Speckle denoising method, system, device and medium based on iterative wavelet transform

Abstract

The application belongs to the technical field of image processing, and relates to a speckle denoising method, system, device and medium based on iterative wavelet transform, which comprises the following steps: wavelet decomposition and reconstruction are performed on an input speckle image to obtain a background estimation image; the background estimation image is compared with a deviation threshold value, and pixels exceeding the threshold value are subjected to truncation processing to generate a residual image; after iteration is updated, the final residual image is taken as the background estimation; the background estimation is subtracted from the original speckle image to obtain a denoised speckle image; and deconvolution operation is performed to obtain a denoised recovered image. The method combines the advantages of wavelet transform, effectively suppresses noise in a complex speckle scene, and finely removes noise on different scales while retaining key details of a target image through preprocessing of a speckle image and the multiscale characteristics of wavelet transform.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology and relates to a speckle denoising method, system, device and medium based on iterative wavelet transform. Background Technology

[0002] In photoelectric imaging, the inherent physical characteristics of the imaging system, such as the dark current effect of the sensor, and distortion factors like aberrations and spherical aberrations of the optical system, all contribute to random noise. This noise severely affects the quality and contrast of speckle reconstructed images, greatly reducing their usability. Therefore, speckle noise removal is of paramount importance for speckle reconstruction, both in research and practical applications.

[0003] In spatial domain methods, total variation (TV)-based methods are among the most versatile, as they can smooth the internal structure of an image while preserving edge information. The first TV-based speckle noise removal model achieves denoising using a constrained optimization method with two Lagrange multipliers.

[0004] In the transform domain, sine and cosine averaging is the most commonly used speckle denoising method due to its simplicity. Based on sine and cosine filtering, many scholars have made improvements, reducing the number of loops in traditional sine and cosine filtering and reducing the loss of grayscale information in the phase diagram.

[0005] Gaussian filtering is also an important filtering method. Based on the properties of the Gaussian function, it smooths images and can effectively remove noise. Its principle is to recalculate the value of each pixel in the image based on the weighted average of its neighboring pixels, where the weights are determined by the Gaussian function.

[0006] Wavelet transform is another important speckle denoising method. It decomposes an image into different feature maps through multi-scale decomposition and uses coefficient shrinkage techniques to achieve good denoising results. In recent years, wavelet transform-based methods have been continuously developed. Among them, the improved total variational denoising method based on shear wavelet transform includes a model that includes denoising regularization terms, weighted edge-preserving regularization terms, and weighted fidelity-preserving terms. By adjusting the weights, noise suppression and edge protection are achieved. The speckle denoising method based on stationary wavelet transform first performs sine and cosine decomposition and stationary wavelet decomposition on the original image, then processes high-frequency coefficients according to threshold rules, and finally performs arctangent operation to obtain the denoised image.

[0007] While diffusion-based methods in the spatial domain have achieved good results, they tend to overfit in high speckle noise conditions, leading to a loss of detail. TV-based methods, due to their inability to capture sufficient signal and noise information, are prone to the "staircase" effect, and their denoising performance varies significantly across different image regions, impacting the overall denoising outcome.

[0008] While sine and cosine averaging denoising is simple and the most commonly used speckle denoising method, it suffers from a problem where the denoising speed decreases with each iteration. Many researchers have improved upon sine and cosine filtering, reducing the number of iterations and minimizing the loss of phase diagram grayscale information. However, these algorithms are based on spatial or transform domain denoising and cannot simultaneously preserve phase transitions and phase details.

[0009] Gaussian filtering, a commonly used smoothing technique, is also applied in speckle noise suppression. However, the speckle recovery results after Gaussian filtering often still retain stripe artifacts, which not only limits its applicability in some fields but also indicates that this method may not be effective enough for certain types of noise.

[0010] Furthermore, wavelet transform methods in the transform domain do not fully consider the spatial information redundancy of the image, making them prone to artifacts and scratches in the image texture, which severely affects the denoising effect. Meanwhile, while speckle denoising methods based on stationary wavelet transform achieve better smoothing results with a higher number of stationary wavelet decomposition levels, they also lose some edge details.

[0011] In summary, the inherent physical characteristics of the imaging system and random noise caused by environmental factors lead to crosstalk in speckle reconstruction, resulting in poor image quality. Random noise not only introduces a large amount of noise components into the speckle, reducing the stability of the reconstruction results, but also easily produces artifacts and noise amplification, making the recovery of image details more difficult. Furthermore, these problems are particularly pronounced in high-noise environments, making it difficult to generate high-fidelity speckle reconstruction images and severely limiting the application of speckle imaging technology in scientific research, biological microscopy, astronomical observation, and satellite remote sensing. Although current speckle denoising technology is developing rapidly in multiple dimensions and through multiple paths, and has achieved certain results, there is still room for optimization and improvement in image quality enhancement. Therefore, developing a novel speckle denoising method that can effectively suppress noise while preserving image details and edge information has significant theoretical and practical value. Summary of the Invention

[0012] To address the problems existing in current technologies, this invention proposes a speckle denoising method, system, device, and medium based on iterative wavelet transform. This method combines the advantages of wavelet transform to effectively suppress noise in complex speckle scenes. By preprocessing the speckle image and utilizing the multi-scale characteristics of wavelet transform, this method can precisely remove noise at different scales while preserving key details of the target image. This method not only possesses strong noise suppression capabilities but also high algorithmic robustness, making it widely applicable in various fields of speckle image restoration, such as biological microscopy, astronomical imaging, computational photography, and remote sensing image processing.

[0013] This invention is achieved through the following technical solution:

[0014] A speckle denoising method based on iterative wavelet transform includes,

[0015] Obtain the original speckle image I 0 Initialize the dynamic deviation threshold T as the input speckle image I. k The average pixel value;

[0016] Input speckle image I k Perform multi-scale wavelet decomposition, retain the lowest frequency coefficients, and set the high frequency coefficients to 0;

[0017] Perform inverse wavelet transform on the lowest frequency coefficients to reconstruct the background estimation image.

[0018] Background estimation image The image is compared with a deviation threshold T, and pixels exceeding the threshold are truncated to generate a residual image I. k ;

[0019] Residual image I k As input for the next round, the updated dynamic bias threshold T is returned until the preset number of iterations is reached, and the final residual image is used as the background estimate.

[0020] Through the original speckle image I 0 Subtract background estimation Obtain denoised speckle image For denoised speckle images Perform deconvolution to obtain the denoised restored image.

[0021] Preferably, the original speckle image contains random noise caused by imaging system noise and environmental interference.

[0022] Preferably, the input speckle image I k Perform multi-scale wavelet decomposition, retain the lowest frequency coefficients, and set the high frequency coefficients to 0, specifically:

[0023] The input speckle image I was processed using a two-dimensional Daubechies-6 wavelet filter. k Perform an n-level decomposition to obtain the high-frequency coefficients {cD1, cD2, ..., cD} of each level. n} and the lowest frequency coefficient cA n Only the lowest frequency coefficient cA is retained. n All other high-frequency coefficients are set to zero.

[0024] Preferably, the lowest frequency coefficients are subjected to inverse wavelet transform to reconstruct the background estimation image. The expression is:

[0025]

[0026] Where IDWT represents the inverse discrete wavelet transform, k represents the current iteration number, is the lowest frequency coefficient cA retained in the k-th iteration n .

[0027] Preferably, the background estimated image is compared with the dynamic deviation threshold T, and pixels exceeding the deviation threshold are truncated. The expression is:

[0028]

[0029] The lowest frequency coefficient cA n is compared with the deviation threshold T. Pixel values exceeding the deviation threshold T are set to the threshold T, and pixel values not exceeding the deviation threshold T remain unchanged, obtaining the final speckle image I k .

[0030] Preferably, the residual image I k is used as the input for the next round, and the dynamic deviation threshold T is updated and returned until the preset number of iterations is reached. Specifically:

[0031] Set the maximum number of iterations to iterations, k represents the current iteration number,

[0032] If k < iterations, the residual image I k is used as the input for the next round, and the deviation threshold T is updated and returned;

[0033] If k = iterations, perform background estimation The expression is:

[0034]

[0035] The denoised speckle image is obtained by subtracting the background estimation 0 from the original speckle image I The expression is:

[0036]

[0037] Preferably, the denoised speckle image is deconvolved to obtain the denoised restored image,

[0038]

[0039] ​In the formula, deconv(·) represents the deconvolution operation, and PSF represents the point spread function of the system.

[0040] A speckle denoising system based on iterative wavelet transform includes,

[0041] The image acquisition module is used to acquire the raw speckle image I. 0 ;

[0042] The dynamic thresholding module is used to initialize the dynamic deviation threshold T, which is used as the input speckle image I. k The average pixel value; and the background estimation image The image is compared with a deviation threshold T, and pixels exceeding the threshold are truncated to generate a residual image I. k ;

[0043] The wavelet processing module is used to process the input speckle image I k Perform multi-scale wavelet decomposition, retain the lowest frequency coefficients, and set the high frequency coefficients to 0;

[0044] The reconstruction module performs inverse wavelet transform on the lowest frequency coefficients to reconstruct the background estimation image.

[0045] The iterative optimization module is used to optimize the residual image I. k As input for the next round, the updated dynamic bias threshold T is returned until the preset number of iterations is reached, and the final residual image is used as the background estimate.

[0046] Image restoration module, used to restore the original speckle image I 0 Subtract background estimation Obtain denoised speckle image For denoised speckle images Perform deconvolution operation to obtain the target restored image.

[0047] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the steps of a speckle denoising method based on iterative wavelet transform as described above.

[0048] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a speckle denoising method based on iterative wavelet transform as described above.

[0049] Compared with the prior art, the present invention has the following beneficial technical effects:

[0050] This invention proposes a speckle denoising method based on iterative wavelet transform. This method accurately estimates background noise and extracts target information through multiple iterations, while introducing a bias threshold adjustment mechanism to more precisely distinguish between background and target regions. This effectively removes noise while better preserving the structural features and details of the image. This innovative method provides a new approach to solving key challenges in speckle denoising and has significant theoretical value and practical application prospects.

[0051] Furthermore, the speckle denoising method based on iterative wavelet transform proposed in this invention can accurately remove the background information of the original speckle while preserving the Gaussian morphology unique to the original speckle, thus preserving the key feature information of the speckle to the greatest extent and laying the foundation for subsequent accurate processing.

[0052] Furthermore, the iterative wavelet transform-based speckle denoising method proposed in this invention can accurately and effectively remove streak artifacts, effectively eliminating this key factor affecting image quality. Simultaneously, the algorithm significantly improves image resolution and contrast; through rigorous mathematical models and optimization strategies, it enables the reconstructed image to achieve higher clarity and fidelity. Compared to traditional methods, the algorithm proposed in this invention better meets the stringent requirements of accuracy and clarity in high-precision image restoration, providing a more reliable and efficient solution for the practical application of related technologies and powerfully promoting the technological development of this field.

[0053] Furthermore, this invention introduces a bias threshold mechanism in iterative wavelet analysis, which effectively avoids misjudging useful target information as background noise and removing it, thereby ensuring the accuracy of the denoising process while preserving the effective information in the image. Attached Figure Description

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart of the speckle noise reduction method proposed in this invention;

[0056] Figure 2 This is a flowchart of the speckle denoising algorithm proposed in this invention;

[0057] Figure 3 This is a schematic diagram of the imaging optical path and overall process in an embodiment;

[0058] Figure 4The reconstruction results of the speckle denoising algorithm are compared in the example. Detailed Implementation

[0059] The technical solution of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] This invention addresses the problem of poor image quality in speckle reconstruction due to significant random noise interference in scattering imaging scenarios. It proposes a speckle denoising method based on iterative wavelet transform for high-quality speckle reconstruction. The specific steps are as follows, and the flowchart is shown below. Figure 1 As shown,

[0061] Step 1: Read the original speckle image

[0062] The original speckle image is read from the computer and denoted as I. 0 The image contains random noise caused by imaging system noise and environmental interference;

[0063] Step 2: Initialize the deviation threshold

[0064] The initial deviation threshold T is set to the average pixel value of the input speckle image I0. In subsequent iterations, the threshold is dynamically updated based on the current pixel mean of the input image.

[0065] Step 3: Multiscale wavelet decomposition

[0066] The input image was decomposed into n levels using a two-dimensional Daubechies-6 wavelet filter. Experiments showed that 7 levels were optimal, yielding the high-frequency coefficients {cD1, cD2, ..., cD} for each level. n} and the lowest frequency coefficient cA n Only the lowest frequency coefficient cA is retained. n All other high-frequency coefficients are set to zero. This decomposition process uses the multi-resolution characteristics of wavelet transform to separate image background noise into low-frequency components;

[0067] Step 4: Inverse wavelet reconstruction of background estimation

[0068] For the retained low-frequency coefficients cA n Perform inverse wavelet transform to reconstruct the estimated background image. The mathematical expression is:

[0069]

[0070] Where IDWT represents the inverse discrete wavelet transform, and k represents the current iteration number. It is the lowest frequency coefficient cA retained in the k-th iteration. n ;

[0071] Step 5: Dynamic threshold correction, such as Figure 2 As shown,

[0072] Background estimation image The data is compared with the deviation threshold T set in step 2, and pixels exceeding the threshold are truncated.

[0073]

[0074] The optimized residual image I is generated by applying the thresholding process. k ;

[0075] Step 6: Iterative Optimization Mechanism

[0076] Set the maximum number of iterations (experiments show that 5-7 iterations are optimal), where k represents the current iteration number. <iterations:

[0077] Residual image I k As input for the next round, return to step 2 to update the dynamic threshold;

[0078] If k = iterations, perform background estimation:

[0079]

[0080] The final denoised speckle image is equal to the original image minus the background estimate:

[0081]

[0082] Step 7: Deconvolution and Reconstruction

[0083] After obtaining noise reduction speckle Then, the intensity distribution image O of the object can be deconvolved from the speckle pattern using a deconvolution method. The denoised image is then recovered, and this is used to determine whether the method used is effective.

[0084]

[0085] Here, deconv(·) represents the deconvolution operation, PSF represents the point spread function of the system, and commonly used deconvolution algorithms include Wiener filtering algorithm, regularized deconvolution algorithm, blind deconvolution algorithm, etc.

[0086] 1. This invention proposes a speckle denoising method based on iterative wavelet transform. Unlike conventional denoising methods, this method uses multi-layer Daubechies-6 wavelet decomposition to accurately separate low-frequency noise components. Through multiple iterations, it achieves progressive estimation of background noise, highlighting key features, and finally removes the estimated background noise from the original signal. This method preserves the true characteristics of the signal to the greatest extent and can robustly remove low-frequency noise. This method has uniqueness at the overall architecture level.

[0087] It should be noted that, besides wavelet transform, other domain-specific transform methods, such as direct sine transform (DCT) and direct Fourier transform (DFT), also possess good frequency domain separation capabilities and can effectively suppress noise within a similar framework. Therefore, the design concept of this invention has strong versatility, and related transform methods can be adapted and replaced according to specific application requirements, thereby further expanding the applicability of this method in different types of image denoising tasks.

[0088] 2. This invention introduces a dynamic threshold constraint mechanism in iterative wavelet analysis. In each iteration, the threshold is updated based on the input image. The deviation of the background estimation is checked based on the threshold. For regions that exceed the set threshold after each iteration, the threshold level is adjusted to avoid erroneously removing useful target information as background noise. This ensures the accuracy of denoising and the retention of effective information in the image, which is a key step in ensuring good denoising results.

[0089] In the speckle denoising method proposed in this invention, the iterative wavelet transform plays a crucial role. Its working mechanism involves accurately extracting target information from the speckle within the transform domain while effectively filtering out background information. This transform domain can also be replaced by the frequency domain; high-frequency and low-frequency information of the speckle can also be extracted through Fourier transform.

[0090] The deviation threshold set in this invention is a soft threshold, meaning it changes with the number of iterations. A hard threshold could be used instead of the method described in this invention in certain specific situations.

[0091] Example 1

[0092] An imaging optical path and overall flowchart of a speckle denoising method based on iterative wavelet transform are shown below. Figure 3 As shown, where:

[0093] (a) The image shows the target “XDU”; (b) The scattering medium, where the beam of light from the target is scattered to form a speckle field; (c) The detector, used to receive the speckle pattern carrying target information; (d) The original speckle directly acquired by the detector, which contains a lot of background interference; (e) The speckle obtained after processing with a speckle denoising method, which is a denoised speckle with noise removed; (f) The point spread function (PSF) of the scattering medium acquired by the detector; (g) The clear restored image obtained by deconvolution reconstruction of the speckle and PSF.

[0094] The original speckle pattern, the best existing Gaussian filtering speckle denoising technique, and the speckle denoising technique proposed in this invention are compared. The speckle images are shown below. Figure 4 As shown in (a), (b), and (c), the three speckle patterns are deconvolved with the acquired PSF to reconstruct the images, and the reconstructed images are shown below. Figure 4 (d)(g)(j) Figure 4 (e)(h)(k) and Figure 4 As shown in (f)(i)(l), the algorithm proposed in this invention can effectively remove stripe artifacts from the reconstructed image while significantly improving the resolution and contrast of the reconstructed image.

[0095] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of a speckle denoising method based on iterative wavelet transform.

[0096] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the speckle denoising method based on iterative wavelet transform in the above embodiments.

[0097] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0098] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can 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, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0099] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0100] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A speckle denoising method based on iterative wavelet transform, characterized in that, include, Obtain the original speckle image I 0 Initialize the dynamic deviation threshold T as the input speckle image I. k The average pixel value; Input speckle image I k Perform multi-scale wavelet decomposition, retaining the lowest frequency coefficients and setting the high frequency coefficients to 0; Perform inverse wavelet transform on the lowest frequency coefficients to reconstruct the background estimation image. Background estimation image The image is compared with a deviation threshold T, and pixels exceeding the threshold are truncated to generate a residual image I. k ; Residual image I k As input for the next round, the updated dynamic bias threshold T is returned until the preset number of iterations is reached, and the final residual image is used as the background estimate. Through the original speckle image I 0 Subtract background estimation Obtain denoised speckle image For denoised speckle images Perform deconvolution to obtain the denoised restored image.

2. The speckle denoising method based on iterative wavelet transform according to claim 1, characterized in that, The original speckle image contains random noise caused by imaging system noise and environmental interference.

3. The speckle denoising method based on iterative wavelet transform according to claim 1, characterized in that, Input speckle image I k Perform multi-scale wavelet decomposition, retain the lowest frequency coefficients, and set the high frequency coefficients to 0, specifically: The input speckle image I was processed using a two-dimensional Daubechies-6 wavelet filter. k Perform an n-level decomposition to obtain the high-frequency coefficients {cD1, cD2, ..., cD} of each level. n } and the lowest frequency coefficient cA n Only the lowest frequency coefficient cA is retained. n All other high-frequency coefficients are set to zero.

4. The speckle denoising method based on iterative wavelet transform according to claim 1, characterized in that, Perform inverse wavelet transform on the lowest frequency coefficients to reconstruct the background estimation image. The expression is: In the formula, IDWT represents the inverse discrete wavelet transform, and k represents the current iteration number. It is the lowest frequency coefficient cA retained in the k-th iteration. n .

5. The speckle denoising method based on iterative wavelet transform according to claim 1, characterized in that, Background estimation image Pixels exceeding the dynamic deviation threshold T are compared and truncated, as expressed by: The lowest frequency coefficient cA n The pixel values ​​are compared with the deviation threshold T. Pixel values ​​exceeding the deviation threshold T are set to the threshold T, while pixel values ​​below the deviation threshold T remain unchanged, thus obtaining the final speckle image I. k .

6. The speckle denoising method based on iterative wavelet transform according to claim 1, characterized in that, Residual image I k As input for the next round, the updated dynamic deviation threshold T is returned until the preset number of iterations is reached, specifically: Set the maximum number of iterations, where k represents the current iteration number. If k < iterations, use the residual image I k as the input for the next round, and return the updated deviation threshold T; If k = iterations, perform background estimation. The expression is:

7. The speckle denoising method based on iterative wavelet transform according to claim 1, characterized in that, For denoised speckle images Perform deconvolution to obtain the denoised restored image. In the formula, deconv(·) represents the deconvolution operation, and PSF represents the point spread function of the system.

8. A speckle denoising system based on iterative wavelet transform, characterized in that, include, The image acquisition module is used to acquire the raw speckle image I. 0 ; The dynamic thresholding module is used to initialize the dynamic deviation threshold T, which is used as the input speckle image I. k The average pixel value; and background estimation image The image is compared with a deviation threshold T, and pixels exceeding the threshold are truncated to generate a residual image I. k ; The wavelet processing module is used to process the input speckle image I k Perform multi-scale wavelet decomposition, retain the lowest frequency coefficients, and set the high frequency coefficients to 0; The reconstruction module performs inverse wavelet transform on the lowest frequency coefficients to reconstruct the background estimation image. The iterative optimization module is used to optimize the residual image I. k As input for the next round, the updated dynamic bias threshold T is returned until the preset number of iterations is reached, and the final residual image is used as the background estimate. Image restoration module, used to restore the original speckle image I 0 Subtract background estimation Obtain denoised speckle image For denoised speckle images Perform deconvolution operation to obtain the target restored image.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of a speckle denoising method based on iterative wavelet transform as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the speckle denoising method based on iterative wavelet transform as described in any one of claims 1-7.

Images