CN114677447A

CN114677447A - Optical microscopic imaging redundant information quantitative measurement method

Info

Publication number: CN114677447A
Application number: CN202210279435.9A
Authority: CN
Inventors: 潘安; 高慧琴
Original assignee: XiAn Institute of Optics and Precision Mechanics of CAS
Current assignee: Wuxi Xiguang Health Industry Co.,Ltd.
Priority date: 2022-03-21
Filing date: 2022-03-21
Publication date: 2022-06-28
Anticipated expiration: 2042-03-21
Also published as: CN114677447B

Abstract

The invention discloses a quantitative measurement method for redundant information of optical microscopic imaging, and further relates to a method for quantitatively measuring the redundant information required by Fourier laminated imaging (FPM) reconstruction. The method defines the imaging flux limit of the FPM technology, provides theoretical guidance for selecting experimental parameters, and provides important quantitative reference basis for related redundant information-based imaging technology, information multiplexing technology and the like.

Description

Optical microscopic imaging redundant information quantitative measurement method

Technical Field

The invention belongs to the technical field of optical information acquisition and processing, and particularly relates to a quantitative measurement method for redundant information of optical imaging, and further relates to a method for quantitatively measuring the redundant information required by reconstruction of a Fourier laminated imaging technology.

Background

The accurate English name of Fourier transform tomography (FPM) technology is named as Fourier transform graphics (FPM) technology, and the FPM technology is invented by Yang et al, California institute of technology, U.S. A, in 2013 and reported in (Zheng G, Horstmeyer R, Yang C.wide-field, high-resolution Fourier transform graphics [ J) ].Nature Photonics,2013,7(9):739-745.]Chinese translation has a number of names such as Fourier ptychographic imagingFourier stack microscopy, fourier stack imaging, fourier stack technique, etc., the term fourier stack imaging being used generically herein. The FPM technology is a promising new generation of computational optical imaging technology, and improves the original Abbe far-field diffraction limit formula lambda/2 NA into lambda/(NA)_illu+NA_obj) Wherein NA is_illuAnd NA_objThe illumination and objective Numerical Apertures (NA) are represented, respectively, and are written in the fourier optical introduction of the professor Goodman (fourth edition). Compared with the traditional bright field microscopy, the FPM technology combines the optical phase recovery technology and the microwave synthetic aperture technology, can realize high resolution and quantitative phase imaging while realizing a large field of view by using a low NA objective lens, and further can realize higher imaging flux.

For an optical system, the temporal resolution, the spatial resolution and the size of a field of view are contradictory trade-off relationships, and the number of pixels distinguishable in a unit time, i.e., the number of empty bandwidth products or the number of effective pixels, is used to describe the flux in units of pixels in academic practice. High imaging throughput (high throughput) defines the number of more significant pixels resolvable per unit time.

The FPM technique requires redundant information of frequency domain overlapping to restore a high resolution image. At present, research on redundant information is being conducted in two aspects of how to improve data redundancy utilization and reduce unnecessary data redundancy, thereby achieving higher imaging throughput. The former results include the LED multiplexing method [ Tian L, Li X, Ramchandran K, et al, multiplexed coded amplification for Fourier transform with LED array microprocessor [ J ]. biometrical Optics Express 2014,5(7): 2376; ], the single exposure method [ Pan A, Shen C, Yao B, et al, Single-shot Fourier transform with optoelectronic semiconductor video camera [ C ]// front in Optics/Laser, proceedings of OSA,2019: FTh3F.4 ] ], the deep learning method [ Xue Y, Cheng S, Li Y, real. transform device-array 618 ] 2019, FTh3F.4 ] ], the deep learning method [ Xue Y, Cheng S, Li, Y, real. prediction-array 2019: FTh3F.4 ] ], the deep learning method [ Xue Y, J.J. ], N.J. ], 14 J.J. ], the latter is aimed at reducing unnecessary data redundancy, such as with sparse LED arrays [ Sun J, Chen Q, Zhang Y, et al. sampling criteria for Fourier transform in object space and frequency space [ J ]. Optics Express,2016,24(14):15765 15781 ], lossy compression algorithms [ Bian L, Suo J, Situ G, et al. content adaptation for Fourier transform [ J ]. Optics Letters,2014,39(23): 6648-. Therefore, the current research knowledge on the information redundancy of the FPM technology and how different parameters influence the redundancy information only comes from the experimental experience of a specific sample under a specific algorithm, just like ' blind people's feeling ', the definition of the redundancy information is still an abstract concept, the quantitative measurement research of a theoretical level is lacked, the definition of the redundancy information is not clear, and the great challenge is brought to the further exploration and promotion of the imaging flux and the imaging efficiency of the current technology.

Disclosure of Invention

The invention aims to provide a general optical imaging redundancy information quantitative measurement method based on Shannon information theory and a redundancy information quantitative measurement method in Fourier laminated imaging technology, and the method solves the problem that in the prior art, the redundancy information lacks theoretical definition and the measurement method causes the imaging flux limit to be unclear.

In order to realize the technical task, the invention adopts the following technical scheme to realize:

a quantitative measurement method for redundant information of optical microscopy imaging comprises the following steps:

step 1: the optical imaging system obtains a coding image sequence set of a sample to be detected through a coder;

step 2: constructing a redundancy measurement model;

and step 3: a series of obtained coded image sequence sets are subjected to a redundancy measurement model to obtain the lumped redundancy information quantity of the coded image sequences and the complexity of a sample to be detected, namely a two-dimensional information entropy; calibrating the utilization rate of a reconstruction algorithm for a decoder by utilizing a redundancy measurement model;

and 4, step 4: completing redundancy judgment through the lumped redundancy information quantity of the coded image sequence, the complexity of a sample to be detected and the utilization rate of a calibrated reconstruction algorithm to obtain an imaging system parameter evaluation scheme;

And 5: outputting, by a decoder, the target image if the imaging system parameter evaluation scheme is executable.

Further, step 5 of the present invention further comprises: and if the parameter evaluation scheme of the imaging system cannot be executed, obtaining the optimized parameter scheme of the imaging system through the optimization module.

Further, the step 2 of constructing a redundancy measurement model comprises the following steps:

αQ≥η

wherein alpha is the utilization rate of the reconstruction algorithm, Q is the lumped redundant information quantity of the coding image sequence, and eta is the complexity of the sample to be measured.

Step 3, respectively calculating the information quantity of each image in the image sequence set to obtain the lumped redundant information quantity of the coding image sequence; calculating the complexity of the sample to be detected by utilizing the two-dimensional information entropy of the image; and (3) operating the redundancy measurement model for multiple times according to the types of different samples to be detected to obtain the inherent algorithm utilization rate based on a specific reconstruction algorithm, and the inherent algorithm utilization rate is used for representing the capability of the reconstruction algorithm related to the decoder for extracting information from the coded image sequence set.

Conventionally, in step 3: for a general optical imaging system, the calculation formula of the lumped redundant information quantity of the coded image sequence is as follows:

wherein: q denotes the total information content of the image sequence set, R denotes the weight for generating redundant information from the encoded image sequence set, which can be expressed in specific different forms such as frequency domain or spatial domain overlapping rate, sampling rate, etc., thereby generating redundant information of a specific domain, Q _iQuantity of information, SNR, characterizing each image_iThe signal-to-noise ratio is shown, and C is the bandwidth of the optical system and is determined by the corresponding parameters of the imaging system.

In particular, in step 3: for the FPM imaging system, the calculation formula of the lumped redundant information quantity of the coding image sequence is as follows:

wherein Q represents the amount of lumped redundant information of the sequence of encoded images, Q_iQuantity of information, R, characterizing each image_overlapRepresenting the frequency domain overlap ratio, SNR_iRepresenting the signal-to-noise ratio, R_camRepresenting the spatial sampling rate and W representing the optical system bandwidth.

Further, the optical system bandwidth W is calculated as follows:

wherein: λ represents wavelength, NA_objRepresenting the objective lens numerical aperture.

In step 3 of the present invention: for a general optical imaging system, the complexity calculation formula of the sample to be measured is as follows:

wherein p is_ijF (i, j)/MN (j is more than or equal to 0 and less than or equal to 255) represents the frequency of the appearance of the image feature binary f (i, j), MN is the image size, i is more than or equal to 0 and less than or equal to 255 is the gray value of the pixel, and j is more than or equal to 0 and less than or equal to 255 is the neighborhood gray average value.

In the step 3: for the FPM imaging system, the complexity calculation formula of the sample to be measured is as follows:

wherein p is_ijF (I, j)/MN (j is more than or equal to 0 and less than or equal to 255), the frequency of the occurrence of the image characteristic binary f (I, j) of the central bright field defocused image I of the sample to be detected is represented, MN is the image size, I is more than or equal to 0 and less than or equal to 255 is the gray value of a pixel, and j is more than or equal to 0 and less than or equal to 255 is the neighborhood gray average value; h _IIndicating that the centre contains phase informationThe information entropy of the out-of-focus image of the bright field image, I, represents the central bright field out-of-focus image containing phase information.

In the step 3:

wherein

Indicating LED_m,nAnd its corresponding illumination wave vector (k)_x,m,n,k_y,m,n) A transmitted light wave field through the sample object function o (x, y); j is an imaginary unit;

representing a quadratic phase factor of the simulation; z is defocus; λ is the illumination wavelength.

Preferably, the defocusing amount z ranges from 1 to 10 μm.

The calculation process of the utilization rate of the standard reconstruction algorithm in the step 3 comprises the following steps: and calculating a reconstruction algorithm utilization rate alpha value under a reconstruction limit by substituting a plurality of groups of samples to be measured of different classes into the redundancy measurement model, wherein when values of the reconstruction algorithm utilization rate alpha values corresponding to the plurality of groups of samples to be measured are all similar, the alpha value is the inherent reconstruction algorithm utilization rate of the reconstruction algorithm.

The reconfiguration limit condition setting process includes: and (3) taking a limit value of the airspace sampling rate, adjusting the light source illumination height and the number of the LED arrays, controlling the illumination NA, and obtaining a reconstruction limit condition when the quality of the reconstructed image meets the set requirement.

In the limit of reconstruction, the illumination NA is preferably 0.6. The root mean square error value of the intensity characterizing the quality requirement of the reconstructed image is preferably 0.01, and the minimum root mean square error of the phase is preferably 0.1.

Further, in step 4, the redundancy judgment process is as follows:

substituting the determined image sequence lumped redundancy information quantity, the complexity of the sample to be detected and the utilization ratio value of the calibration reconstruction algorithm into a redundancy measurement model when the alpha Q valueWhen the value is more than or equal to the sample complexity eta, the high-resolution image can be accurately reconstructed by the current sample under the imaging system parameter, namely, the imaging system parameter evaluation scheme is executable; when the value of alpha Q is smaller than the sample complexity eta, the information quantity provided by the current system is represented, the reconstruction requirement is not met, accurate reconstruction cannot be carried out, and the imaging system parameter evaluation scheme cannot be executed; when the value of alpha Q is equal to eta, the redundancy information quantity Q obtained by calculation at the moment_limitA limit value for the amount of information needed to reconstruct accurately for the current sample.

And 4, obtaining the feasibility of whether the imaging system parameters can realize the reconstruction of the high-resolution image by combining the redundancy judgment according to the imaging system parameter evaluation scheme, and marking the specific imaging system parameters which can not realize the reconstruction of the high-resolution image.

The imaging system parameters comprise one or more of wavelength, objective lens NA, illumination NA, spatial sampling rate, frequency domain overlapping rate, image sequence set, noise level, and sample complexity and algorithm utilization rate.

And 4, the parameter evaluation scheme of the imaging system also comprises the steps of outputting the space bandwidth product limit (SBP) of the imaging system,

wherein: FN is the number of views, mag is the magnification, NA_synTo synthesize the numerical aperture, i.e. the numerical aperture NA of the objective lens_objAnd the illumination numerical aperture NA_illuAnd, NA_sys＝NA_obj+NA_illu。

In the step 5, the optimization module utilizes the redundancy measurement model to obtain the optimization scheme of one or more specific imaging system parameters including wavelength, objective lens NA, illumination NA, spatial sampling rate, frequency domain overlapping rate, image sequence set, noise level, sample complexity and algorithm utilization rate.

The reconstruction algorithm related to the invention selects one or more of a Gauss-Newton algorithm, a GS algorithm, an EPRY-FPM algorithm, a wavelength multiplexing algorithm, an LED multiplexing algorithm or a reconstruction algorithm based on a deep learning model.

The invention also provides an electronic device, comprising a processor and a memory;

the memory is used for storing;

the processor is used for executing the disclosed quantitative measuring method for the redundant information of the micro-optical imaging.

Meanwhile, the method related to the present invention may also form a computer program product, which includes a computer program and/or instructions, and the computer program and/or instructions are executed by a processor to execute the steps of the disclosed method for quantitatively measuring redundant information in micro-optical imaging.

The technical scheme adopted by the invention is that a Shannon information theory thought in a general communication system model is transferred to a general optical imaging system, an optical microscopic imaging is combined, a redundant information measurement model is established particularly aiming at the influence of the selection of FPM imaging system parameters on an experimental result, the influence factors comprise imaging system parameters such as wavelength, objective lens NA, spatial sampling rate, frequency domain overlapping rate, quantity of non-repeated images, noise level (signal-to-noise ratio), sample complexity and algorithm utilization rate, wherein the imaging parameters participate in the calculation of the redundant information quantity, the algorithm utilization rate and the sample complexity form constraint conditions, and when the algorithm utilization rate multiplied by the redundant information quantity is more than or equal to the sample complexity, a sample can be accurately reconstructed. If and only if the constraint takes a limit, the corresponding amount of redundant information is the limit of redundant information that the imaging system accurately reconstructs for the sample.

Compared with the prior art, the invention has the following beneficial effects:

1) a set of quantitative measurement method for general optical microscopic imaging and further aiming at redundant information required by an FPM imaging system is provided, a theoretical basis is provided for selection of optimal experimental parameters, the imaging flux limit of the technology is determined, the redundant information can be utilized to the maximum extent, the imaging flux is improved, and even the limit is approached;

2) Before the actual experiment parameters are determined but the experiment is not carried out, whether the sample can be accurately reconstructed can be definitely judged, namely whether the parameters of the current imaging system are executable or not is evaluated, so that the experiment time is saved, and the parameters of which the evaluation results are non-executable are explained and optimized;

3) the method also provides quantitative reference basis for related redundant information-based imaging technology, information multiplexing technology and the like.

Drawings

Fig. 1 is a flow chart of a method for quantitatively measuring redundant information in optical microscopy imaging.

Fig. 2 is a theoretical model diagram based on shannon information theory according to the present invention, wherein fig. 2a is a general communication system model, and fig. 2b is a fourier transform imaging system model;

FIG. 3 is a flow diagram of a redundant information metrics module implementation.

FIG. 4 is a flow chart of a Gaussian Newton reconstruction algorithm employed in the exemplary embodiment;

FIG. 5 shows the calibration results under the extreme conditions of the simulation experiment under a 2X/0.08 NA objective lens. (fig. 5a1, fig. 5b1) are the true intensity and phase information of the sample, respectively, (fig. 5a2, fig. 5b2) are the high-resolution intensity and phase maps reconstructed by the algorithm under the limit condition, respectively, (fig. 5a3, fig. 5b3) are the mean square error of the intensity and the mean square error of the phase varying with the height h, respectively, and the marked points are the height and the mean square error result corresponding to the selected limit position;

FIG. 6 is the calibration result of the simulation experiment limit condition corresponding to 4X/0.1 NA objective lens;

FIG. 7 is the calibration result of the simulation experiment limit condition corresponding to the objective lens with 10X/0.25 NA;

FIG. 8 shows the results of the calibration of the experimental limit conditions for USAF resolution plates and cotton leaf cross-section samples under 4X/0.1 NA objective. (fig. 8a1, fig. 8b1) are respectively experimentally collected USAF and cotton leaf transverse normal incidence Low Resolution (LR) central bright field images, (fig. 8a 2-8 a4) are reconstructed high resolution intensity images and corresponding redundant information quantity Q at 39mm, 41mm, 43mm heights, (fig. 8b 2-8 b4) are reconstructed high resolution intensity images and corresponding redundant information quantity Q at 43mm, 45mm, 48mm heights, and the frame highlights the data set where the calibrated redundancy limit is located;

FIG. 9 is a graph of simulated noise level (Gaussian noise within 10%) versus algorithm utilization noise fluctuation;

FIG. 10 is a summary of the calibration results for the redundancy limits under different sets of experimental conditions, where the last column is the statistical result for the algorithm utilization under different noise;

FIG. 11 is the utilization of other reconstruction algorithms that have been calibrated;

FIG. 12 is a graph of experimental results of a redundant quantitative metrology model explaining the USAF resolution plate sample data at 80mm height under a 2X/0.1 NA objective. Fig. 12a1 is a low resolution central bright field image acquired experimentally, the spectrum of the high resolution image reconstructed in fig. 12a2, (fig. 12b1, fig. 12b2) are the intensity and phase images reconstructed, respectively, (fig. 12c1, fig. 12c2) are the pupil intensity and phase images reconstructed, respectively.

Detailed Description

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

Referring to fig. 1, the present invention provides a general quantitative measurement method for redundant information in optical microscopy imaging, which comprises the following steps:

step 2: constructing a redundancy measurement model;

and step 3: a series of obtained coded image sequence sets are subjected to a redundancy measurement model to obtain the lumped redundancy information quantity of the coded image sequences and the complexity of a sample to be detected; utilizing a redundancy measurement model to calibrate the utilization rate of the algorithm of the decoder;

and 4, step 4: completing redundancy judgment through the lumped redundancy information quantity of the coded image sequence, the complexity of a sample to be detected and the utilization rate of a calibration reconstruction algorithm to obtain an imaging system parameter evaluation scheme;

Furthermore, step 5 further comprises: and if the imaging system parameter evaluation scheme can not be executed, obtaining the optimized parameter scheme of the imaging system through the optimization module.

The following describes the technical principles of the present invention in detail:

the Technical conception and the implementation process of the invention are based on Shannon information theory (Shannon C. agricultural chemical the order of communication. the Bell System Technical Journal,1948,27(3): 379-. ]. As shown in fig. 2a, in particular, the general communication process can be described as the message is modulated or encoded by the transmitter to form a signal, the signal affected by noise during transmission is received by the receiver and demodulated or decoded to obtain a message, and finally the message is delivered to the sink. In a general communication system model, shannon uses boltzmann form S ═ kln (w) of thermodynamic entropy for reference, and proposes that information entropy is used to describe the amount of information required by a system, including two important formulas of information entropy h (x) (formula 1) and channel capacity C (formula 2). H, C, W respectively represents information entropy, channel capacity and bandwidth, S/N is signal-to-noise ratio, and P is_iIs the probability of message i.

The larger the information entropy value is, the more chaotic the system is represented, and the larger the information quantity required by the system is. The channel capacity limits the information throughput per unit of time, i.e. the maximum information throughput of the system.

The method is based on a general communication system model, carries out similar modeling on a general optical imaging system, and can describe the process that a coder (data acquisition means such as multiplexing technology and the like) acquires a series of coded image sequence sets, then a decoder (such as a preprocessing algorithm, a reconstruction algorithm and the like) processes the source data, and finally a high-quality target image is obtained. The invention can realize redundancy measurement of most optical imaging means and give result prediction and scheme optimization. Taking the FPM imaging process as an example, as shown in fig. 2b, the information after the light source (LED) irradiates the sample is collected by the detector (CCD/sCMOS) to obtain a low resolution image sequence set (LR) containing noise, and then the low resolution image sequence set is reconstructed by using the reconstruction algorithm to obtain a high resolution image. Since the discipline of fourier optics is itself based on the discipline of signals and systems, one-dimensional temporal signals are generalized to two-dimensional space. It is therefore possible to migrate the entropy of information describing the communication system to the optical imaging system. In contrast, the FPM technique is not a simple "what you see is what you get" technique, which adds one step to the reconstruction process of reconstructing a series of low resolution images into high resolution images, compared to the general communication system model.

From the experimental experience associated with optical microscopy imaging, such as the FPM technique, the following basic principles are known:

1) the redundant information refers to frequency domain information redundancy generated in a sub-aperture overlapping area in a frequency domain or related spatial domain information redundancy generated in spatial domain oversampling;

2) numerical aperture NA of objective lens_objSpatial sampling rate R_camFrequency domain overlap ratio R_overlapWavelength λ, number of non-repeating angular illuminated low resolution images n, signal to noise ratio SNR_i(i ═ 1,2, …, n), a preprocessing algorithm, and the like, are main parameters that affect the FPM reconstructed image quality;

3) numerical aperture NA of objective lens_objThe larger the sample is, the clearer the observed sample is, the larger the amount of information provided is, the less the amount of information required is, and therefore, the larger the amount of information is, the more inversely the amount of information is;

4) spatial sampling rate R_camAnd frequency domain overlap ratio R_overlapApproximate linear compensation relationship between each other within a certain range [ Sun J, Qian C, Zhang Y, et al].Optics Express,2016,24(14):15765.]Space domain sampling rate R_camAnd frequency domain overlap ratio R_overlapAre all in direct proportion to the total redundant information quantity;

5) wavelength lambda and objective numerical aperture NA_objBoth relate to the optical system bandwidth W, for a circular pupil:

6) the more n low-resolution images without repeated angle illumination, the more redundant information is provided, and the larger the information amount is, and the repeated low-resolution images do not have additional redundant information and do not contribute to the final imaging result improvement, so the additional information amount is not provided;

7) SNR (Signal-to-noise ratio)_iAnd (i ═ 1,2, …, n) is directly related to the preprocessing algorithm, which can be regarded as a priori information about the noise, so that the signal-to-noise ratio of each low-resolution image can be directly calculated. Because the FPM technical reconstruction algorithm has certain robustness and can effectively improve the imaging quality after simple preprocessing, the images before and after preprocessing are respectively regarded as the image with noise pollution and the image without noise pollution completely. In practice, the preprocessed image cannot be completely regarded as an image without noise pollution, the more accurate preprocessing method is, the more accurate the signal-to-noise ratio is, but according to the test result, the influence of the preprocessing method on the final evaluation model is very small, so that the requirements of simple algorithm, easy operability and no parametrization are considered, and only the preprocessing method of subtracting the background image is used;

8) like a general communication system, the FPM system is a linear space (time) -invariant system.

According to the basic principle and the mutual restriction relation among the parameters, a formula 2 is used for reference, and a calculation model (formula 3, formula 4) of the image sequence lumped redundancy information amount of the FPM imaging system is constructed, wherein Q represents the low-resolution image sequence lumped redundancy information amount obtained by the whole imaging system, and Q represents the low-resolution image sequence lumped redundancy information amount obtained by the whole imaging system _iQuantity of information, Q, characterizing each image_iThe form of the method is basically the same as that of the formula 2, the method is slightly different in expression form, the logarithm with the base 2 is selected, a physicist likes a natural logarithm, and the difference of constants is mainly negligible; considering that the sampling rate is not necessarily 2 times, the 2-time coefficient of equation 2 is replaced by Rcam, R_camRepresenting the spatial sampling rate; for FPM imaging systems, which primarily use redundant information generated by overlap to reconstruct high resolution results, Q is Q_iIs multiplied by Roverlap, R_overlapRepresenting the frequency domain overlap ratio, W representing the optical system bandwidth, SNR_iRepresenting the signal-to-noise ratio.

Q_i＝W ln(1+SNR_i) (3)

The sample complexity η is defined taking into account the sample differences, the more complex the sample the more redundant information is certainly needed. Thus, optical microscopy imaging, such as the FPM technique, does not have an extreme overlap ratio applicable to any sample. In imaging systems, similar to entropy, image entropy is typically used to measure the degree of information redundancy [ Xi L, Liu G, Ni J].IEEE Transactions on Aerospace andElectronic Systems,1999,35(4):1240-1252.]. Compared with the one-dimensional entropy of the image, the two-dimensional entropy of the image can not only describe the average information content of the sample, but also measure the spatial characteristics, so that the two-dimensional entropy calculation H is carried out on the sample image I _IThe sample complexity η (equation 5 and equation 6) is characterized. Wherein p is_ijThe frequency of the appearance of the characteristic binary group f (I, j) in the pixel of the image I is MN, I is greater than or equal to 0 and less than or equal to 255 is the gray value of the pixel, and j is greater than or equal to 0 and less than or equal to 255 is the average value of the neighborhood gray values. The complexity calculation formula of the sample to be detected is as follows:

p_ij＝f(i,j)/MN(0≤j≤255) (6)

then, for the FPM imaging system, considering that the loss of the sample phase causes the imaging system not to be linear to the image intensity, the image I in the above equation uses the off-focus image liner of the central bright field image containing the phase information to calculate the sample complexity η (see equation 7) to obtain more accurate results.

Wherein p is_ijF (I, j)/MN (j is more than or equal to 0 and less than or equal to 255) represents the frequency of the appearance of the image characteristic binary f (I, j) of the central bright field defocused image I of the sample to be detected, MN is the image size, I is more than or equal to 0 and less than or equal to 255 is the gray value of a pixel, and j is more than or equal to 0 and less than or equal to 255 is the neighborhood gray average value; h_IThe entropy of information representing an out-of-focus image map of a central bright field image I containing phase information, I representing the central bright field image containing phase information.

See equation 8 for LED_m,nAnd its corresponding illumination wave vector (k)_x,m,n,k_y,m,n) The transmitted light wave field passing through the sample object function o (x, y) is

Wherein (x, y) represents object plane coordinates, (k) _x,k_y) Representing a two-dimensional frequency domain coordinate system, (m, n) representing the position of the illuminating LED, and j being an imaginary unit. By introducing a quadratic phase factor in the transmitted light wave field

Defocusing is achieved, where z is the amount of defocus and λ is the illumination wavelength.

For non-parametrization, the general value range of the defocus z in the experiment is: 1-10 mu m, the invention makes certain approximation, according to the test result, the influence of the defocusing amount on the image entropy is less than 0.2bit and can be ignored, so that z can select 3 mu m in the simulation and actual experiment, and other values are also possible. (m, n) selecting the center position of the LED array, and finally obtaining a defocusing image I of the central bright field image through Fourier inverse transformation and mode square.

Further, the invention proposes to construct a redundancy measurement model as follows:

αQ≥η (9)

wherein alpha is the utilization rate of the reconstruction algorithm, Q is the lumped redundant information quantity of the sequence of the coded image (low-resolution image), and eta is the complexity of the sample to be measured.

And (3) operating the redundancy measurement model for multiple times according to the category of the sample to be detected to obtain the utilization rate of the inherent reconstruction algorithm based on the sample to be detected of different categories, wherein the utilization rate is used for representing the capability of the algorithm related to the decoder for extracting information from the low-resolution image set.

We know that different reconstruction algorithms have different information extraction rates, which is why some algorithms can reconstruct only a few images or even a single image. The utilization rate of the reconstruction algorithm is defined as alpha, the utilization rate of the reconstruction algorithm alpha is used as an inherent attribute of the reconstruction algorithm, the capability of the algorithm for extracting information from a coded image sequence set (low-resolution image set) is represented, and when the reconstruction algorithm is determined, the alpha is a fixed value theoretically.

Therefore, a plurality of groups of samples to be tested of different types are substituted into the redundancy measurement model, the alpha value of the utilization rate of the reconstruction algorithm is calculated under the reconstruction limit, and when the alpha values of the utilization rate of the reconstruction algorithm corresponding to the plurality of groups of samples to be tested are all similar, the alpha value is the inherent utilization rate of the reconstruction algorithm.

The calibration method of the reconstruction algorithm utilization rate alpha comprises the following steps of comparing a plurality of groups of alpha obtained by calculation under reconstruction limits under different experimental conditions, namely taking equal sign to (formula 9) to calculate and compare alpha of different groups, and when alpha values are all similar, conservatively selecting the minimum alpha calibration value as the algorithm utilization rate.

And finishing redundancy judgment through the lumped redundancy information quantity of the coded image (low-resolution image) sequence, the complexity of the sample to be detected and the utilization rate of the calibration algorithm to obtain the parameter evaluation scheme of the imaging system. Substituting the determined image sequence lumped redundancy information quantity, the complexity of the sample to be detected and the calibration algorithm utilization value into a redundancy measurement model, and when the value of alpha Q is more than or equal to the sample complexity eta, indicating that the current sample can accurately reconstruct a high-resolution image under the imaging system parameter, namely that the imaging system parameter evaluation scheme is executable; when the value of alpha Q is less than the sample complexity eta, the redundant information quantity provided by the current system is represented and the weight is not satisfied Constructing requirements, namely, the imaging system parameter evaluation scheme cannot be executed, and cannot be accurately reconstructed; when the value of alpha Q is equal to eta, the redundant information quantity Q obtained by calculation at the moment_limitA limit value for the amount of information needed to reconstruct accurately for the current sample.

Specifically, for the FPM imaging technique, the redundancy measure module implements the flow shown in fig. 3. The redundancy judgment needs to be carried out by three inputs, namely, calculating the information quantity of a single image for the collected low-resolution original image sequence set, obtaining the total redundant information quantity Q according to the corresponding weight of the single image, calculating the sample complexity eta according to the two-dimensional entropy of the defocused image of the central bright field, calibrating the algorithm utilization rate alpha for a decoder (mainly referred to as a reconstruction algorithm in FPM), and carrying out redundancy judgment according to the three input quantities.

The parameter evaluation scheme of the imaging system in the step 4 of the invention obtains the feasibility of whether the parameters of the imaging system can realize the reconstruction of the high-resolution image or not by combining redundancy judgment, and marks the parameters of the specific imaging system which can not realize the reconstruction of the high-resolution image. The imaging system parameters included in the evaluation scheme at least comprise one or more of wavelength, objective lens NA, illumination NA, spatial sampling rate, frequency domain overlapping rate, image sequence set, noise level, and sample complexity and algorithm utilization rate.

Meanwhile, the imaging system parameter evaluation scheme also comprises a space bandwidth product limit SBP of the output imaging system,

wherein: FN is the number of views, mag is the magnification, NA_synFor the synthetic NA, the numerical aperture NA of the objective lens is calculated_objAnd the illumination numerical aperture NA_illuAnd, NA_sys＝NA_obj+NA_illu。

For the FPM imaging system, the imaging flux is given by the three tradeoffs of time, spatial resolution, and field of view. After the reconstruction algorithm and the related system parameters are determined, the time-space bandwidth product limit of the imaging system can be further calculated according to the ratio of the imaging space bandwidth product SBP to the total time of the image data collected by the detector.

In step 5 of the invention, if the parameter evaluation scheme of the imaging system cannot be executed, the optimized parameter scheme of the imaging system is obtained through the optimization module. The optimization module obtains an optimization scheme of one or more specific imaging system parameters including wavelength, objective lens NA, spatial sampling rate, frequency domain overlapping rate, image sequence set, noise level, sample complexity and algorithm utilization rate by using the redundancy measurement model.

DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION

According to the overall technical scheme of the invention, in order to verify the feasibility and the correctness of the model, the FPM imaging system based on the Gauss-Newton algorithm is selected as an embodiment (the algorithm flow is shown in figure 4) to be verified in a simulation group and an experiment group respectively. Of course, the reconstruction algorithm also comprises one or more of GS, EPRY-FPM, a wavelength multiplexing algorithm or an LED multiplexing algorithm and a deep learning reconstruction algorithm based on a physical model.

The simulation group carries out simulation experiments under three groups of different objective lenses of 2X/0.08 NA, 4X/0.1 NA and 10X/0.25 NA on the same sample. The analog noise parameters are set according to experimental experience, and Gaussian noise and Poisson noise are simultaneously set for the low-resolution image set to describe photon shot noise and readout noise in the imaging process. Wherein, the Gaussian noise is 5%, the Gaussian noise is mainly added in the dark field image, and the mean value of the Poisson noise is calculated by 10 from the pixel value³Scaling results and adding the scales in the bright and dark field images. And operating the redundancy measurement model for multiple times according to the category of the sample to be measured to obtain the inherent algorithm utilization rate based on the samples to be measured of different categories. And substituting a plurality of groups of samples to be tested of different types into the redundancy measurement model, calculating the utilization rate alpha value of the reconstruction algorithm under the reconstruction limit condition, and when the utilization rates alpha values of the plurality of groups of reconstruction algorithms are similar, determining the alpha value as the inherent algorithm utilization rate of the samples to be tested of different types.

The reconfiguration limit condition setting process includes: the spatial domain sampling rate is limited, the light source irradiation height and the number of LED arrays are adjusted, and the illumination numerical aperture NA is controlled_illuWhen heavyWhen the quality of the composition image meets the set requirement, the composition image is the reconstruction limit condition, and specifically:

1) For the sampling rate R_camCarrying out limit value taking to just meet the Nyquist law, wherein the value is close to 1;

2) adjusting the illumination NA by adjusting the height and number of LED arrays, and selecting the illumination numerical aperture NA_illuIn the vicinity of 0.6, the reconstructed sample can ensure the complete resolution state, and the fixed synthetic numerical aperture NA is kept_synSize to ensure a similar level of reconstruction resolution;

3) computing the minimum mean square root error of intensity_rmseSum phase minimum root mean square error P_rmseTo 1, pair_rmseIs limited to 0.01, for P_rmseThe observation of (2) is based on 0.1;

4) on the basis of meeting 1)2)3), finding the limit height, and obtaining the limit overlapping rate to obtain a formula 8 and the like. At this time, η/Q is used_limitAnd calibrating the algorithm utilization rate alpha under each group of simulation conditions. On one hand, the sample complexity η is the amount of information required to restore the sample to be accurate, so that enough illumination NA needs to be provided for waiting for the left side in formula 8; on the other hand, the actual illumination intensity of the LED will decay rapidly (proportional to cos) as the angle θ increases⁴θ)[Pan A,Zhang Y,Wen K,et al.Subwavelength resolution Fourier ptychography with hemispherical digital condensers[J].Optics Express,2018,26(18):23119-23131.]Due to the attenuated signal-to-noise ratio, the dark-field image can be easily submerged in noise, the image with higher angle has almost no signal-to-noise ratio, the information content is 0, and the illumination NA of 0.6 meets the sample requirement used in the simulation and experiment.

The simulation experiment result of the 2 x/0.08 NA objective lens is shown in figure 5, and the simulation limit parameters are as follows: wavelength is 0.532 μm, pixel size of the detector is 6.5 μm, illumination NA is controlled around 0.6, number of low resolution images is 17X 17, and LED array spacing is 4 mm. The limit of the sampling rate is 1.0231, the height is 60mm, the overlapping rate is 48.63 percent, and the limit redundant information quantity Q_limitAt 31.6749 bits, the algorithm utilization is 24.09%.

The simulation experiment result of the 4 x/0.1 NA objective lens is shown in figure 6, and the simulation limit parameters are as follows: the wavelength is 0.532 μm, the pixel size of the detector is 10.5 μm, the illumination NA is controlled to be around 0.6, the number of low resolution images is 13X 13, and the spacing is 4 mm. The limit of the sampling rate is 1.0133, the height is 48mm, the overlapping rate is 48.69 percent, and the limit redundant information quantity Q_limitAt 31.1027 bits, the algorithm utilization is 24.54%.

The simulation experiment result of the 10X/0.25 NA objective lens is shown in FIG. 7, and the simulation limit parameters are as follows: the wavelength is 0.532 μm, the pixel size of the detector is 10.5 μm, the illumination NA is controlled to be around 0.6, the number of low resolution images is 5 × 5, and the spacing is 8 mm. The limit of the sampling rate is 1.0133, the height is 38mm, the overlapping rate is 49.06 percent, and the limit redundant information quantity Q_limitAt 31.9397 bits, the algorithm utilization is 23.89%.

Experimental group reconstruction experiments were performed on resolution and cotton leaf transected samples, and the resolution plate USAF experiments are shown in fig. 8 a. The main parameters of the imaging system are 4X/0.1 NA objective, green wavelength 0.51808 μm, LED spacing of 4mm, number of low resolution images of 13X 13, detector pixel size of 3.75 μm, and USAF sample complexity of 6.1250bit under 4X/0.1 NA objective. The height data of 41mm, 43mm and 45mm are respectively displayed, and the contrast reconstruction result is observed, and the result shows that the height of 43mm can be restored to 10 groups 1 at a sampling rate of 2.7631, and the height of 43mm can not be accurately reconstructed under 41mm, so that the limit position of the resolution plate data is 43mm, the limit of the redundant information of the imaging system is marked to be 24.6502bit, and the algorithm utilization rate is 24.85%.

Cotton leaf transection experiment As shown in FIG. 8b, the main parameters of the imaging system were 4X/0.1 NA objective, green wavelength 0.51808 μm, LED spacing of 4mm, number of low-resolution images of 13X 13, and detector pixel size of 7.5 μm. The sample complexity of a cotton leaf transverse cutting sample under a 4 x/0.1 NA objective lens is 8.1087bit, the data with the heights of 43mm, 45mm and 48mm are respectively displayed, the reconstruction result is observed, the reconstruction result with the height of 45mm is a limit position, the corresponding redundant information limit is 33.1191bit, and the algorithm utilization rate is 24.48%. Cotton leaf transection samples were more complex than the USAF data, requiring more redundancy for reconstruction at 45mm height than the USAF, consistent with the empirical assumption that samples of higher complexity require more redundant information to reconstruct a high resolution image.

By comparing the alpha calibration results at the extreme positions under different experimental conditions, alpha fluctuation in a small range is found, and the noise level directly influences the signal-to-noise ratio level of the image and the noise also has random fluctuation, so that the algorithm utilization rate obtained by calculation has small fluctuation. Thus, it can still be considered that different sets of experimental results satisfy the expected assumptions, and the α values are only relevant to the algorithm itself.

By observing the fluctuation of the Gaussian noise within 10% to alpha, as shown in FIG. 9, the statistical values of the alpha variation with noise in the three simulation cases of 2X/0.08 NA, 4X/0.1 NA and 10X/0.25 NA are obtained

24.30% ± 0.56%, 24.46% ± 0.44%, 24.29% ± 0.64%, respectively, can all meet within 24 ± 1%, the simulation results meet the expected assumption, and the α value is only related to the algorithm. The small amplitude deviation is because noise variation affects the SNR, but the LED array height or the overlap ratio corresponding to the limit condition of the noise variation also varies accordingly, so that the Q value is kept in a relatively balanced range. Similarly, the utilization rate of the statistical algorithm for carrying out multiple experiments on the resolution sample and the cotton leaf transverse cutting sample respectively can also meet the requirement

Variation within the range of 24. + -. 1%.

For the sake of conservation in practical experiments, the minimum statistical value α may be selected as the information utilization rate of the gauss-newton reconstruction algorithm, that is, α is 23%. Similarly, the quantitative metric model can determine the utilization of the remaining algorithms.

Here we present the utilization results of the partial algorithm, as shown in fig. 11. Including gauss [ Yeh L-H, Dong J, Zhong J, et al. Experimental robustness of Fourier transform phase transform Algorithms ] GS [ Gerchberg R W, Saxton W O. phase prediction from image and diffusion plane in electron-microscopy [ J ] optical, 1971,34(3): 275-inch ], EPRY-FPM [ Qu X, Zhong G, Yang C. embedded pulse function for Fourier transform [ J ] optical, 2014,22(5): 4960-inch-LED 2], wavelength multiplexing Algorithms [ light W, Jong D, S. transform for Fourier transform coefficients [ J ] optical, 2014,22(5): 4960-inch-LED 2], wavelength multiplexing Algorithms [ light W, JUNG D, S. 12. transform ] K [ 11. transform J ] optical, 7. multiplexing [ 11. transform J ] optical, 7. transform J ] N, K [ 11. transform J ] N, K-phase transform, K ] N, K [ 11, J ] multiplexing Algorithms [ 12, K ] of LED [ 12, K ] N, K-wavelength transform [ 12, K ] N, K-L-K [ 7, K ] multiplexing Algorithms [ 7, K ] of LED [ 12, K ] S, K-11, K-L-K [ 9, K-G, K-P [ 7, K-P ] multiplexing Algorithms of the multiplexing algorithm [ 12, K [ 7, K [ 12, K [ 7, K, S,2014,5(7), 2376 and 2389, and the like. The EPRY-FPM, wavelength multiplexing and LED multiplexing algorithm reconstruction and the Gauss-Newton algorithm have the same algorithm core, so that the algorithm utilization rate is obviously consistent. And replacing the reconstruction algorithm of the simulation group with a GS algorithm to obtain the corresponding algorithm utilization rate alpha of 24.09%, 24.53% and 23.88%, and still meeting the range of 24 +/-1%.

After the algorithm utilization rate is calibrated, the model can be used for predicting the FPM imaging redundant information quantity of other samples, and theoretical guidance is provided for experimental parameters.

After the utilization rate of the Gaussian Newton reconstruction algorithm is determined, a USAF resolution plate sample under an objective lens of 2 x/0.1 NA and under 80mm is selected to perform a reconstruction experiment so as to verify whether the redundancy model prediction result of the method can provide guidance for experiment parameters, and the result is shown in FIG. 12. Other experimental parameters were the same as those in the experimental group. The USAF sample complexity is calculated to be 4.5049bit, the redundancy information amount limit required by the system to accurately reconstruct the high-resolution image can be estimated to be 18.5540bit according to the redundancy quantitative measurement model, but the redundancy information amount calculated by the group of data is 5.9305bit which is far less than the model estimation limit value 18.5540bit, namely the current experimental parameters do not meet the formula 6, and the model can not accurately reconstruct the group of experimental parameters when predicting. The phase in the actual reconstruction result in fig. 12 cannot be recovered, and the pupil reconstruction fails, i.e., the imaging experiment conforms to the redundancy model estimation. In this regard, we can present an optimized parameter scheme for the imaging system: A4X/0.1 NA objective lens may be substituted for a 2X/0.1 NA objective lens in an imaging system to increase the sampling rate R _camAnd increase the number of image acquisitions n to increase the redundancyThe respiration rate Q.

The redundancy measurement model in the method not only provides theoretical guidance for experimental parameters, but also can further define the imaging flux limit, and the space-time bandwidth limit can be calculated according to the total time of the used detector for collecting images, for example, the theoretical limit of the space bandwidth product SBP of the experimental group imaging system is calculated to be 6 hundred million pixels.

The embodiment of the method for reconstructing the FPM by aiming at the Gauss-Newton algorithm successfully and quantitatively measures the redundancy theoretical limit of the FPM imaging system and further provides feasibility for measuring the imaging flux limit of the FPM imaging system. Similarly, the imaging model in the method is applicable to all high-resolution imaging technologies, and can theoretically explain other high-resolution reconstruction algorithms, such as mPIE [ Maiden A, Johnson D, Li P. Further improvements to the physiological iterative engine [ J ]. Optica,2017,4(7): 736. 745 ], deep learning algorithms [ Thanh, Nguyen, Yujia, et al. deep learning approach for Fourier iterative approach [ J ]. Optics express,2018 ]. Compared with the traditional reconstruction algorithm, in order to achieve the limit of redundant information required by accurate reconstruction, the number of images actually required by accurate reconstruction of the deep learning algorithm is smaller than that of the traditional FPM algorithm, and the deep learning algorithm can be explained from the level of the utilization rate of the algorithm, namely the utilization rate of the deep learning algorithm is higher than that of the traditional algorithm.

The invention also provides an electronic device, which comprises a processor and a memory;

the memory is used for storing;

the processor is configured to execute the method according to any one of the above embodiments through calling.

The invention also provides a computer program product comprising a computer program and/or instructions which, when executed by a processor, performs the steps of the method of any of the above embodiments.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, apparatus, or computer program product. Accordingly, the present invention may take the form of an entirely software embodiment, an entirely hardware embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention is in the form of a computer program product that may be embodied on one or more computer-usable storage media having computer-usable program code embodied therewith. And such computer-usable storage media include, but are not limited to: various media capable of storing program codes, such as a usb disk, a portable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk Memory, a Compact Disc Read-Only Memory (CD-ROM), and an optical Memory.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a system for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction system which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the present invention and its equivalent technology, it is intended that the present invention also include such modifications and variations.

Claims

1. A quantitative measurement method for redundant information of optical microscopy imaging is characterized by comprising the following steps: the method comprises the following steps:

And 2, step: constructing a redundancy measurement model;

and 3, step 3: a series of obtained coded image sequence sets are subjected to a redundancy measurement model to obtain the lumped redundancy information quantity of the coded image sequences and the complexity of a sample to be detected; utilizing a redundancy measurement model to calibrate the utilization rate of a reconstruction algorithm for a decoder;

2. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 1, wherein: the step 5 further comprises: and if the imaging system parameter evaluation scheme can not be executed, obtaining the optimized parameter scheme of the imaging system through the optimization module.

3. The method for quantitatively measuring the redundant information in micro-optical imaging according to claim 1 or 2, wherein: the construction of the redundancy measurement model comprises the following steps:

αQ≥η

4. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 3, wherein: step 3, respectively calculating the information quantity of each image in the image sequence set to obtain the lumped redundant information quantity of the coding image sequence; calculating the complexity of the sample to be detected by utilizing the entropy of the image information; and running the redundancy measurement model for multiple times according to the category of the sample to be detected to obtain the utilization rate of the inherent reconstruction algorithm based on the sample to be detected of different categories, wherein the utilization rate is used for representing the capability of the reconstruction algorithm related to the decoder to extract information from the coded image sequence set.

5. The method for quantitatively measuring the redundant information in optical microscopy imaging as set forth in claim 3, wherein: in the step 3: for a general optical imaging system, the calculation formula of the lumped redundant information quantity of the coded image sequence is as follows:

wherein: q represents the total information content of the image sequence set, R_overlapRepresenting the frequency domain overlap ratio, Q_iQuantity of information, SNR, characterizing each image_iThe signal-to-noise ratio is shown, and C is the bandwidth of the optical system and is determined by the corresponding parameters of the imaging system.

6. The method for quantitatively measuring the redundant information in optical microscopy imaging as set forth in claim 5, wherein: in the step 3: for the FPM imaging system, the calculation formula of the lumped redundant information quantity of the coding image sequence is as follows:

Wherein Q represents a coded picture sequenceLumped redundant information quantity, Q_iQuantity of information, R, characterizing each image_overlapRepresenting the frequency domain overlap ratio, SNR_iRepresenting the signal-to-noise ratio, R_camDenotes the spatial sampling rate and W denotes the optical system bandwidth.

7. The method for quantitatively measuring the redundant information in optical microscopy imaging as set forth in claim 6, wherein: the calculation method of the optical system bandwidth W is as follows:

8. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 3, wherein: in the step 3: for a general optical imaging system, the complexity calculation formula of the sample to be measured is as follows:

9. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 8, wherein: in the step 3: for the FPM imaging system, the complexity calculation formula of the sample to be measured is as follows:

wherein p is_ijF (i, j)/MN (j is more than or equal to 0 and less than or equal to 255) represents that the test is to be performedThe frequency of appearance of image feature binary f (I, j) of the central bright field defocused image I of the sample is MN, wherein MN is the image size, I is greater than or equal to 0 and less than or equal to 255 is the gray value of a pixel, and j is greater than or equal to 0 and less than or equal to 255 is the neighborhood gray average value; h _IInformation entropy of an out-of-focus image map of a central bright field image I containing phase information is shown, and I shows a central bright field out-of-focus image containing phase information.

10. The method for quantitatively measuring redundant information in micro-optical imaging according to claim 9, wherein: i described in step 3:

wherein

Indicating LED_m,nAnd its corresponding illumination wave vector (k)_x,m,n,k_y,m,n) The transmitted light wave field through the sample object function o (x, y); j is an imaginary unit;

11. The method for quantitatively measuring redundant information in optical microscopy imaging as defined in claim 10, wherein: the value range of the defocusing amount z is 1-10 mu m.

12. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 3, wherein: the calculation process of the utilization rate of the standard reconstruction algorithm in the step 3 is as follows: and calculating a reconstruction algorithm utilization rate alpha value under a reconstruction limit by substituting a plurality of groups of samples to be measured of different classes into the redundancy measurement model, wherein when values of the reconstruction algorithm utilization rate alpha values corresponding to the plurality of groups of samples to be measured are all similar, the alpha value is the inherent reconstruction algorithm utilization rate of the reconstruction algorithm.

13. The method for quantitatively measuring the redundant information in micro-optical imaging according to claim 12, wherein: the reconfiguration limit condition setting process comprises the following steps: and (3) taking a limit value of the airspace sampling rate, adjusting the light source illumination height and the number of the LED arrays, controlling the illumination NA, and obtaining a reconstruction limit condition when the quality of the reconstructed image meets the set requirement.

14. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 13, wherein: in the reconstruction limit condition, the illumination NA is preferably 0.6.

15. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 13, wherein: in the reconstruction limit condition, the root mean square error value of the intensity representing the quality requirement of the reconstructed image is preferably 0.01, and the minimum root mean square error of the phase is preferably 0.1.

16. The method for quantitatively measuring the redundant information in the micro-optical imaging according to claim 3, wherein: in step 4, the redundancy judgment process is as follows:

substituting the determined image sequence lumped redundancy information quantity, the complexity of the sample to be detected and the calibration reconstruction algorithm utilization value into a redundancy measurement model, and when the value of alpha Q is more than or equal to the sample complexity eta, indicating that the current sample can accurately reconstruct a high-resolution image under the imaging system parameter, namely that the imaging system parameter evaluation scheme is executable; when the value of the alpha Q is smaller than the sample complexity eta, the information quantity provided by the current system is represented, the reconstruction requirement is not met, accurate reconstruction cannot be carried out, and namely, the parameter evaluation scheme of the imaging system cannot be executed; when the value of alpha Q is equal to eta, the redundancy information quantity Q obtained by calculation at the moment _limitA limit value for the amount of information needed to reconstruct accurately for the current sample.

17. The method for quantitatively measuring redundant information in micro-optical imaging according to claim 16, wherein: and 4, the parameter evaluation scheme of the imaging system obtains the feasibility of whether the parameters of the imaging system can realize the reconstruction of the high-resolution image or not by combining redundancy judgment, and marks the parameters of the specific imaging system which can not realize the reconstruction of the high-resolution image.

18. The method for quantitatively measuring redundant information in micro-optical imaging according to claim 16 or 17, wherein: the imaging system parameters comprise one or more of wavelength, objective lens NA, illumination NA, spatial sampling rate, frequency domain overlapping rate, image sequence set, noise level, sample complexity and algorithm utilization rate.

19. The method for quantitatively measuring redundant information in optical microscopy imaging as set forth in claim 17, wherein: the step 4 of the imaging system parameter evaluation scheme further comprises outputting the space bandwidth product limit SBP of the imaging system,

wherein: FN is the number of fields of view, mag is the magnification, NA_synFor synthesizing numerical aperture, i.e. objective numerical aperture NA_objAnd the illumination numerical aperture NA _illuAnd, NA_sys＝NA_obj+NA_illu。

20. The method for quantitatively measuring the redundant information in optical microscopy imaging as set forth in claim 2, wherein: and 5, the optimization module obtains an optimization scheme of one or more specific imaging system parameters including wavelength, objective lens NA, spatial sampling rate, frequency domain overlapping rate, image sequence set, noise level, sample complexity and algorithm utilization rate by using the redundancy measurement model.

21. The method for quantitatively measuring the redundant information in optical microscopy imaging as set forth in claim 4 or 12, wherein: the reconstruction algorithm is one or more of a Gauss Newton algorithm, a GS algorithm, an EPRY-FPM algorithm, a wavelength multiplexing algorithm, an LED multiplexing algorithm or a deep learning reconstruction algorithm based on a physical model.

22. An electronic device comprising a processor and a memory;

the memory is used for storing;

the processor is used for executing the quantitative measurement method for the redundant information of the micro-optical imaging according to any one of claims 1 to 21 through calling.

23. A computer program product comprising a computer program and/or instructions, characterized in that the computer program and/or instructions, when executed by a processor, implement the steps of the method for quantitative measurement of redundant information for microscopy optical imaging according to any one of claims 1 to 21.