Underwater image enhancement and deblurring method and system based on deep iteration

Through deep iterative optimization of network and fuzzy kernel processing, the problems of insufficient defuzzing capabilities and prior modeling in underwater image enhancement are solved, and efficient image enhancement in complex underwater environments are achieved.

CN118446913BActive Publication Date: 2024-11-22UNIV OF SCI & TECH BEIJING +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410442528.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-12
Publication Date
2024-11-22
Estimated Expiration
2044-04-12

AI Technical Summary

Technical Problem

Existing underwater image enhancement methods lack the ability to defuzz and require specific modeling of priors, making it difficult to accurately model in complex underwater environments.

Method used

Using the underwater image enhancement and defuzzing method based on depth iteration, the underwater imaging model with defuzzing extension is transformed into multivariate joint optimization problem, and the augmented Lagrangian method and alternating The direction multiplier method decomposes the problem, and through deep iterative optimization network solution, a fuzzy kernel is introduced to deal with image blur problem.

Benefits of technology

There is no need to set a prior in advance, effectively overcome the defects of difficulty in accurately modeling a priori, solve the problems of image blur and lack of details, and achieve efficient image enhancement through the deep neural network composed of iterative optimization modules.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118446913B_ABST
    Figure CN118446913B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for underwater image enhancement and deblurring based on deep iteration, which relates to the technical field of underwater image enhancement, including: color correction of an initial underwater image to obtain a corrected underwater image; constructing an underwater imaging model based on the corrected underwater image, and introducing fuzzy terms into the underwater imaging model to obtain a deblurred and extended underwater imaging model; based on the deblurred and extended underwater imaging model, transforming the image enhancement problem of the initial underwater image into a multivariate joint optimization problem; based on an augmented Lagrangian method and an alternating direction multiplier method, decomposing the multivariate joint optimization problem into a plurality of sub-problems; solving the plurality of sub-problems based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image. The present invention alleviates the technical problem that it is difficult to accurately model each prior in a complex underwater environment in the prior art.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of underwater image enhancement, and in particular to an underwater image enhancement and deblurring method and system based on depth iteration. Background Art

[0002] Due to the complexity and particularity of the underwater environment, the images acquired by underwater imaging equipment have problems such as color distortion, low contrast, and blurred structures and details. According to the principle of underwater optical imaging, the amount of light captured by underwater images is mainly composed of three components: direct component (the amount reflected by the object and not scattered), forward scattered component (light reflected by the object and scattered at a small angle), and backscattered component (light reflected by floating particles instead of the target object, which enters the camera after scattering). The underwater image is considered to be the linear weighted sum of these three components. Forward scattering can easily cause problems such as blurring and distortion of underwater images, while backscattering can cover up the structure and details of underwater images. Therefore, underwater image quality improvement (enhancement and deblurring) is a key underwater imaging technology that improves visibility and restores structural details, and has been widely used in many underwater fields such as computational photography, robotic vision, and geomorphic surveying.

[0003] In order to solve the above-mentioned underwater quality degradation problem, researchers have proposed a series of underwater image enhancement methods. Among them, the deep unfolding network based on priors shows significant performance in improving underwater image quality. However, the underwater imaging models in these methods do not have deblurring capabilities, and specific modeling of relevant priors is required. Summary of the invention

[0004] In order to solve the above technical problems existing in the prior art, an embodiment of the present invention provides an underwater image enhancement and deblurring method and system based on depth iteration. The technical solution is as follows:

[0005] On the one hand, a method for underwater image enhancement and deblurring based on deep iteration is provided, the method comprising: performing color correction on an initial underwater image to obtain a corrected underwater image; constructing an underwater imaging model based on the corrected underwater image, and introducing fuzzy terms into the underwater imaging model to obtain an underwater imaging model after deblurring and expansion; based on the underwater imaging model after deblurring and expansion, transforming the image enhancement problem of the initial underwater image into a multivariable joint optimization problem; based on an augmented Lagrangian method and an alternating direction multiplier method, decomposing the multivariable joint optimization problem into multiple sub-problems; solving the multiple sub-problems based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image.

[0006] Furthermore, color correction is performed on the initial underwater image to obtain a corrected underwater image, including: respectively calculating the average value and variance of multiple color channels of the initial underwater image; the multiple color channels include a red channel, a green channel and a blue channel; based on the average value and the variance, calculating the maximum value and the minimum value of each color channel; and color correction is performed on the initial underwater image based on the maximum value, the minimum value and a statistical color correction method to obtain the corrected underwater image.

[0007] Further, based on the mean value and the variance, the maximum value and the minimum value of each color channel are calculated, including:

[0008]

[0009]

[0010] Where c∈{R,G,B}, and are the mean and variance of the c channel, and represent the maximum value and minimum value of the c channel respectively, μ is a parameter for controlling image dynamics, R is the red channel, G is the green channel, and B is the blue channel; a color correction method based on the maximum value, the minimum value and statistics is used to perform color correction on the initial underwater image, including:

[0011]

[0012] Among them, I c is the initial underwater image of channel c, This is the underwater image after correction of the c channel.

[0013] Furthermore, the underwater imaging model after deblurring and expansion includes:

[0014] I(x)=J(x)t(x)+B(1-t(x))+(J(x)·t(x))*g(x)+e(x)

[0015] Among them, I represents the underwater image after correction, J represents the underwater image after enhancement and deblurring, B represents global background light, t represents transmission rate, g represents blur kernel, * represents convolution operation, e represents noise, and x represents the current pixel.

[0016] Furthermore, the multivariable joint optimization problem includes:

[0017]

[0018] Wherein, G represents the matrix form of the blur kernel, f(J) is an implicit sparse representation of the underwater image J prior after the enhancement and deblurring, h(t) is an implicit sparse representation of the transmission rate t prior, It is the L2 norm sparse constraint on the blur kernel G, and α is the weight balance coefficient.

[0019] Furthermore, based on the augmented Lagrangian method and the alternating direction multiplier method, the multivariable joint optimization problem is decomposed into multiple sub-problems, including: solving the multivariable joint optimization problem based on the augmented Lagrangian method to obtain a target solution equation; the target solution equation includes multiple variables to be jointly optimized; based on the alternating direction multiplier method, decomposing the target solution equation into multiple sub-problems regarding the multiple variables to be jointly optimized.

[0020] On the other hand, a deep iteration-based underwater image enhancement and deblurring system is also provided, the system comprising: a color correction module, a model extension module, a transformation module, a decomposition module and a solution module; wherein the color correction module is used to perform color correction on an initial underwater image to obtain a corrected underwater image; the model extension module is used to construct an underwater imaging model based on the corrected underwater image, and introduce fuzzy terms into the underwater imaging model to obtain a deblurred and extended underwater imaging model; the transformation module is used to transform the image enhancement problem of the initial underwater image into a multivariable joint optimization problem based on the deblurred and extended underwater imaging model; the decomposition module is used to decompose the multivariable joint optimization problem into multiple sub-problems based on an augmented Lagrangian method and an alternating direction multiplier method; the solution module is used to solve the multiple sub-problems based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image.

[0021] Furthermore, the color correction module is also used to: respectively calculate the average value and variance of multiple color channels of the initial underwater image; the multiple color channels include a red channel, a green channel and a blue channel; based on the average value and the variance, calculate the maximum value and the minimum value of each color channel; perform color correction on the initial underwater image based on the maximum value, the minimum value and a statistical color correction method to obtain the corrected underwater image.

[0022] Furthermore, the decomposition module is also used to: solve the multivariable joint optimization problem based on the augmented Lagrangian method to obtain a target solution equation; the target solution equation includes multiple variables to be jointly optimized; based on the alternating direction multiplier method, decompose the target solution equation into multiple sub-problems regarding the multiple variables to be jointly optimized.

[0023] On the other hand, a computer-readable storage medium is provided, in which a program code is stored. The program code can be called by a processor to execute the method provided by the present invention.

[0024] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: compared with the existing deep expansion network underwater image enhancement method, the advantage of the present invention is that there is no need to set the specific form of the prior in advance, which effectively overcomes the defect that the respective priors in complex underwater environments are difficult to accurately model; at the same time, the introduction of the blur kernel can effectively solve key problems such as image blur and missing details; in addition, the iterative solution method is expanded into a deep neural network composed of a series of iterative optimization modules, and the solution is accurately and efficiently updated in an end-to-end network full-parameter learning manner, so as to obtain the final ideal enhanced image. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 is a flow chart of an underwater image enhancement and deblurring method based on depth iteration provided by an embodiment of the present invention;

[0027] Figure 2 is a flow chart of another underwater image enhancement and deblurring method based on depth iteration provided by an embodiment of the present invention;

[0028] Figure 3 It is a schematic diagram of an underwater image enhancement and deblurring system based on depth iteration provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0030] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0031] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0032] Embodiment 1

[0033] Figure 1 FIG. 1 is a flowchart of an underwater image enhancement and deblurring method based on depth iteration according to an embodiment of the present invention. Figure 1 As shown, the method specifically comprises the following steps:

[0034] Step S102, color correction is performed on the initial underwater image to obtain a corrected underwater image.

[0035] Step S104: construct an underwater imaging model based on the corrected underwater image, and introduce a fuzzy term into the underwater imaging model to obtain a defuzzified and expanded underwater imaging model.

[0036] Step S106, based on the underwater imaging model after deblurring and expansion, the image enhancement problem of the initial underwater image is transformed into a multivariable joint optimization problem.

[0037] Step S108, based on the augmented Lagrangian method and the alternating direction multiplier method, decompose the multivariable joint optimization problem into multiple sub-problems.

[0038] Step S110, solving multiple sub-problems based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image.

[0039] Figure 2 FIG. 1 is a flowchart of another method for underwater image enhancement and deblurring based on depth iteration provided according to an embodiment of the present invention. Figure 2 As shown, firstly, an initial underwater image I is input as input data.

[0040] Then, the initial underwater image is color corrected to obtain a corrected underwater image, specifically, including:

[0041] The average values ​​and variances of multiple color channels of the initial underwater image are calculated respectively; the multiple color channels include a red channel, a green channel, and a blue channel;

[0042] Based on the mean and variance, calculate the maximum and minimum values ​​of each color channel;

[0043] The color correction method based on maximum value, minimum value and statistics is used to perform color correction on the initial underwater image to obtain a corrected underwater image.

[0044] The specific calculation process is as follows:

[0045]

[0046]

[0047] Where c∈{R,G,B}, and are the mean and variance of the c channel, and Respectively represent the maximum and minimum values ​​of the c channel; μ is a parameter for controlling the image dynamics. Preferably, in the embodiment of the present invention, μ is set to 2.2; R is the red channel, G is the green channel, and B is the blue channel. The statistical color correction method is used to perform color correction on the initial underwater image:

[0048]

[0049] Among them, I c is the initial underwater image of channel c, This is the underwater image after correction of the c channel.

[0050] Specifically, the underwater imaging model after deblurring and expansion includes:

[0051] I(x)=J(x)t(x)+B(1-t(x))+(J(x)·t(x))*g(x)+e(x) (4)

[0052] Among them, I represents the underwater image after correction, J represents the underwater image after enhancement and deblurring, B represents the global background light, t represents the transmission rate, g represents the blur kernel, * represents the convolution operation, e represents the noise, and x represents the current pixel.

[0053] The method provided by the embodiment of the present invention introduces a fuzzy term into the underwater imaging model and performs underwater image deblurring processing.

[0054] Specifically, step S106 is specifically expressed as follows:

[0055]

[0056] Among them, G represents the matrix form of the blur kernel, f(J) is an implicit sparse representation of the underwater image J prior after enhancement and deblurring, and h(t) is an implicit sparse representation of the transmission rate t prior. is to perform L2 norm sparse constraint on blur kernel G, and α is a weight balancing coefficient. Preferably, in the embodiment of the present invention, α is set to 5.

[0057] In the embodiment of the present invention, through the above conversion, there is no need to determine the specific function forms of related a priori such as f and h in advance.

[0058] Specifically, step S108 also includes the following steps:

[0059] Step S1081, solving the multivariable joint optimization problem based on the augmented Lagrangian method to obtain a target solution equation; the target solution equation includes multiple variables to be jointly optimized;

[0060] Step S10082, based on the alternating direction multiplier method, decompose the target solution equation into multiple sub-problems regarding multiple variables to be jointly optimized.

[0061] Specifically, the multivariable joint optimization problem is first solved by the augmented Lagrangian method. The specific calculation is as follows:

[0062]

[0063] Among them, Λ and Υ are two Lagrange multiplier variables, ρ and μ are non-negative penalty balance parameters, and <, > represent matrix inner products. At the same time, let M = J, H = t, and use two auxiliary variables M and H to approximate J and t.

[0064] In the embodiment of the present invention, it is impossible to directly solve the joint optimization of the eight variables in formula (6). The alternating direction multiplier method is used to decompose the joint optimization problem into a series of sub-problems for solving. The specific implementation process is as follows:

[0065] Subproblem B: Fix J, t, G, M, H, Λ, Υ of the nth iteration, and update the background light B as follows:

[0066]

[0067] Where n is the nth iteration.

[0068] Subproblem J: Fix t, G, M, H, Λ, Υ of the nth iteration and B of the n+1th iteration, and the ideal enhanced image J is updated as:

[0069]

[0070] Subproblem t: Fix G, M, H, Λ, Υ of the nth iteration and B, J of the n+1th iteration, and update the transmission rate t to:

[0071]

[0072] M and H subproblems: fix G, M, H, Λ, Υ of the nth iteration and B, J, t of the n+1th iteration, and update the auxiliary variables M and H as follows:

[0073]

[0074]

[0075] In order to overcome the limitation of the existing method that needs to set the specific forms of J and t in advance, a deep neural network is used to solve the regularization function prox c and prox v In formulas (10) and (11), prox c and prox v The deep neural network is learned by training real underwater image datasets.

[0076] In addition, according to the augmented Lagrangian method, the Lagrangian multiplier matrices Λ and Υ are updated:

[0077] Λ n+1 =Λ n +ρ(J n+1 -M n+1 ) (12)

[0078] γ n+1 =γ n +μ(t n+1 -H n+1 ) (13)

[0079] For the G sub-problem, fix J, t, B, G, M, H, Λ, Υ for the n+1th time, and update the fuzzy kernel G as:

[0080]

[0081] However, the blur kernel solved by formula (14) using the image grayscale value usually contains noise and artifacts. Therefore, the present invention uses image gradient information to more effectively and accurately estimate the blur kernel, and uses the data fidelity term of the L2 norm to quickly calculate and solve it as follows:

[0082]

[0083]

[0084] in, is the fast Fourier transform symbol, is the inverse operator of the fast Fourier transform, is the transpose operator of the fast Fourier transform. A multi-scale strategy is used to accurately and effectively estimate the blur kernel. Specifically, by constructing a pyramid sequence of blurred images, the blur kernel with the minimum resolution is first estimated, and then it is iterated and updated until the highest resolution is reached.

[0085] like Figure 2 As shown, step S108 is expanded into a deep iterative optimization network. After iterative updates N=5 times, the ideal image J outputted for the 5th time is used as the underwater image after enhancement and deblurring.

[0086] From the above description, it can be seen that the embodiment of the present invention provides an underwater image enhancement and deblurring method based on deep iteration to overcome the problem that it is difficult to accurately model the priori in complex underwater environments, and effectively deal with the blur problem by introducing a blur kernel; in addition, the training ability of the neural network is used to obtain the best auxiliary variable update solution effect. At the same time, the present invention expands the iterative solution method into a deep neural network composed of a series of iterative optimization modules, and accurately and efficiently iterates and updates through the full parameter learning method of the end-to-end network, so as to obtain the final ideal enhanced image.

[0087] Embodiment 2

[0088] Figure 3 is a schematic diagram of an underwater image enhancement and deblurring system based on depth iteration provided according to an embodiment of the present invention. Figure 3 As shown, the system includes: a color correction module 10, a model expansion module 20, a transformation module 30, a decomposition module 40 and a solution module 50.

[0089] Specifically, the color correction module 10 is used to perform color correction on the initial underwater image to obtain a corrected underwater image.

[0090] The model extension module 20 is used to construct an underwater imaging model based on the corrected underwater image, and introduce fuzzy terms into the underwater imaging model to obtain a defuzzified and extended underwater imaging model.

[0091] The transformation module 30 is used to transform the image enhancement problem of the initial underwater image into a multivariable joint optimization problem based on the underwater imaging model after deblurring and expansion.

[0092] The decomposition module 40 is used to decompose the multivariable joint optimization problem into multiple sub-problems based on the augmented Lagrangian method and the alternating direction multiplier method.

[0093] The solution module 50 is used to solve multiple sub-problems based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image.

[0094] Specifically, the color correction module 10 is also used for:

[0095] The average values ​​and variances of multiple color channels of the initial underwater image are calculated respectively; the multiple color channels include a red channel, a green channel, and a blue channel;

[0096] Based on the mean and variance, calculate the maximum and minimum values ​​of each color channel;

[0097] The color correction method based on maximum value, minimum value and statistics is used to perform color correction on the initial underwater image to obtain a corrected underwater image.

[0098] Specifically, the underwater imaging model after deblurring and expansion includes:

[0099] I(x)=J(x)t(x)+B(1-t(x))+(J(x)·t(x))*g(x)+e(x)

[0100] Among them, I represents the underwater image after correction, J represents the underwater image after enhancement and deblurring, B represents the global background light, t represents the transmission rate, g represents the blur kernel, * represents the convolution operation, e represents the noise, and x represents the current pixel.

[0101] Specifically, multivariable joint optimization problems include:

[0102]

[0103] Among them, G represents the matrix form of the blur kernel, f(J) is an implicit sparse representation of the underwater image J prior after enhancement and deblurring, and h(t) is an implicit sparse representation of the transmission rate t prior. It is the L2 norm sparse constraint on the blur kernel G, and α is the weight balance coefficient.

[0104] Specifically, the decomposition module 40 is further used for:

[0105] The multivariable joint optimization problem is solved based on the augmented Lagrangian method to obtain the target solution equation; the target solution equation includes multiple variables to be jointly optimized;

[0106] Based on the alternating direction multiplier method, the target solution equation is decomposed into multiple sub-problems regarding multiple variables to be jointly optimized.

[0107] The embodiment of the present invention further provides a computer-readable storage medium, in which program codes are stored. The program codes can be called by a processor to execute the method in the first embodiment.

[0108] It should be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM).

[0109] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0110] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0111] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0112] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0113] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0114] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0115] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0116] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0117] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for underwater image enhancement and deblurring based on deep iteration, characterized in that: The method comprises: Performing color correction on the initial underwater image to obtain the corrected underwater image; comprising: respectively calculating the average value and variance of multiple color channels of the initial underwater image; the multiple color channels include a red channel, a green channel and a blue channel; calculating the maximum value and the minimum value of each color channel based on the average value and the variance; performing color correction on the initial underwater image based on the maximum value, the minimum value and a statistical color correction method to obtain the corrected underwater image; Wherein, based on the mean value and the variance, calculating the maximum value and the minimum value of each color channel includes: Where c∈{R,G,B}, and are the mean and variance of the c channel, and represent the maximum and minimum values ​​of the c channel respectively, μ is a parameter for controlling image dynamics, R is the red channel, G is the green channel, and B is the blue channel; Performing color correction on the initial underwater image based on the maximum value, the minimum value and a statistical color correction method includes: Among them, I c is the initial underwater image of channel c, is the underwater image after correction of the c channel; Building an underwater imaging model based on the corrected underwater image, and introducing a fuzzy term into the underwater imaging model to obtain a defuzzified and expanded underwater imaging model; The underwater imaging model after deblurring and expansion includes: I(x)=J(x)t(x)+B(1-t(x))+(J(x)·t(x))*g(x)+e(x) Wherein, I represents the underwater image after correction, J represents the underwater image after enhancement and deblurring, B represents global background light, t represents transmission rate, g represents blur kernel, * represents convolution operation, e represents noise, and x represents current pixel point; Based on the underwater imaging model after deblurring and expansion, the image enhancement problem of the initial underwater image is transformed into a multivariable joint optimization problem; The multivariable joint optimization problem includes: Wherein, G represents the matrix form of the blur kernel, f(J) is an implicit sparse representation of the underwater image J prior after the enhancement and deblurring, h(t) is an implicit sparse representation of the transmission rate t prior, It is the L2 norm sparse constraint on the blur kernel G, and α is the weight balance coefficient; Based on the augmented Lagrangian method and the alternating direction multiplier method, the multivariable joint optimization problem is decomposed into multiple sub-problems; The multiple sub-problems are solved based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image.

2. The method according to claim 1, characterized in that Based on the augmented Lagrangian method and the alternating direction multiplier method, the multivariable joint optimization problem is decomposed into multiple sub-problems, including: Solving the multivariable joint optimization problem based on the augmented Lagrangian method to obtain a target solution equation; the target solution equation includes a plurality of variables to be jointly optimized; Based on the alternating direction multiplier method, the target solution equation is decomposed into a plurality of sub-problems regarding the plurality of variables to be jointly optimized.

3. An underwater image enhancement and deblurring system based on depth iteration, the system being used to implement the method according to any one of claims 1 to 2, characterized in that: The system comprises: a color correction module, a model expansion module, a transformation module, a decomposition module and a solution module; wherein, The color correction module is used to perform color correction on the initial underwater image to obtain a corrected underwater image; The model extension module is used to construct an underwater imaging model based on the corrected underwater image, and introduce a fuzzy term into the underwater imaging model to obtain a defuzzified and extended underwater imaging model; The transformation module is used to transform the image enhancement problem of the initial underwater image into a multivariable joint optimization problem based on the underwater imaging model after deblurring and expansion; The decomposition module is used to decompose the multivariable joint optimization problem into multiple sub-problems based on the augmented Lagrangian method and the alternating direction multiplier method; The solution module is used to solve the multiple sub-problems based on a deep iterative optimization network to obtain an enhanced and deblurred underwater image.

4. The system according to claim 3, characterized in that The color correction module is further used for: Respectively calculating the average value and variance of multiple color channels of the initial underwater image; the multiple color channels include a red channel, a green channel and a blue channel; Based on the mean value and the variance, calculating the maximum value and the minimum value of each color channel; The initial underwater image is color corrected by a color correction method based on the maximum value, the minimum value and statistics to obtain the corrected underwater image.

5. The system according to claim 3, characterized in that The decomposition module is further used for: Solving the multivariable joint optimization problem based on the augmented Lagrangian method to obtain a target solution equation; the target solution equation includes a plurality of variables to be jointly optimized; Based on the alternating direction multiplier method, the target solution equation is decomposed into a plurality of sub-problems regarding the plurality of variables to be jointly optimized.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code, and the program code can be called by a processor to execute the method as described in any one of claims 1-2.