A method and system for underwater target clear imaging based on Mueller matrix and decomposition

Abstract

The invention discloses a method and system for clear imaging of underwater targets based on Mueller matrix sum decomposition, comprising: building a Mueller matrix underwater optical imaging system to obtain a Mueller matrix image of an underwater target, performing Mueller matrix sum decomposition on the image, and obtaining sum decomposition parameters; performing matrix filtering according to the sum decomposition parameters to obtain a new Mueller matrix, and calculating polarization entropy and depolarization index of the acquired image; establishing a polarization entropy background distribution model to automatically select a background area, setting the light intensity average value of the automatically selected background area to the backscattered light intensity at infinity; establishing a depolarization coefficient correction model, and correcting the depolarization coefficient with an optimal coefficient when the information entropy of the restored image is maximum; calculating and obtaining a restored image of the underwater target according to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity; and performing histogram equalization on the restored image of the underwater target to obtain a final clear image of the underwater target.

Description

Technical Field

[0001] The present invention belongs to the technical field of underwater polarization imaging, and in particular relates to a method and system for underwater target clear imaging based on Mueller matrix and decomposition. Background Art

[0002] Underwater imaging technology has important applications in many fields such as marine resource exploration and underwater archaeology. However, due to the influence of water absorption and scattering, the image contrast obtained by underwater imaging technology is reduced, detail information is lost, and the imaging quality is seriously affected. In underwater imaging technology, polarization imaging technology improves the dimension of target information and makes full use of the partial polarization characteristics of backscattered light to restore images affected by water absorption and scattering. At present, underwater polarization imaging technology mainly relies on images in four basic polarization directions to capture and analyze the polarization characteristics of light. Although this method can reveal certain characteristics of the underwater environment to a certain extent, it does not make full use of the polarization information of light. Since light is affected by various factors such as scattering and absorption when propagating underwater, it is impossible to fully analyze these complex physical processes by relying solely on images in four polarization directions. Summary of the invention

[0003] Purpose of the invention: In order to solve the problem that the existing underwater polarization imaging technology does not fully utilize the polarization information of light, the present invention proposes a method and system for underwater target clarity imaging based on Mueller matrix and decomposition. The method of the present invention analyzes the depolarization characteristics from the perspective of matrix and decomposition and more comprehensively analyzes and utilizes polarization information.

[0004] Technical solution: A method for underwater target clear imaging based on Mueller matrix and decomposition, comprising the following steps:

[0005] Build a Mueller matrix underwater optical imaging system;

[0006] In the constructed Mueller matrix underwater optical imaging system, multiple polarization images of underwater targets are obtained; based on the multiple polarization images of underwater targets, a Mueller matrix is ​​obtained, and Mueller matrix sum decomposition is performed on it to obtain sum decomposition parameters; matrix filtering is performed according to the sum decomposition parameters to obtain a new Mueller matrix;

[0007] According to the decomposition parameters, the polarization entropy and depolarization index are calculated;

[0008] A polarization entropy background distribution model is established and used to automatically select a background area, and the average light intensity of the automatically selected background area is set to be the backscattered light intensity at infinity;

[0009] A depolarization coefficient correction model is established, information entropy is determined as the image quality evaluation index function, and the depolarization coefficient is corrected with the optimal coefficient when the information entropy of the restored image is the maximum value to obtain the corrected depolarization coefficient.

[0010] On the basis of the classic underwater physical imaging model, the first element of the new Mueller matrix is ​​used as the total light intensity of the underwater image obtained by the Mueller matrix underwater optical imaging system. The restored image of the underwater target is calculated according to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity.

[0011] The restored image of the underwater target is histogram equalized to obtain the final clear image of the underwater target.

[0012] Further, the Mueller matrix underwater optical imaging system includes an LED red light source, a first polarizer, a first 1 / 4 wave plate, a transparent water tank, an underwater target, a scattering medium, a second 1 / 4 wave plate, a second polarizer and a CCD detector;

[0013] The scattering medium is used to simulate a turbid water environment, and the underwater target and the scattering medium are both located in a transparent water tank;

[0014] The LED red light source emits a starting light beam, which passes through the first polarizer and the first quarter wave plate in sequence, and then passes through the transparent water tank to irradiate the underwater target. The target reflected light reflected by the underwater target and the backscattered light scattered by the scattering medium pass through the second quarter wave plate and the second polarizer in sequence, and then are irradiated onto the CCD detector.

[0015] The polarization directions of the first polarizer and the second polarizer are both horizontal. Furthermore, in the constructed Mueller matrix underwater optical imaging system, multiple polarization images of underwater targets are obtained; based on the multiple polarization images of the underwater targets, a Mueller matrix is ​​obtained, and the Mueller matrix is ​​summed and decomposed to obtain sum and decomposition parameters; the specific operations include:

[0016] The first 1 / 4 wave plate and the second 1 / 4 wave plate are rotated at a set rotation speed ratio to obtain multiple polarization images of the target object;

[0017] Based on multiple polarization images of the target object, the Mueller matrix is ​​obtained according to the double-rotation wave plate method;

[0018] Using the Pauli matrix σ as the basis, the Mueller matrix is ​​mapped to the Hermitian matrix H(M), the Mueller matrix and decomposition are realized, and the decomposition parameters are obtained:

[0019]

[0020] In the formula, m ij is the Mueller matrix element; σ i represents the ith Pauli matrix, σ j represents the j-th Pauli matrix.

[0021] Furthermore, the matrix filtering is performed according to the decomposition parameters to obtain a new Mueller matrix, and the specific operations include:

[0022] According to the eigenvalues ​​of the Hermitian matrix H(M) and the eigenvectors corresponding to the eigenvalues ​​of the Hermitian matrix H(M), matrix filtering is performed to obtain a new Mueller matrix M′, which is expressed as:

[0023]

[0024] In the formula, λ i is the eigenvalue of the Hermitian matrix H(M), M i is the eigenvalue λ of the Hermitian matrix H(M) i The corresponding eigenvector u i The matrix restored by the outer product is expressed as:

[0025]

[0026] In the formula, m 11 represents the first element of the Mueller matrix, represents Hermitian conjugation.

[0027] Further, the polarization entropy and depolarization index are calculated based on the decomposition parameters, specifically including:

[0028] According to the eigenvalue λ of the Hermitian matrix H(M) i Calculate the polarization entropy S(H), expressed as:

[0029]

[0030] According to the eigenvalue λ of the Hermitian matrix H(M) i Calculate the depolarization index P V , expressed as:

[0031]

[0032] Where P 1 , P 2 , P 3 is the polarization purity index, and its expression is:

[0033]

[0034] In the formula, m 11 Represents the first element of the Mueller matrix.

[0035] Furthermore, the establishment and use of the polarization entropy background distribution model to automatically select a background area, and setting the average light intensity of the automatically selected background area to the backscattered light intensity at infinity, specifically includes:

[0036] For an image, each pixel has a corresponding polarization entropy S(H). The polarization entropy of each pixel is calculated to obtain the polarization entropy image corresponding to the image.

[0037] The polarization entropy background distribution model determines the background area range according to the median and interquartile range of the polarization entropy image, and selects the maximum connected graph of pixels in the background area range with a box as the selected background area;

[0038] Calculate the average light intensity of the background area as the backscattered light intensity A at infinity ∞ .

[0039] Furthermore, in the depolarization coefficient correction model, the depolarization coefficient is initialized and expressed as:

[0040] Dep(x)=1-P V .

[0041] Furthermore, the information entropy is determined as an image quality evaluation index function, which is expressed as:

[0042]

[0043] In the formula, p i Indicates the probability of a certain grayscale appearing in the image;

[0044] The depolarization coefficient is corrected by using the optimal coefficient when the information entropy of the restored image is at the maximum value, and the corrected depolarization coefficient (λ) is obtained. optimal Dep(x), where (λ) optimal =argmax{Entropy(L(x))}, L(x) is the restored image of the underwater target, and Entropy(L(x)) represents the information entropy of L(x).

[0045] Furthermore, based on the classical underwater physical imaging model, the first element of the new Mueller matrix is ​​used as the total light intensity of the underwater image acquired by the Mueller matrix underwater optical imaging system. According to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity, the restored image of the underwater target is calculated, which is expressed as:

[0046]

[0047] Among them, m′ 11 is the first element of the new Mueller matrix.

[0048] The present invention discloses an underwater target clear imaging system based on Mueller matrix and decomposition, comprising:

[0049] The Mueller matrix acquisition module is used to acquire multiple polarization images of underwater targets in the constructed Mueller matrix underwater optical imaging system; based on the multiple polarization images of the underwater targets, the Mueller matrix is ​​obtained;

[0050] The matrix decomposition processing module is used to perform Mueller matrix sum decomposition on the Mueller matrix to obtain sum decomposition parameters; perform matrix filtering according to the sum decomposition parameters to obtain a new Mueller matrix; and calculate polarization entropy and depolarization index according to the sum decomposition parameters;

[0051] The background area selection module is used to establish and use the polarization entropy background distribution model to automatically select the background area, and set the average light intensity of the automatically selected background area to the backscattered light intensity at infinity;

[0052] A depolarization coefficient calculation module is used to establish a depolarization coefficient correction model, determine information entropy as an image quality evaluation index function, and correct the depolarization coefficient with the optimal coefficient when the information entropy of the restored image is the maximum value to obtain a corrected depolarization coefficient;

[0053] The underwater target clear image calculation module is used to calculate the underwater target restoration image based on the classic underwater physical imaging model, taking the first element of the new Mueller matrix as the total light intensity of the underwater image obtained by the Mueller matrix underwater optical imaging system, and according to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity; the underwater target restoration image is histogram equalized to obtain the final underwater target clear image.

[0054] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0055] (1) The present invention introduces polarization entropy, which has advantages in target detection, classification and imaging system performance optimization by quantifying the polarization characteristics of underwater targets;

[0056] (2) The present invention introduces the Mueller matrix, which can fully characterize the change in polarization state after the interaction between light and matter, and can obtain more comprehensive polarization information;

[0057] (3) The present invention introduces polarization entropy for background area selection. The background area can be automatically selected through the polarization entropy background distribution model, and the depolarization index is calculated using the Mueller matrix and the polarization purity index obtained by decomposition to achieve underwater clear imaging. The depolarization characteristics can be analyzed from the perspective of matrix and decomposition, and the polarization information can be used for more comprehensive analysis and calculation, combined with various parameters of the Mueller matrix, to greatly improve the underwater imaging quality;

[0058] (4) The present invention can automatically select the background area, use the Mueller matrix and the polarization purity index obtained by decomposition to calculate the depolarization index to achieve underwater clear imaging, and provide a more comprehensive image quality improvement solution by combining multiple parameters of the Mueller matrix for comprehensive analysis and calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A flowchart of a method for underwater target clear imaging based on Mueller matrix and decomposition according to the present invention;

[0060] Figure 2 A model diagram of the Mueller matrix underwater optical imaging system constructed for the present invention;

[0061] Figure 3 This is a diagram showing the effect of high and low concentration experiments using the method of the present invention, wherein: Figure 3 (a) is the total light intensity image underwater. Figure 3 (b) is a diagram showing the effect of the method of the present invention. DETAILED DESCRIPTION

[0062] The technical solution of the present invention is now further described in conjunction with the accompanying drawings and embodiments.

[0063] like Figure 1 As shown, this embodiment discloses a method for underwater target clear imaging based on Mueller matrix and decomposition, which mainly includes the following steps:

[0064] Step 1: Build a Mueller matrix underwater optical imaging system, obtain the Mueller matrix and perform sum decomposition, perform matrix filtering based on the parameters obtained by sum decomposition, and calculate the polarization entropy and depolarization index; the specific operations include:

[0065] Figure 2 A Mueller matrix underwater optical imaging system is shown, in which an LED red light source 1 emits a starting light beam, which passes through a first polarizer 2, whose polarization direction is set to the horizontal direction, and then passes through a rotatable first 1 / 4 wave plate 3, and then through a transparent water tank 4, and finally irradiates a non-uniform polarization characteristic target 6 and a scattering medium 5. To simulate a turbid water environment, 40 ml of milk is added to the water. The non-uniform polarization characteristic target 6 is composed of a metal coin with a high polarization degree characteristic and a plastic magic cube with a low polarization degree characteristic bonded together. When the starting light beam irradiates the non-uniform polarization characteristic target 6, the non-uniform polarization characteristic target 6 will reflect the target reflected light, and the scattering medium 5 will scatter the backscattered light. These target reflected light and backscattered light are processed by a rotatable second 1 / 4 wave plate 7 and a second polarizer 8, and then irradiated onto a CCD detector 9 for recording and analysis. The polarization direction of the second polarizer is horizontal.

[0066] The total light intensity of the underwater image received by the CCD detector 9 has the following relationship with the target reflected light and backscattered light:

[0067] I=D+B

[0068] D(x)=L(x)·t(x)

[0069] B(x)=A ∞ ·[1-t(x)]

[0070] Where I is the total light intensity of the underwater image, D is the target reflected light, B is the backscattered light, L(x) is the restored image, t(x) is the transmittance map, A ∞ is the backscattered light intensity at infinity, D(x) is the target reflected light image, and B(x) is the backscattered light image;

[0071] The first 1 / 4 wave plate 3 and the second 1 / 4 wave plate 7 are rotated at a rotation speed ratio of 5:1 by using an electric rotation stage to take 36 polarization sub-images. In this embodiment, the polarization sub-images are taken by a Daheng Optoelectronics MER2-2000-19U3M / CL camera, and the Mueller matrix is ​​obtained according to the double rotation wave plate method. The Mueller matrix is ​​decomposed, and the Pauli matrix σ is first used as the basis, and then the Mueller matrix is ​​mapped to the Hermitian matrix H(M):

[0072]

[0073] In the formula, m ij is the Mueller matrix element; σ i represents the ith Pauli matrix, σ j represents the j-th Pauli matrix.

[0074] According to H(M), its eigenvalues ​​and eigenvectors are obtained, and the new Mueller matrix M′ is calculated. This process is matrix filtering:

[0075]

[0076] In the formula, λ i is the eigenvalue of the Hermitian matrix H(M), M i is the eigenvalue λ i The corresponding eigenvector u i The matrix restored by the outer product is expressed as:

[0077]

[0078] In the formula, m 11 represents the first element of the Mueller matrix, represents Hermitian conjugation.

[0079] Calculate the polarization entropy S(H), the expression is:

[0080]

[0081] Calculate the depolarization index P V , the expression is:

[0082]

[0083] Where P 1 , P 2 , P 3 is the polarization purity index, and its expression is:

[0084]

[0085] In the formula, m 11 Represents the first element of the Mueller matrix.

[0086] Step 2: Select the first element of the new Mueller matrix M′ as the total light intensity, select the background area through the polarization entropy background distribution model and estimate the backscattered light intensity A at infinity ∞ , using the depolarization index P V Calculate the depolarization coefficient; the specific operations include:

[0087] For an image, each pixel has a corresponding polarization entropy S(H). By calculating the polarization entropy of each pixel, the polarization entropy image corresponding to the image can be obtained. The polarization entropy background distribution model is constructed according to the median and interquartile range of the polarization entropy image. Specifically, the polarization entropy background distribution model uses the median and interquartile range of the polarization entropy image for weighted calculation to determine a range.

[0088] The background area is automatically selected using the polarization entropy background distribution model; specifically, the maximum connected graph of pixels in the background area is selected with a box and used as the selected background area.

[0089] Calculate the average light intensity of the background area as the backscattered light intensity A at infinity ∞ , which is to calculate the average brightness of the pixels in the selected area.

[0090] Step 3: Establish a depolarization coefficient correction model. First, initialize the depolarization coefficient Dep(x), which is expressed as:

[0091] Dep(x)=1-P V

[0092] Select information entropy Entropy as the evaluation index of image quality, select the coefficient when Entropy is the maximum value to correct the depolarization coefficient Dep(x), and obtain the transmittance map; the specific operations include:

[0093] Information entropy is expressed as:

[0094]

[0095] In the formula, p i It indicates the probability of a certain grayscale appearing in the image, which can be obtained from the grayscale histogram. The greater the information entropy of the image, the richer the pixel grayscale contained in the image and the more uniform the grayscale distribution.

[0096] The objective function is expressed as:

[0097] (λ) optimal =argmax{Entropy(L(x))}

[0098] Where L(x) is the restored image of the underwater target, Entropy(L(x)) represents the information entropy of L(x), (λ) optimal represents the optimal parameters.

[0099] The optimal solution of the objective function is found by adopting a step-by-step approach, with a value range of 0 to 1 and a step size of 0.05.

[0100] The final optimal parameter (λ) optimal is 0.55. In this embodiment, the maximum information entropy of the restored image of the target is 7.2025, and the information entropy of the total light intensity map is 5.4977.

[0101] The coefficient when Entropy is at its maximum value is selected to correct the depolarization coefficient Dep(x) and obtain the transmittance graph, which is expressed as:

[0102] t(x)=1-(λ) optimal Dep(x)

[0103] Where t(x) is the transmittance graph.

[0104] Step 4: Based on the obtained transmittance graph, the total light intensity of the underwater image received by the CCD detector 9 and the backscattered light intensity A at infinity ∞ , based on the classic underwater physical imaging model, the clear image L(x) of the underwater target is calculated and expressed as:

[0105]

[0106] Among them, m 1 ' 1 is the first element of the new Mueller matrix M′, which is the total light intensity received at pixel x of the CCCD detector 9.

[0107] The underwater target restoration image L(x) is histogram equalized to obtain the final image.

[0108] To verify the effectiveness of this embodiment, an experiment was conducted in turbid water. Figure 3 (a) is a total light intensity image in turbid water. Subjectively, the image is not very clear and the image details are seriously lost. After being processed by the method of this embodiment, the image clarity and visibility are significantly improved. Figure 3 As shown in (b), it is possible to realize automatic restoration of underwater images of targets with complex polarization characteristics and is suitable for underwater environments with different turbidity.

[0109] Information entropy (Entorpy), image enhancement evaluation criteria (EME) and average gradient (AG) are used to objectively evaluate the quality of the restored image. The larger the value, the higher the image quality. The results are shown in Table 1.

[0110] Table 1 Quality evaluation of restored images of complex polarization targets in turbid water

[0111] EME Entorpy AG Total light intensity diagram 9.7026 5.4977 0.0078 After adopting the method of this embodiment 33.4185 7.2025 0.0378

[0112] It can be seen from the table that the restored image obtained by the method of this embodiment is significantly improved in various objective evaluation indicators compared with the original total light intensity image, which verifies the effectiveness of the method of this embodiment in achieving clear underwater imaging.

[0113] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0114] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for underwater target clear imaging based on Mueller matrix and decomposition, characterized by: The following steps are involved: Build a Mueller matrix underwater optical imaging system; In the constructed Mueller matrix underwater optical imaging system, multiple polarization images of underwater targets are obtained; based on the multiple polarization images of underwater targets, a Mueller matrix is ​​obtained, and Mueller matrix sum decomposition is performed on it to obtain sum decomposition parameters; matrix filtering is performed according to the sum decomposition parameters to obtain a new Mueller matrix; According to the decomposition parameters, the polarization entropy and depolarization index are calculated; A polarization entropy background distribution model is established and used to automatically select a background area, and the average light intensity of the automatically selected background area is set to be the backscattered light intensity at infinity; A depolarization coefficient correction model is established, information entropy is determined as the image quality evaluation index function, and the depolarization coefficient is corrected with the optimal coefficient when the information entropy of the restored image is the maximum value to obtain the corrected depolarization coefficient. On the basis of the classic underwater physical imaging model, the first element of the new Mueller matrix is ​​used as the total light intensity of the underwater image obtained by the Mueller matrix underwater optical imaging system. The restored image of the underwater target is calculated according to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity. Perform histogram equalization on the restored image of the underwater target to obtain the final clear image of the underwater target; The Mueller matrix underwater optical imaging system includes an LED red light source, a first polarizer, a first 1 / 4 wave plate, a transparent water tank, an underwater target, a scattering medium, a second 1 / 4 wave plate, a second polarizer and a CCD detector; The scattering medium is used to simulate a turbid water environment, and the underwater target and the scattering medium are both located in a transparent water tank; The LED red light source emits a starting light beam, which passes through the first polarizer and the first quarter wave plate in sequence, and then passes through the transparent water tank to irradiate the underwater target. The target reflected light reflected by the underwater target and the backscattered light scattered by the scattering medium pass through the second quarter wave plate and the second polarizer in sequence, and then are irradiated onto the CCD detector. The polarization directions of the first polarizer and the second polarizer are both horizontal; In the constructed Mueller matrix underwater optical imaging system, multiple polarization images of underwater targets are obtained; Based on multiple polarization images of underwater targets, the Mueller matrix is ​​obtained, and the Mueller matrix is ​​decomposed to obtain the decomposition parameters; the specific operations include: The first 1 / 4 wave plate and the second 1 / 4 wave plate are rotated at a set rotation speed ratio to obtain multiple polarization images of the target object; Based on multiple polarization images of the target object, the Mueller matrix is ​​obtained according to the double-rotation wave plate method; Using the Pauli matrix σ as the basis, the Mueller matrix is ​​mapped to the Hermitian matrix H(M), the Mueller matrix and decomposition are realized, and the decomposition parameters are obtained: In the formula, m ij is the Mueller matrix element; σ i represents the ith Pauli matrix, σ j represents the jth Pauli matrix; The establishment and use of the polarization entropy background distribution model to automatically select a background area, and setting the average light intensity of the automatically selected background area to the backscattered light intensity at infinity, specifically includes: For an image, each pixel has a corresponding polarization entropy. The polarization entropy of each pixel is calculated to obtain the polarization entropy image corresponding to the image. The polarization entropy background distribution model determines the background area range according to the median and interquartile range of the polarization entropy image, and selects the maximum connected graph of pixels in the background area range with a box as the selected background area; Calculate the average light intensity of the background area as the backscattered light intensity at infinity; The matrix filtering is performed according to the decomposition parameters to obtain a new Mueller matrix. The specific operations include: According to the eigenvalues ​​of the Hermitian matrix H(M) and the eigenvectors corresponding to the eigenvalues ​​of the Hermitian matrix H(M), matrix filtering is performed to obtain a new Mueller matrix M′, which is expressed as: In the formula, λ i is the eigenvalue of the Hermitian matrix H(M), M i is the eigenvalue λ of the Hermitian matrix H(M) i The corresponding eigenvector u i The matrix restored by the outer product is expressed as: In the formula, m 11 represents the first element of the Mueller matrix, represents Hermitian conjugation.

2. The underwater target clear imaging method based on Mueller matrix and decomposition according to claim 1, characterized in that: The polarization entropy and depolarization index are calculated based on the decomposition parameters, specifically including: According to the eigenvalue λ of the Hermitian matrix H(M) i Calculate the polarization entropy S(H), expressed as: According to the eigenvalue λ of the Hermitian matrix H(M) i Calculate the depolarization index P V , expressed as: In the formula, P1, P2, and P3 are polarization purity indicators, and the expression is: In the formula, m 11 Represents the first element of the Mueller matrix.

3. The underwater target clear imaging method based on Mueller matrix and decomposition according to claim 1, characterized in that: In the depolarization coefficient correction model, the depolarization coefficient is initialized and expressed as: Dep(x)=1-P V 。 4. The underwater target clear imaging method based on Mueller matrix and decomposition according to claim 3 is characterized in that: The information entropy is determined as the image quality evaluation index function, which is expressed as: In the formula, p i Indicates the probability of a certain grayscale appearing in the image; The depolarization coefficient is corrected by using the optimal coefficient when the information entropy of the restored image is at the maximum value, and the corrected depolarization coefficient (λ) is obtained. optimal Dep(x), where (λ) optimal =argmax{Entropy(L(x))}, L(x) is the restored image of the underwater target, and Entropy(L(x)) represents the information entropy of L(x).

5. The underwater target clear imaging method based on Mueller matrix and decomposition according to claim 4, characterized in that: Based on the classical underwater physical imaging model, the first element of the new Mueller matrix is ​​used as the total light intensity of the underwater image acquired by the Mueller matrix underwater optical imaging system. According to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity, the restored image of the underwater target is calculated, which is expressed as: Among them, m1′1 is the first element of the new Mueller matrix.

6. An underwater target clear imaging system based on Mueller matrix and decomposition, characterized by: include: The Mueller matrix acquisition module is used to acquire multiple polarization images of underwater targets in the constructed Mueller matrix underwater optical imaging system; Based on multiple polarization images of underwater targets, the Mueller matrix is ​​obtained; A matrix decomposition processing module is used to perform Mueller matrix sum decomposition on the Mueller matrix to obtain sum decomposition parameters; Perform matrix filtering according to the sum decomposition parameters to obtain a new Mueller matrix; and calculate polarization entropy and depolarization index according to the sum decomposition parameters; A background area selection module is used to establish and use a polarization entropy background distribution model to automatically select a background area, and set the average light intensity of the automatically selected background area to the backscattered light intensity at infinity; A depolarization coefficient calculation module is used to establish a depolarization coefficient correction model, determine information entropy as an image quality evaluation index function, and correct the depolarization coefficient with the optimal coefficient when the information entropy of the restored image is the maximum value to obtain a corrected depolarization coefficient; The underwater target clear image calculation module is used to calculate the underwater target restoration image based on the classic underwater physical imaging model, taking the first element of the new Mueller matrix as the total light intensity of the underwater image obtained by the Mueller matrix underwater optical imaging system, and according to the corrected depolarization coefficient, the total light intensity of the underwater image and the backscattered light intensity at infinity; Perform histogram equalization on the restored image of the underwater target to obtain the final clear image of the underwater target; The Mueller matrix underwater optical imaging system includes an LED red light source, a first polarizer, a first 1 / 4 wave plate, a transparent water tank, an underwater target, a scattering medium, a second 1 / 4 wave plate, a second polarizer and a CCD detector; The scattering medium is used to simulate a turbid water environment, and the underwater target and the scattering medium are both located in a transparent water tank; The LED red light source emits a starting light beam, which passes through the first polarizer and the first quarter wave plate in sequence, and then passes through the transparent water tank to irradiate the underwater target. The target reflected light reflected by the underwater target and the backscattered light scattered by the scattering medium pass through the second quarter wave plate and the second polarizer in sequence, and then are irradiated onto the CCD detector. The polarization directions of the first polarizer and the second polarizer are both horizontal; In the constructed Mueller matrix underwater optical imaging system, multiple polarization images of underwater targets are obtained; Based on multiple polarization images of underwater targets, the Mueller matrix is ​​obtained, and the Mueller matrix is ​​decomposed to obtain the decomposition parameters; the specific operations include: The first 1 / 4 wave plate and the second 1 / 4 wave plate are rotated at a set rotation speed ratio to obtain multiple polarization images of the target object; Based on multiple polarization images of the target object, the Mueller matrix is ​​obtained according to the double-rotation wave plate method; Using the Pauli matrix σ as the basis, the Mueller matrix is ​​mapped to the Hermitian matrix H(M), the Mueller matrix and decomposition are realized, and the decomposition parameters are obtained: In the formula, m ij is the Mueller matrix element; σ i represents the ith Pauli matrix, σ j represents the jth Pauli matrix; The establishment and use of the polarization entropy background distribution model to automatically select a background area, and setting the average light intensity of the automatically selected background area to the backscattered light intensity at infinity, specifically includes: For an image, each pixel has a corresponding polarization entropy. The polarization entropy of each pixel is calculated to obtain the polarization entropy image corresponding to the image. The polarization entropy background distribution model determines the background area range according to the median and interquartile range of the polarization entropy image, and selects the maximum connected graph of pixels in the background area range with a box as the selected background area; Calculate the average light intensity of the background area as the backscattered light intensity at infinity; The matrix filtering is performed according to the decomposition parameters to obtain a new Mueller matrix. The specific operations include: According to the eigenvalues ​​of the Hermitian matrix H(M) and the eigenvectors corresponding to the eigenvalues ​​of the Hermitian matrix H(M), matrix filtering is performed to obtain a new Mueller matrix M′, which is expressed as: In the formula, λ i is the eigenvalue of the Hermitian matrix H(M), M i is the eigenvalue λ of the Hermitian matrix H(M) i The corresponding eigenvector u i The matrix restored by the outer product is expressed as: In the formula, m 11 represents the first element of the Mueller matrix, represents Hermitian conjugation.

Images