the camera 12 by the imaging lens 11 to record the light intensity, wherein, the plane formed by scanning the light spot incident through the acousto-optic deflector 52, the scanning lens 6, the collecting lens 7 and the first microscope objective 8 on the sample 9 is perpendicular to the light path of the second microscope objective 10, the imaging lens 11 and the camera 12.
The light sheet microimaging device is provided with a first microobjective 8 used for illuminating/exciting fluorescence and a second microobjective 10 used for detecting, wherein the first microobjective 8 and the second microobjective 10 are arranged in a mutually perpendicular mode, so that an incident illumination light path where the first microobjective 8 is located is mutually perpendicular to a fluorescence emission light path where the second microobjective 10 is located, and therefore the area needing to be observed is ensured to be selectively excited, ineffective exposure is reduced, and photobleaching and phototoxicity are remarkably reduced.
Although the present disclosure has been described above, the scope of the present disclosure is not limited thereto. Various changes and modifications may be effected therein by one of ordinary skill in the pertinent art without departing from the spirit and scope of the present disclosure, and these changes and modifications are intended to be within the scope of the present disclosure.
Claims (10)
1. An adaptive aberration correction method, comprising:
establishing a Zernike polynomial characterization model of aberration;
solving an approximately optimal solution of the Zernike polynomial representation model by using a genetic algorithm;
further solving the global optimal solution of the Zernike polynomial representation model by using a random parallel gradient descent algorithm by taking the approximate optimal solution as a starting point;
and calculating a phase to be corrected based on the global optimal solution, wherein the phase to be corrected is an aberration correction parameter.
2. The adaptive aberration correction method of claim 1, wherein the zernike polynomial characterization model is a linear combination of zernike polynomials of order 5-36.
3. The adaptive aberration correction method of claim 1 or 2, wherein said solving a near-optimal solution of said zernike polynomial representation model using a genetic algorithm comprises:
coding the Zernike coefficients in the Zernike polynomial characterization model, and taking the coded Zernike coefficients as gene values;
setting the number of groups to be M, wherein each group comprises N Zernike coefficients, and selecting N random numbers in an aberration correction range as each Zernike coefficient, wherein N is equal to the polynomial term number of the Zernike polynomial expression model;
based on Zernike coefficients of M populations, phase distribution corresponding to each population is sequentially calculated and sequentially applied to a wavefront corrector, and image evaluation functions under the phase distribution corresponding to each population are respectively calculated, wherein the image evaluation functions are fitness functions of the genetic algorithm;
reserving the individuals with the highest fitness in the M populations to directly enter the next generation, selecting individuals from the M populations based on a preset cross rule to perform cross operation to obtain new individuals, and selecting individuals from the M populations based on a preset variation rule to perform variation operation;
judging whether a preset iteration termination condition is met or not;
if the preset iteration termination condition is not met, returning to execute the step of sequentially calculating the phase distribution corresponding to each population based on the Zernike coefficients of the M populations;
and if the preset iteration termination condition is met, taking the individual with the highest fitness in the current population as the approximate optimal solution.
4. The adaptive aberration correction method according to claim 3, wherein said encoding the Zernike coefficients in the Zernike polynomial characterization model comprises:
and floating-point coding is carried out on the Zernike coefficients in the Zernike polynomial characterization model.
5. The adaptive aberration correction method according to claim 3, wherein the crossover operation is performed using a uniform crossover operator, and the mutation operation is a Gaussian mutation operation.
6. The adaptive aberration correction method according to claim 3, wherein said further solving the globally optimal solution of the Zernike polynomial representation model using a stochastic parallel gradient descent algorithm starting from the near optimal solution comprises:
setting the near-optimal solution to a ═ a5,a6,a7,...a36]Generating a random vector [ a ] following a Bernoulli distribution5,a6,a7,...a36]；
Calculating a Zernike coefficient vector disturbed by the two parties according to a preset formula, wherein the preset formula is as follows:
a+＝a+＝[a5+a5,a6+a6,a7+a7,...a36+a36]，
a-＝a-＝[a5-a5,a6-a6,a7-a7,...a36-a36]，
by vector a+And a-Generating corresponding phase distributions psi + and psi-according to the Zernike polynomial characterization model, applying psi + and psi-to the wavefront corrector respectively, and calculating image evaluation functions J under psi + and psi-+And J-Wherein the image evaluation function J is the fitness function;
calculating the difference between the values of the two image evaluation functions, Δ J ═ J+-J-；
Updating the Zernike coefficient vector based on a difference between values of the two image merit functions;
judging whether a preset iteration termination condition is met or not;
if the preset iteration termination condition is not met, returning to execute the random vector which is generated and obeys the Bernoulli distribution;
and if the preset iteration termination condition is met, calculating corresponding phase distribution by the updated Zernike coefficient vector according to the Zernike polynomial representation model, wherein the phase distribution is the global optimal solution.
7. An adaptive aberration correction method according to any of claims 4-6, characterized in that the image evaluation function comprises an image grey scale variance function and/or an image gradient sum function.
8. An optical sheet microscopy imaging device based on the adaptive aberration correction method according to any one of claims 1 to 7, comprising: a correction controller (13) that calculates aberration correction parameters based on the adaptive aberration correction method according to any one of claims 1 to 7, a spatial light modulator (51) that functions as a wavefront corrector for correcting aberrations based on the aberration correction parameters, and a camera (12) for imaging a sample.
9. The light sheet microscopic imaging device according to claim 8, further comprising an acousto-optic deflector (52), wherein the spatial light modulator (51) is further configured to generate different coefficients of phase distribution of the defocused mode to focus light at different depths of the sample, thereby realizing longitudinal scanning of the sample, and the acousto-optic deflector (52) is configured to deflect light at different angles, thereby realizing transverse scanning of the sample.
10. The light sheet microscopy imaging device of claim 9, further comprising: the laser device comprises a laser device (1), a first shaping lens (2), a second shaping lens (3), a polaroid (4), a scanning lens (6), a collecting lens (7), a first microscope objective (8), a second microscope objective (10) and an imaging lens (11), wherein the light path of the second microscope objective (10), the imaging lens (11) and the camera (12) is orthogonal and vertical to the light path of the acousto-optic deflector (52), the scanning lens (6), the collecting lens (7) and the first microscope objective (8).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010587864.3A CN111580271B (en) | 2020-06-24 | 2020-06-24 | Self-adaptive aberration correction method and light sheet microscopic imaging device based on same |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010587864.3A CN111580271B (en) | 2020-06-24 | 2020-06-24 | Self-adaptive aberration correction method and light sheet microscopic imaging device based on same |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111580271A true CN111580271A (en) | 2020-08-25 |
| CN111580271B CN111580271B (en) | 2022-02-18 |
Family
ID=72111320
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010587864.3A Active CN111580271B (en) | 2020-06-24 | 2020-06-24 | Self-adaptive aberration correction method and light sheet microscopic imaging device based on same |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111580271B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327327A (en) * | 2020-11-05 | 2021-02-05 | 哈尔滨工业大学 | Laser radar imaging quality detection method and system based on genetic algorithm |
| CN112601015A (en) * | 2020-12-10 | 2021-04-02 | 烟台艾睿光电科技有限公司 | Automatic focusing method and automatic focusing system |
| CN114545621A (en) * | 2022-03-15 | 2022-05-27 | 中国科学院光电技术研究所 | Deformed reflector structure optimization method based on genetic algorithm |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504229A (en) * | 2014-09-19 | 2015-04-08 | 杭州电子科技大学 | Intelligent bus scheduling method based on hybrid heuristic algorithm |
| US9134538B1 (en) * | 2013-02-06 | 2015-09-15 | Massachusetts Institute Of Technology | Methods, systems, and apparatus for coherent beam combining |
| CN105589210A (en) * | 2016-03-10 | 2016-05-18 | 中国科学院光电技术研究所 | Digital synthetic aperture imaging method based on pupil modulation |
| CN106873152A (en) * | 2017-01-11 | 2017-06-20 | 浙江大学 | A kind of high speed aberration correcting method based on machine learning |
| CN107506723A (en) * | 2017-08-21 | 2017-12-22 | 河海大学 | A kind of deficient based on artificial neural network and genetic algorithms determines Mixture matrix method of estimation |
| CN107894662A (en) * | 2017-11-08 | 2018-04-10 | 北京邮电大学 | A kind of improvement SPGD algorithms for OAM beam phase reparations |
| CN110309482A (en) * | 2019-05-16 | 2019-10-08 | 中国科学院西安光学精密机械研究所 | A kind of rapid convergence, high-precision phase recovery method |
| CN110349095A (en) * | 2019-06-14 | 2019-10-18 | 浙江大学 | Learn the adaptive optics wavefront compensation method of prediction wavefront zernike coefficient based on depth migration |
| CN110389119A (en) * | 2019-07-19 | 2019-10-29 | 浙江大学 | Quick self-adapted optical scanner micro imaging system and method based on machine learning |
| CN110567959A (en) * | 2019-09-17 | 2019-12-13 | 哈工大机器人(中山)无人装备与人工智能研究院 | Self-adaptive aberration correction image scanning microscopic imaging method and device |
| CN110824697A (en) * | 2019-11-21 | 2020-02-21 | 重庆工商大学 | Self-adaptive optical system combining artificial beacon and wavefront-free detection |
| CN110941100A (en) * | 2019-11-29 | 2020-03-31 | 哈尔滨工业大学 | Multi-photon excitation combined multi-mode array type scanning imaging device |
| CN111008945A (en) * | 2019-12-31 | 2020-04-14 | 哈工大机器人(中山)无人装备与人工智能研究院 | Multi-image-quality-parameter self-adaptive aberration correction method and device based on machine learning |
| CN111077085A (en) * | 2019-12-31 | 2020-04-28 | 哈工大机器人(中山)无人装备与人工智能研究院 | Wavefront-free detector self-adaptive point scanning imaging method based on deep learning |
-
2020
- 2020-06-24 CN CN202010587864.3A patent/CN111580271B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9134538B1 (en) * | 2013-02-06 | 2015-09-15 | Massachusetts Institute Of Technology | Methods, systems, and apparatus for coherent beam combining |
| CN104504229A (en) * | 2014-09-19 | 2015-04-08 | 杭州电子科技大学 | Intelligent bus scheduling method based on hybrid heuristic algorithm |
| CN105589210A (en) * | 2016-03-10 | 2016-05-18 | 中国科学院光电技术研究所 | Digital synthetic aperture imaging method based on pupil modulation |
| CN106873152A (en) * | 2017-01-11 | 2017-06-20 | 浙江大学 | A kind of high speed aberration correcting method based on machine learning |
| CN107506723A (en) * | 2017-08-21 | 2017-12-22 | 河海大学 | A kind of deficient based on artificial neural network and genetic algorithms determines Mixture matrix method of estimation |
| CN107894662A (en) * | 2017-11-08 | 2018-04-10 | 北京邮电大学 | A kind of improvement SPGD algorithms for OAM beam phase reparations |
| CN110309482A (en) * | 2019-05-16 | 2019-10-08 | 中国科学院西安光学精密机械研究所 | A kind of rapid convergence, high-precision phase recovery method |
| CN110349095A (en) * | 2019-06-14 | 2019-10-18 | 浙江大学 | Learn the adaptive optics wavefront compensation method of prediction wavefront zernike coefficient based on depth migration |
| CN110389119A (en) * | 2019-07-19 | 2019-10-29 | 浙江大学 | Quick self-adapted optical scanner micro imaging system and method based on machine learning |
| CN110567959A (en) * | 2019-09-17 | 2019-12-13 | 哈工大机器人(中山)无人装备与人工智能研究院 | Self-adaptive aberration correction image scanning microscopic imaging method and device |
| CN110824697A (en) * | 2019-11-21 | 2020-02-21 | 重庆工商大学 | Self-adaptive optical system combining artificial beacon and wavefront-free detection |
| CN110941100A (en) * | 2019-11-29 | 2020-03-31 | 哈尔滨工业大学 | Multi-photon excitation combined multi-mode array type scanning imaging device |
| CN111008945A (en) * | 2019-12-31 | 2020-04-14 | 哈工大机器人(中山)无人装备与人工智能研究院 | Multi-image-quality-parameter self-adaptive aberration correction method and device based on machine learning |
| CN111077085A (en) * | 2019-12-31 | 2020-04-28 | 哈工大机器人(中山)无人装备与人工智能研究院 | Wavefront-free detector self-adaptive point scanning imaging method based on deep learning |
Non-Patent Citations (6)
| Title |
|---|
| 何杰铃: "基于自适应光学的激光束整形系统研究", 《中国博士学位论文全文数据库》 * |
| 张素文: "一种基于GA-BP混合算法的模糊神经网络控制器", 《电气自动化》 * |
| 朱明: "《数据挖掘第2版》", 30 November 2008 * |
| 杨平: "基于泽尼克模式系数的自适应光学遗传算法", 《中国激光》 * |
| 杨慧珍: "自适应光学系统随机并行下降控制算法仿真与分析", 《光学学报》 * |
| 陶文华: "基于混合遗传算法的模糊神经网络控制器", 《中国人工智能学会第9届全国学术年会暨中国人工智能学会成立二十周年庆祝大会论文集》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327327A (en) * | 2020-11-05 | 2021-02-05 | 哈尔滨工业大学 | Laser radar imaging quality detection method and system based on genetic algorithm |
| CN112327327B (en) * | 2020-11-05 | 2022-08-05 | 哈尔滨工业大学 | Laser radar imaging quality detection method and system based on genetic algorithm |
| CN112601015A (en) * | 2020-12-10 | 2021-04-02 | 烟台艾睿光电科技有限公司 | Automatic focusing method and automatic focusing system |
| CN114545621A (en) * | 2022-03-15 | 2022-05-27 | 中国科学院光电技术研究所 | Deformed reflector structure optimization method based on genetic algorithm |
| CN114545621B (en) * | 2022-03-15 | 2023-09-22 | 中国科学院光电技术研究所 | Deformed reflector structure optimization method based on genetic algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111580271B (en) | 2022-02-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111580271B (en) | Self-adaptive aberration correction method and light sheet microscopic imaging device based on same | |
| WO2014041660A1 (en) | Optical modulation control method, control program, control device, and laser light irradiation device | |
| JP4790567B2 (en) | Aberration measurement method, aberration correction method and electron microscope using Ronchigram | |
| JP5749553B2 (en) | Light modulation control method, control program, control device, and laser light irradiation device | |
| US20150323787A1 (en) | System, method and computer-accessible medium for depth of field imaging for three-dimensional sensing utilizing a spatial light modulator microscope arrangement | |
| JP2016532883A (en) | Method of correcting imaging aberrations with laser scanning microscopes and in particular with high resolution scanning microscopy | |
| US11971368B2 (en) | Determination method, elimination method and apparatus for electron microscope aberration | |
| CN107121775B (en) | Large-field-angle multilayer conjugate adaptive optical focusing and microscopic system and method | |
| JP5188846B2 (en) | Aberration correction apparatus and aberration correction method for scanning transmission electron microscope | |
| CN109212735A (en) | High-speed adaptive optics ring-shaped light spot based on machine learning corrects system and method | |
| TW202129680A (en) | Method of determining aberrations in images obtained by a charged particle beam tool, method of determining a setting of a charged particle beam tool, and charged particle beam tool | |
| CN111008945A (en) | Multi-image-quality-parameter self-adaptive aberration correction method and device based on machine learning | |
| JP4615886B2 (en) | Scanning optical microscope | |
| JP4207232B2 (en) | Charged beam exposure system | |
| KR20200077429A (en) | Optical alignment correction using convolutional neural network evaluation of a beam image | |
| JP2022512244A (en) | Real-time detection and correction of system response | |
| TW202242800A (en) | Optical correction coefficient prediction method, optical correction coefficient prediction device, machine learning method, machine learning preprocessing method, and trained learning model | |
| CN115356839A (en) | Bessel two-photon microscope illumination light path biological tissue aberration detection method | |
| Yang et al. | An adaptive laser beam shaping technique based on a genetic algorithm | |
| CN111819500B (en) | Ultrafast laser manufacturing method and system | |
| CN114556182A (en) | Method and device for aberration correction in fluorescence microscopy | |
| WO2022113366A1 (en) | Method for generating trained model, image processing method, image transforming device, and program | |
| Yang et al. | Remote focusing with dynamic aberration elimination by model-based adaptive optics | |
| Li et al. | Experimental Study on Wavefront Distortion Correction of 40-Elements Adaptive Optical System Based on SPGD Algorithm | |
| TWI802893B (en) | Method and system for eliminating zero-order diffraction light |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |