A method, system, storage medium and product for scatter imaging

Abstract

A method, system, storage medium, and product for scatter imaging, belonging to the field of image processing technology, include the following steps: acquiring a target speckle image; calculating a target amplitude based on the target speckle image; constructing an initial estimate based on the target amplitude; constructing an initial support model; iterating based on the initial support model and the initial estimate to obtain an iterative support model and an iterative estimate; iterating by shrinking the iterative support model to drive the iterative estimate, obtaining the target phase, and reconstructing the target image. This method, which uses the iterative support model to shrink the iterative estimate, can reduce the number of reconstruction iterations to less than 60. The method is insensitive to iteration parameters such as constraint coefficients, exhibits high robustness, and has a fast algorithmic calculation speed.

Description

Technical Field

[0001] The present invention belongs to the field of image processing technology, and in particular relates to a method, system, storage medium and product for scatter imaging. Background Art

[0002] In the biomedical field, there is an increasing demand for microscopy to diagnose and treat lesions. However, non-transparent biological tissue scatters light, hindering direct imaging of the interior of the organism. Traditional optical imaging systems cannot acquire images of targets in the presence of scattering light. Therefore, the non-invasive imaging of unknown targets obscured by biological tissue is of great research value in biomedicine, microscopy, and other fields.

[0003] Scatter imaging based on speckle correlation is an emerging optical imaging technology that successfully achieves non-invasive imaging through strongly scattering media, effectively addressing the problem of optical path obstruction in microscopic imaging. Based on the principle of speckle autocorrelation, the amplitude of the image of a target obscured by the scattering medium can be derived from the autocorrelation of the speckle pattern. Phase recovery algorithms can then be used to image the obscured target.

[0004] For example, Chinese patent CN112950731B discloses a single-frame method for imaging through scattering media in the presence of strong background interference. By processing the speckle autocorrelation image, the contrast of the speckle autocorrelation is further improved, enabling target reconstruction from speckle in the presence of strong background interference. Chinese patent CN111795949B also discloses a method and apparatus for anti-scattering imaging. This method reconstructs the original target image using Fourier domain amplitude information and high-precision Fourier domain phase information, improving phase recovery accuracy and enhancing noise immunity.

[0005] However, the existing speckle correlation-based imaging technology has a slow reconstruction speed, which is difficult to meet the real-time requirements of medical detection and cannot be applied in rapidly changing dynamic scattering scenarios. Summary of the Invention

[0006] The present invention aims to provide a scattering imaging method, which solves the problem of slow speckle image reconstruction speed mentioned in the above background technology by driving iterative estimation through iterative support model contraction.

[0007] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0008] A method for scatter imaging, comprising the following steps:

[0009] Acquire target speckle image;

[0010] Calculate the target amplitude according to the target speckle image;

[0011] Based on the target amplitude, construct an initial estimate; construct an initial support model;

[0012] An iterative support model and an iterative estimate are obtained according to the initial support model and the initial estimate. The iterative estimate is driven to iterate by shrinking the iterative support model to obtain a target phase, so as to reconstruct a target image.

[0013] Furthermore, the initial support model is a non-zero matrix, and the initial estimate is expressed by the following formula:

[0014] E0(u,v0=|Amp|·exp[iφ0(u,v)];

[0015] Where E0(u,v) is the initial estimate, Amp is the target amplitude, i is the imaginary unit, and φ0(u,v) is the random phase.

[0016] Furthermore, the iterative estimation includes time domain input estimation and time domain output estimation;

[0017] The iterative support model contraction driven iterative estimation is iterated, comprising:

[0018] Iterate the time domain input estimation through the constraint formula;

[0019] The constraint formula is as follows:

[0020]

[0021] in, is the time domain input estimate for the k+1th iteration; is the time domain input estimate for the kth iteration; is the time domain output estimate of the kth iteration; s k is the iterative support model of the kth iteration; α and β are constraint coefficients; when k = 0, is the inverse Fourier transform of the initial estimate, s0 is the initial support model;

[0022] The time domain output estimate of the k-th iteration is expressed as follows:

[0023]

[0024] in, is the time domain output estimate of the kth iteration; F -1 is the inverse Fourier transform operator, Amp is the target amplitude, i is the imaginary unit; φ k (u,v) is Phase;

[0025] The iterative support model of the k-th iteration is expressed by the following formula:

[0026] s k =s k-1 ·ρ k ·f k ;

[0027] Among them, s k is the iterative support model of the kth iteration, s k-1 is the iterative support model of the k-1th iteration, f k is the fuzzy kernel function of the kth iteration, ρ k To support the update function;

[0028] The support update function is as follows:

[0029]

[0030] Among them, ρ k To support the update function, is the time domain input estimate for the k-1th iteration, and ε is the threshold parameter.

[0031] Furthermore, the fuzzy kernel function is expressed by the following formula:

[0032]

[0033] in, is the fuzzy kernel function of the kth iteration, r k is the blur kernel length of the kth iteration, σ k is the standard deviation of the blur kernel at the kth iteration;

[0034] The blur kernel length of the k-th iteration is expressed as follows:

[0035]

[0036] Among them, r k is the blur kernel length of the kth iteration, k is the number of current iterations, and j is the total number of iterations;

[0037] The blur kernel standard deviation of the kth iteration is expressed as follows:

[0038] σ k =max(0.5,0.99 k ·σ0);

[0039] Among them, σ k is the standard deviation of the blur kernel at the kth iteration, k is the number of current iterations, and σ0 is the initial value of the blur kernel standard deviation.

[0040] Furthermore, the step of acquiring the target speckle image includes:

[0041] Acquire an initial speckle image;

[0042] Perform logarithmic transformation and Fourier transform on the initial speckle image to obtain the speckle frequency domain image;

[0043] performing a filtering operation on the speckle frequency domain image to obtain a speckle filtered image;

[0044] Performing inverse Fourier transform and exponential transform on the speckle filter image to obtain a target speckle image.

[0045] Furthermore, the filtering operation on the speckle frequency domain image is performed by a high-pass filter;

[0046] The high-pass filter is shown below:

[0047] H(u,v)=(α-β)exp(1-exp{-c[D 2 (u,v) / D0 2 ]})+β;

[0048] Where H(u,v) is a high-pass filter, α is a parameter for adjusting high-frequency components, β is a parameter for adjusting low-frequency components, exp is an exponential transformation operator, c is a high-pass filter sharpening coefficient, and D(u,v) is the distance from the coordinate (u,v) to the filter center.

[0049] The high-pass filter sharpening coefficient is obtained by the following steps, including:

[0050] Perform autocorrelation operation on the target speckle image to obtain a speckle autocorrelation image;

[0051] Determining the coordinates of a boundary point of the target speckle image, and using the coordinates of a boundary point of the target speckle image as the reference point coordinates;

[0052] In the form of an iterative cycle, the target speckle image is filtered using a step-by-step sharpening coefficient to obtain several initial filtered images;

[0053] Perform autocorrelation on several initial filtered images to obtain an initial autocorrelation image, and record the intensity value corresponding to the reference point coordinate in the initial autocorrelation image. When the intensity value corresponding to the reference point coordinate reaches the minimum value, the corresponding sharpening coefficient is the optimal sharpening coefficient of the target speckle image;

[0054] The reference point coordinates are calculated using the following formula:

[0055]

[0056] in, is the Laplace operator, is the target speckle image, is the autocorrelation operator, x0 and y0 are the specified x and y coordinates, θ th is the threshold coefficient.

[0057] Furthermore, the target amplitude is obtained by calculating the autocorrelation of the target.

[0058] A system for scatter imaging, comprising:

[0059] A speckle image acquisition module, used for acquiring a target speckle image;

[0060] A target amplitude calculation module is used to calculate the target amplitude;

[0061] The target phase calculation module is used to construct an initial estimate based on the target amplitude; construct an initial support model; and, based on the initial support model and the initial estimate, obtain an iterative support model and an iterative estimate, and iterate the iterative estimate by shrinking the iterative support model to obtain target phase information to reconstruct the target image.

[0062] A computer-readable storage medium stores a computer program / instruction thereon, which performs the steps of the method for scatter imaging when executed by a processor.

[0063] A computer program product comprises a computer program, which implements the steps of the method for scatter imaging when executed by a processor.

[0064] The present invention has the following advantages: the iterative process convergence is driven by the contraction of the iterative support model, the iterative support model can quickly fit the shape of the occluded sample based on the iterative process, and drive the phase recovery process convergence in a positive feedback manner, shortening the number of reliable reconstruction iterations to within 60 times. In addition, this method is insensitive to iterative parameters such as constraint coefficients, has high robustness, and fast algorithm calculation speed.

[0065] At the same time, the present application can significantly improve the spectrum signal-to-noise ratio obtained in the autocorrelation process by adaptively setting the optimal Gaussian filter sharpening coefficient, and has adaptability and robustness to improve processing speed and processing accuracy.

[0066] Other features and advantages of the present invention will be disclosed in detail in the following specific embodiments and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A flow chart of the method for scatter imaging of the present application;

[0068] Figure 2 is a schematic structural diagram of an imaging device;

[0069] Figure 3 is the initial speckle image;

[0070] Figure 4 is the target speckle image;

[0071] Figure 5 is the speckle autocorrelation image;

[0072] Figure 6 Target speckle image for determining reference points;

[0073] Figure 7 are several initial filtered images;

[0074] Figure 8 is the image of the initial support model;

[0075] Figure 9 is the initial estimated image;

[0076] Figure 10 is the image of the reconstructed iterative support model.

[0077] Explanation of the marks in the figure: 1. Incoherent monochromatic LED; 2. Stage; 3. Sample to be tested; 4. Blocking biological tissue; 5. Microscope objective; 6. Pupil aperture; 7. Tube lens; 8. Imaging sensor; 9. Field programmable gate array; 10. Display. DETAILED DESCRIPTION

[0078] In order to better understand the purpose, structure and function of the present invention, the present invention is further described in detail below with reference to the accompanying drawings. In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral whole; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances.

[0079] A method for scatter imaging, such as Figure 1 As shown, the following steps are included:

[0080] S1: Acquire target speckle image;

[0081] S2: Calculate the target amplitude based on the target speckle image;

[0082] S3: Calculate the target phase according to the target speckle image and the target amplitude to reconstruct the target image.

[0083] In this embodiment, the step S1 of acquiring the target speckle image includes the following steps:

[0084] S11: Acquire an initial speckle image; in this embodiment, the initial speckle image is as follows Figure 3 shown.

[0085] S12: Preprocess the initial speckle image to obtain a target speckle image; in this embodiment, the target speckle image is as follows: Figure 4 shown.

[0086] In this embodiment, Figure 2 As shown, the initial speckle image in step S11 can be obtained by an imaging device.

[0087] The imaging device includes, from top to bottom, an incoherent monochromatic LED 1, a stage 2, a sample to be measured 3, a shielded biological tissue 4, a microscope objective 5, a pupil aperture 6, a tube lens 7, an imaging sensor 8, a field programmable gate array 9 (FPGA) and a display 10.

[0088] A stage 2 is disposed on the incoherent monochromatic LED 1. The stage 2 is used to place a sample 3 to be tested and a shielded biological tissue 4. Above the stage 2, a microscope objective lens 5, a pupil aperture 6, a tube lens 7, and an imaging sensor 8 are disposed in order from bottom to top. The imaging sensor 8 is communicatively connected to the field programmable gate array 9, which is in turn communicatively connected to the display 10.

[0089] The field programmable gate array 9 controls the incoherent monochromatic LED 1 to illuminate the sample to be detected; the light carrying the information of the sample to be detected is blocked and scattered by the biological tissue. The outgoing light passing through the biological tissue passes through the microscope objective lens 5, the pupil aperture 6 and the tube lens 7 in sequence and is captured by the imaging sensor 8 to obtain the initial speckle image of the current sample in real time.

[0090] The step S12 pre-processes the initial speckle image to obtain a target speckle image, including the following steps:

[0091] S121: performing logarithmic transformation and Fourier transformation on the initial speckle image to obtain a speckle frequency domain image;

[0092] S122: performing a filtering operation on the speckle frequency domain image to obtain a speckle filtered image;

[0093] S123: Performing inverse Fourier transform and exponential transform on the speckle filter image to obtain a target speckle image.

[0094] The step S121 can be expressed by the following formula:

[0095]

[0096] Where I(u,v) is the speckle frequency domain image, F is the Fourier transform operator, ln is the logarithmic transform operator, is the initial speckle image. The speckle frequency domain image is the frequency domain form of the speckle image.

[0097] In step S122 , a filtering operation is performed on the speckle frequency domain image using a high-pass filter.

[0098] The high-pass filter is shown below:

[0099] H(u,v)=(α-β)exp(1-exp{-c[D 2 (u,v) / D0 2 ]})+β;

[0100] Among them, H(u,v) is a high-pass filter, α is a parameter for adjusting the high-frequency component, β is a parameter for adjusting the low-frequency component, exp is an exponential transformation operator, c is the high-pass filter sharpening coefficient, and D(u,v) is the distance from the coordinate (u,v) to the center of the filter.

[0101] The high-pass filter sharpening coefficient is obtained by the following steps, including:

[0102] An autocorrelation operation is performed on the target speckle image to obtain a speckle autocorrelation image. In this embodiment, the speckle autocorrelation image is as follows: Figure 5 As shown;

[0103] Determine the boundary point coordinates of the target speckle image, and use the boundary point coordinates of the target speckle image as the reference point coordinates, such as Figure 6 As shown;

[0104] In the form of an iterative cycle, the target speckle image is filtered using a step-by-step sharpening coefficient to obtain several initial filtered images, such as Figure 7 As shown;

[0105] Autocorrelation is performed on several initial filtered images to obtain an initial autocorrelation image. The intensity values ​​corresponding to the reference point coordinates in the initial autocorrelation image are recorded. When the intensity values ​​corresponding to the reference point coordinates reach a minimum value, the corresponding sharpening coefficient is the optimal sharpening coefficient for the target speckle image. In this embodiment, a corresponding speckle autocorrelation matrix can be obtained from the speckle autocorrelation image. The intensity values ​​corresponding to the reference point coordinates are the values ​​corresponding to the reference point coordinates at the corresponding position on the speckle autocorrelation matrix.

[0106] The reference point coordinates are calculated using the following formula:

[0107]

[0108] in, is the Laplace operator, is the target speckle image, is the autocorrelation operator, x0 and y0 are the specified x and y coordinates, θ th is the threshold coefficient.

[0109] Optionally, in this embodiment, x0=m / 2+2, y0=n / 2+2;

[0110] Wherein, m is the height of the speckle autocorrelation image, and n is the width of the speckle autocorrelation image. In this embodiment, by setting x0=m / 2+2 and y0=n / 2+2, it is possible to prevent cross-distributed noise in the autocorrelation domain from affecting the selection of the reference point.

[0111] This application uses self-correcting feature extraction and amplification to capture low-contrast speckle images in non-darkroom environments. It can perform self-correcting filtering optimization on low-contrast, low-signal-to-noise ratio speckle images captured in complex imaging environments, effectively separating and amplifying the characteristic components of the speckle pattern while suppressing its noise components. Furthermore, the optimal sharpening coefficient can be adaptively set, ultimately significantly improving the spectral signal-to-noise ratio obtained during the autocorrelation process, demonstrating both adaptability and robustness.

[0112] Optionally, the step S1 of acquiring the target speckle image may be directly acquiring the speckle image after or without preprocessing.

[0113] The target amplitude in step S2 is obtained by calculating the autocorrelation of the target.

[0114] Specifically, the autocorrelation of the sample to be tested may be calculated according to the target speckle image to obtain the target amplitude;

[0115] As follows:

[0116]

[0117] Where Amp is the target amplitude, F is the Fourier transform operator, is the target speckle image, is the autocorrelation operator.

[0118] Calculating the target phase according to the target speckle image and the target amplitude in step S3 includes the following steps:

[0119] S31: constructing an initial estimate based on the target amplitude; constructing an initial support model based on the target speckle image;

[0120] S32: Obtain an iterative support model and an iterative estimate based on the initial support model and the initial estimate, and perform iteration by shrinking the iterative support model to drive the iterative estimate to obtain a target phase.

[0121] like Figure 8As shown, the initial support model in step S31 is a non-zero matrix, and the size of the initial support model is close to half the size of the speckle autocorrelation image or the speckle autocorrelation matrix. Preferably, the initial support model is a non-zero matrix half the size of the speckle autocorrelation image. The initial estimation is as follows Figure 9 shown.

[0122] The initial estimation in step S31 is expressed by the following formula:

[0123] E0(u,v)=|Amp|·exp[iφ0(u,v)];

[0124] Where E0(u,v) is the initial estimate, Amp is the target amplitude, i is the imaginary unit, and φ0(u,v) is the random phase. 2 =-1.

[0125] The iterative estimation in step S32 includes time domain input estimation and time domain output estimation; the initial support model and the initial estimation are iterated through a constraint formula.

[0126] The constraint formula is as follows:

[0127]

[0128] in, is the time domain input estimate for the k+1th iteration; is the time domain input estimate for the kth iteration; is the time domain output estimate of the kth iteration; s k is the iterative support model of the kth iteration; α and β are constraint coefficients. In this embodiment, when k=0, is the inverse Fourier transform of the initial estimate, s0 is the initial support model, and α and β are empirical parameters, generally ranging from 0.7 to 0.8. Meaning In the matrix s k Median value.

[0129] The time domain output estimate of the k-th iteration is expressed as follows:

[0130]

[0131] in, is the time domain output estimate of the kth iteration; F -1 is the inverse Fourier transform operator, Amp is the target amplitude, i is the imaginary unit; φ k (u,v) is Phase;

[0132] The initial support model is iterated using the following formula:

[0133] s k =s k-1 ·ρ k ·f k ;

[0134] Among them, s k is the support model of the kth iteration, s k-1 is the support model of the k-1th iteration, f k is the fuzzy kernel function of the kth iteration, ρ k To support the update function;

[0135] The support update function is as follows:

[0136]

[0137] Among them, ρ k To support the update function, is the time domain input estimate for the k-1th iteration, and ε is the threshold parameter.

[0138] The fuzzy kernel function is expressed as follows:

[0139]

[0140] in, is the fuzzy kernel function of the kth iteration, r k is the blur kernel length of the kth iteration, σ k is the standard deviation of the blur kernel at the kth iteration;

[0141] The blur kernel length of the k-th iteration is expressed as follows:

[0142]

[0143] Among them, r k is the blur kernel length of the kth iteration, k is the number of current iterations, and j is the total number of iterations;

[0144] The blur kernel standard deviation of the kth iteration is expressed as follows:

[0145] σ k =max(0.5,0.99 k ·σ0);

[0146] Among them, σ k is the blur kernel standard deviation of the kth iteration, k is the number of current iterations, and σ0 is the initial value of the blur kernel standard deviation. In this embodiment, σ0=5.

[0147] Specifically, the step S32 obtains an iterative support model and an iterative estimate based on the initial support model and the initial estimate, and iterates the iterative estimate by shrinking the iterative support model to obtain the target phase, including the following steps:

[0148] S321: Perform an inverse Fourier transform on the frequency domain estimate to obtain a time domain input estimate. The frequency domain estimate includes an initial estimate.

[0149] When the frequency domain estimate is the initial estimate, the initial time domain input estimate, i.e., the time domain input estimate of the 0th iteration, is obtained by inverse Fourier transform, which is expressed by the following formula:

[0150]

[0151] Among them, E0(u,v) is the initial estimate, is the initial time domain input estimate. The frequency domain estimate is converted into a time domain estimate through inverse Fourier transform.

[0152] S322: Replace the current time domain input estimated amplitude with the target amplitude to obtain the current time domain output estimate. As shown in the following formula,

[0153] in, is the time domain input estimate of the kth iteration, F -1 is the inverse Fourier transform operator, Amp is the target amplitude, i is the imaginary unit, φ k (u,v) is phase.

[0154] When k = 0, it is the time domain input estimation of the 0th iteration,

[0155] S323: Build the current iteration support model k When k=0, s0 is the initial support model.

[0156] like Figure 10 As shown in Figure 2, it is the reconstructed iterative support model.

[0157] S324: Iterate the current iteration estimation by driving the current iteration estimation with the constraint formula through the current iteration support model.

[0158] When using the initial time domain input estimate When iterating, we get

[0159] S325: Obtain the phase of the current time domain input estimate according to the current time domain input estimate.

[0160] S326: Repeat steps S322-S325 to perform iterations. After a specified number of iterations, the final phase in step S325 is output as the target phase.

[0161] The reconstructed target image can be reconstructed through the target phase and target amplitude, which is a prior art and will not be described in detail in this application.

[0162] The present invention provides a method and device for real-time imaging in biological tissue occlusion scenarios, which successfully achieves real-time imaging of biological tissue occlusion scenarios in a non-darkroom environment. It can efficiently and high-quality reconstruct the image of the occluded sample through the low-contrast speckle image captured in a complex imaging environment. The present invention has low requirements for the imaging scene and a fast reconstruction speed.

[0163] By shrinking the iterative support model and driving the convergence of the iterative process, the iterative support model can quickly fit the shape of the occluded sample based on the iterative process, drive the convergence of the phase recovery process in a positive feedback manner, and shorten the number of reliable reconstruction iterations to less than 60. This method is insensitive to iterative parameters such as the constraint coefficient, has high robustness, and fast algorithm calculation speed.

[0164] This invention combines software and hardware computing to optimize the imaging link, integrating optical computational analysis into the FPGA's image algorithm layer, significantly improving computational speed. The imaging method achieves reconstruction time of less than 100 milliseconds, meeting the needs of real-time imaging of video streams exceeding 10 frames.

[0165] A system for scatter imaging, comprising:

[0166] A speckle image acquisition module, used for acquiring a target speckle image;

[0167] A target amplitude calculation module is used to calculate the target amplitude;

[0168] The target phase calculation module is used to construct an initial estimate based on the target amplitude; construct an initial support model; and, based on the initial support model and the initial estimate, obtain an iterative support model and an iterative estimate, and iterate the iterative estimate by shrinking the iterative support model to obtain target phase information to reconstruct the target image.

[0169] A computer-readable storage medium stores a computer program / instruction thereon, which performs the steps of the method for scatter imaging when executed by a processor.

[0170] A computer program product comprises a computer program, which implements the steps of the method for scatter imaging when executed by a processor.

[0171] It will be understood that the present invention is described by way of some embodiments, and it will be appreciated by those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are intended to be protected by the present invention.

[0172] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A method for scatter imaging, characterized in that: The steps include: Acquire target speckle image; Calculate the target amplitude according to the target speckle image; Based on the target amplitude, construct an initial estimate; construct an initial support model; Obtaining an iterative support model and an iterative estimate based on the initial support model and the initial estimate, and performing iteration by driving the iterative estimate through contraction of the iterative support model to obtain a target phase to reconstruct a target image; Wherein, the iterative estimation includes time domain input estimation and time domain output estimation; The iterative support model contraction driven iterative estimation is iterated, comprising: Iterate the time domain input estimation through the constraint formula; The constraint formula is as follows: ； in, is the time domain input estimate for the k+1th iteration; is the time domain input estimate for the kth iteration; is the time domain output estimate of the kth iteration; is the iterative support model for the kth iteration; and is the constraint coefficient; when k=0, is the inverse Fourier transform of the initial estimate, is the initial support model; The time domain output estimate of the k-th iteration is expressed as follows: ； in, is the time domain output estimate of the kth iteration; is the inverse Fourier transform operator, is the target amplitude, is an imaginary unit; for Phase; The iterative support model of the k-th iteration is expressed by the following formula: ； in, is the iterative support model for the kth iteration, is the iterative support model for the k-1th iteration, is the fuzzy kernel function of the kth iteration, To support the update function; The support update function is as follows: ； in, To support the update function, is the time domain input estimate for the k-1th iteration, is the threshold parameter.

2. The method for scatter imaging according to claim 1, characterized in that The initial support model is a non-zero matrix, and the initial estimate is expressed by the following formula: ； in, is the initial estimate, is the target amplitude, is the imaginary unit, is a random phase.

3. The method for scatter imaging according to claim 1, wherein: The fuzzy kernel function is expressed as follows: ； in, is the fuzzy kernel function of the kth iteration, is the blur kernel length of the kth iteration, is the standard deviation of the blur kernel at the kth iteration; The blur kernel length of the k-th iteration is expressed as follows: ； in, is the blur kernel length of the kth iteration, is the number of current iterations, is the total number of iterations; The blur kernel standard deviation of the kth iteration is expressed as follows: ； in, is the standard deviation of the blur kernel at the kth iteration, is the number of current iterations, is the initial value of the blur kernel standard deviation.

4. The method for scattering imaging according to any one of claims 1 to 3, characterized in that: The obtaining of the target speckle image comprises: Acquire an initial speckle image; Perform logarithmic transformation and Fourier transform on the initial speckle image to obtain the speckle frequency domain image; performing a filtering operation on the speckle frequency domain image to obtain a speckle filtered image; Performing inverse Fourier transform and exponential transform on the speckle filter image to obtain a target speckle image.

5. The method for scatter imaging according to claim 4, characterized in that: The filtering operation on the speckle frequency domain image is performed by a high-pass filter; The high-pass filter is shown below: ； in, is a high-pass filter, To adjust the parameters of the high frequency components, To adjust the parameters of the low-frequency component, is the exponential transformation operator, is the high-pass filter sharpening coefficient, For coordinates The distance to the filter center; The high-pass filter sharpening coefficient is obtained by the following steps, including: Perform autocorrelation operation on the target speckle image to obtain a speckle autocorrelation image; Determining the coordinates of a boundary point of the target speckle image, and using the coordinates of a boundary point of the target speckle image as the reference point coordinates; In the form of an iterative cycle, the target speckle image is filtered using a step-by-step sharpening coefficient to obtain several initial filtered images; Perform autocorrelation on several initial filtered images to obtain an initial autocorrelation image, and record the intensity value corresponding to the reference point coordinate in the initial autocorrelation image. When the intensity value corresponding to the reference point coordinate reaches the minimum value, the corresponding sharpening coefficient is the optimal sharpening coefficient of the target speckle image; The reference point coordinates are calculated using the following formula: ； in, is the Laplace operator, is the target speckle image, is the autocorrelation operator, and is designated and coordinate, is the threshold coefficient.

6. The method for scatter imaging according to claim 5, characterized in that: The target amplitude is obtained by calculating the autocorrelation of the target.

7. A system for scatter imaging, characterized in that: include: A speckle image acquisition module, used for acquiring a target speckle image; A target amplitude calculation module is used to calculate the target amplitude; The target phase calculation module is used to construct an initial estimate and an initial support model based on the target amplitude; and obtaining an iterative support model and an iterative estimate based on the initial support model and the initial estimate, and performing iteration by driving the iterative estimate through contraction of the iterative support model to obtain target phase information to reconstruct a target image; Wherein, the iterative estimation includes time domain input estimation and time domain output estimation; The iterative support model contraction driven iterative estimation is iterated, comprising: Iterate the time domain input estimation through the constraint formula; The constraint formula is as follows: ； in, is the time domain input estimate for the k+1th iteration; is the time domain input estimate for the kth iteration; is the time domain output estimate of the kth iteration; is the iterative support model for the kth iteration; and is the constraint coefficient; when k=0, is the inverse Fourier transform of the initial estimate, Initial support model; The time domain output estimate of the k-th iteration is expressed as follows: ； in, is the time domain output estimate of the kth iteration; is the inverse Fourier transform operator, is the target amplitude, for Phase; The iterative support model of the k-th iteration is expressed by the following formula: ； in, is the iterative support model for the kth iteration, is the iterative support model for the k-1th iteration, is the fuzzy kernel function of the kth iteration, To support the update function; The support update function is as follows: ； in, To support the update function, is the time domain input estimate for the k-1th iteration, is the threshold parameter.

8. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images