An underwater polarization imaging method based on target light polarization degree restriction

Abstract

The present invention discloses an underwater polarization imaging method based on the target light polarization degree limitation condition, including: acquiring the collinear polarization image I<supgt;∥< / supgt; and the cross-polarization image I<supgt;⊥< / supgt>, establishing a target light imaging model and a backscattered light imaging model; using the mutual information function to characterize the correlation between the target light imaging image and the backscattered light imaging image, establishing an underwater image contrast evaluation function of the target light imaging image, taking the boundary-limited target light polarization degree and the cut-off frequency in the backscattered light imaging model as independent variables, taking the mutual information function and the underwater image contrast evaluation function as objective functions, and using a bi-objective optimization algorithm to find the optimal solution set that satisfies the minimum mutual information and the maximum underwater image contrast evaluation function of the target light imaging image; substituting the optimal solution set into the target light imaging model in sequence to obtain an underwater restoration image set, and taking the underwater restoration image with the maximum image quality evaluation value from the underwater restoration image set as the final underwater restoration image.

Description

Technical Field

[0001] The invention relates to the technical field of underwater polarization imaging, and in particular to an underwater polarization imaging method based on a target light polarization degree restriction condition. Background Art

[0002] Underwater visual tasks are affected by water absorption and scattering of turbid media, resulting in reduced image contrast and difficulty in distinguishing details. The main problem is the interference of backscattered light. To solve this problem, underwater polarization imaging technology makes full use of the partial polarization characteristics of backscattered light and successfully restores images affected by the underwater environment. The use of this technology helps to reduce scattering interference in underwater imaging, improve image quality, and enhance the ability to observe and understand the underwater environment.

[0003] At present, people use the difference between the target reflected light and backscattered light in the frequency domain to suppress the backscattered light. However, the current frequency-domain-based underwater polarization imaging method can only be applied to the underwater image restoration of targets with a single polarization characteristic, and cannot achieve effective restoration of targets with complex polarization characteristics. In addition, the single objective function and the optimization without limiting the polarization degree of the target light lead to overexposure or too dark restored images, which cannot improve the final imaging quality. At the same time, the need for prior knowledge and human-computer interaction in traditional underwater polarization imaging limits the scope of application of underwater imaging technology, and cannot achieve automatic restoration of underwater polarization images. Summary of the invention

[0004] Purpose of the invention: To solve the problems that traditional underwater polarization imaging methods can only be applied to underwater image restoration of targets with single polarization characteristics, single objective function and optimization without limiting the polarization degree of target light lead to overexposure or too dark restored images, and the current underwater polarization imaging methods require prior knowledge and human-computer interaction, which limits the application scope of underwater imaging technology and cannot realize automatic restoration of underwater polarization images. The present invention proposes an underwater polarization imaging method based on target light polarization degree restriction conditions, which can realize automatic restoration of underwater images for targets with low polarization characteristics, high polarization targets and mixed targets with high and low polarization characteristics, effectively solves the problems of human-computer interaction and image background area limitation in traditional underwater polarization imaging methods, and is applicable to low turbidity and high turbidity underwater environments, promoting the application and development of underwater polarization imaging technology.

[0005] Technical solution: An underwater polarization imaging method based on the target light polarization degree restriction condition includes the following steps:

[0006] Step 1: Obtain collinear polarization image I ∥ and the cross-polarization image I ⊥ , the total light intensity of the underwater image is the collinear polarization image I ∥ With the cross-polarization image I ⊥ The sum of the collinear polarization images I∥ By the collinear target light intensity T ∥ and the collinear backscattered light intensity B ∥ Composition, cross polarization image I ⊥ By the cross target light intensity T ⊥ and the cross-backscattered light intensity B ⊥ Composition; According to the collinear target light intensity T ∥ With the cross target light intensity T ⊥ Calculate the target light polarization degree P obj , according to the collinear backscattered light intensity B ∥ and the cross-backscattered light intensity B ⊥ Calculate the polarization degree P of the backscattered light scat , thereby establishing the target light imaging model and the backscattered light imaging model;

[0007] Step 2: Based on the polarization degree range of the backscattered light and the polarization degree range of the target light, a target light polarization degree boundary model is established;

[0008] Step 3: The mutual information function is used to characterize the correlation between the target light imaging image and the backscattered light imaging image, and an underwater image contrast evaluation function of the target light imaging image is established. The boundary-constrained target light polarization degree and the cutoff frequency in the backscattered light imaging model are used as independent variables, and the mutual information function and the underwater image contrast evaluation function are used as objective functions. A dual-objective optimization algorithm is used to find the optimal solution set that satisfies the minimum mutual information and the maximum underwater image contrast evaluation function of the target light imaging image;

[0009] Step 4: Establish an image quality evaluation function, and select the optimal solution that can maximize the calculation result of the image quality evaluation function from the optimal solution set; bring the optimal solution into the target light imaging model to obtain the final imaging result.

[0010] Further, in step 1, the collinear target light intensity T ∥ With the cross target light intensity T ⊥ Calculate the target light polarization degree P obj , according to the collinear backscattered light intensity B ∥ and the cross-backscattered light intensity B ⊥ Calculate the polarization degree P of the backscattered light scat , in order to establish the target light imaging model and the backscattered light imaging model, specifically including:

[0011] According to the collinear target light intensity T ∥ With the cross target light intensity T ⊥ Calculate the target light polarization degree P obj , expressed as:

[0012]

[0013] According to the collinear backscattered light intensity B ∥ and the cross-backscattered light intensity B ⊥ Calculate the polarization degree P of the backscattered light scat , expressed as:

[0014]

[0015] The target light imaging model is expressed as:

[0016]

[0017] The backscattered light imaging model is expressed as:

[0018]

[0019] Among them, F -1 , F is the inverse Fourier transform and Fourier transform, D(u,v) represents the distance between two points in the spectrum, D 0 is the cut-off frequency.

[0020] Furthermore, based on the target information light intensity model, the polarization degree range of the backscattered light and the target light polarization degree range, a target light polarization degree boundary model is established, which is expressed as:

[0021]

[0022] In the formula, (·) min To find the minimum value.

[0023] Furthermore, the mutual information function is used to characterize the correlation between the target light imaging image and the backscattered light imaging image, which is expressed as:

[0024]

[0025] Where MI(T,B) is the mutual information, prob(t,b) is the joint probability distribution function, prob(t) is the first edge probability distribution function obtained according to the grayscale value corresponding to the target light imaging model; prob(b) is the second edge probability distribution function obtained according to the grayscale value corresponding to the backscattered light imaging model;

[0026] in:

[0027]

[0028]

[0029]

[0030] Wherein, t(i) represents the data obtained by dividing the fixed value pixel number of the target light imaging model by the total number of pixels, and b(i) represents the data obtained by dividing the fixed value pixel number of the backscattered light imaging model by the total number of pixels.

[0031] Furthermore, the underwater image contrast evaluation function of the target light imaging image is expressed as:

[0032]

[0033] In the formula, the target light image is divided into k 1 ,k 2 blocks with sequence number (k,l), Θ is the Kronecker operator, L max,k,l ,L min,k,l are the maximum and minimum values ​​in the block with sequence number (k, l) in the target light image.

[0034] Furthermore, the method uses the boundary-constrained target light polarization degree and the cutoff frequency in the backscattered light imaging model as independent variables, the mutual information function and the underwater image contrast evaluation function as objective functions, and uses a dual-objective optimization algorithm to find the optimal solution set that satisfies the minimum mutual information and the maximum underwater image contrast evaluation function of the target light imaging image, specifically including:

[0035] Taking the mutual information function and the underwater image contrast evaluation function as the objective function, a dual-objective genetic optimization algorithm is used to solve the problem, and the following is obtained:

[0036]

[0037] Among them, (P obj ,D 0 ) optimal is the optimal solution set; Dual-objective-Optimal dual-objective optimization algorithm.

[0038] Furthermore, the image quality evaluation function is expressed as:

[0039]

[0040] The underwater restored image is divided into N×M blocks in the (k, l) dimensions and marked with serial numbers ω. and They represent the maximum and minimum values ​​of the light intensity in the ωth block of the image respectively, and q is a constant used to correct the denominator to a positive value.

[0041] Furthermore, the above-mentioned selecting the optimal solution that can maximize the calculation result of the image quality evaluation function from the optimal solution set; bringing the optimal solution into the target light imaging model to obtain the final imaging result specifically includes:

[0042] From the optimal solution set (P obj ,D 0 ) optimal The optimal solution that can maximize the value of the image quality evaluation function calculation result is selected, which is expressed as:

[0043]

[0044] Among them, P obj optimal Denotes the optimal target light polarization degree, D 0 optimal represents the optimal cutoff frequency;

[0045] The optimal solution Bring it into the target light imaging model to get the final imaging result, namely:

[0046]

[0047] Among them, T opt Represents the final imaging result.

[0048] Beneficial effects: Compared with the prior art, the present invention collects colinear polarization images and cross-polarization images to obtain a target light imaging model and a backscattered light imaging model; establishes a target light polarization boundary model based on the global backscattered light polarization; takes the target light polarization with limited boundaries and the cutoff frequency in the backscattered light imaging model as independent variables, takes the minimum mutual information between the target light imaging image and the backscattered light imaging image and the maximum underwater image contrast function of the target light imaging image as evaluation indicators, adopts a dual-objective optimization algorithm, and establishes an optimal solution set of the target light polarization and the cutoff frequency; sequentially brings the optimal solution set into the target light imaging model to obtain an underwater restored image set, takes the image quality evaluation function as the evaluation indicator, traverses the underwater restored image set, and selects the image with the largest evaluation index value as the final underwater restored image; it has the following advantages:

[0049] (1) The method of the present invention introduces a target light polarization boundary model, which can realize automatic restoration of underwater images for targets with low polarization characteristics, high polarization characteristics, and mixed targets with high and low polarization characteristics. It effectively solves the problems of human-computer interaction and image background area limitation in traditional underwater polarization imaging methods, and is suitable for low turbidity and high turbidity underwater environments, promoting the application and development of underwater polarization imaging technology.

[0050] (2) The method of the present invention can realize the restoration of targets with complex polarization characteristics and targets in water bodies with different turbidity. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flow chart of the present invention;

[0052] Figure 2 Model diagram of the underwater active polarization imaging system built for the present invention;

[0053] Figure 3 This is a comparison chart of the effects; Figure 3 (a) and (c) are the total light intensity images underwater. Figure 3 (b) and (d) are the effect diagrams obtained by actually applying the method of the present invention. DETAILED DESCRIPTION

[0054] The technical solution of the present invention is now further described in conjunction with the accompanying drawings and embodiments.

[0055] Embodiment 1:

[0056] like Figure 1 As shown, this embodiment proposes an underwater polarization imaging method based on the target light polarization degree restriction condition, which mainly includes the following steps:

[0057] Step 1: Build Figure 2 An underwater active polarization imaging simulation system is shown, in which an LED red light source 8 emits a starting light beam, which passes through a first polarizer 7, with its polarization direction set to the horizontal direction, and then passes through a transparent water tank 3, and finally irradiates a non-uniform polarization characteristic target 5 and suspended particles 4. To simulate a turbid water environment, 5 ml and 10 ml of milk are added to the water respectively. The target 5 is made of a metal coin with a high polarization characteristic and a plastic Rubik's cube with a low polarization characteristic bonded together to form a target with complex polarization characteristics. When the starting light beam irradiates the target 5, the target 5 will reflect the target light, while the suspended particles 4 will scatter the backscattered light. After being processed by the second polarizer 2, these target lights and backscattered lights are irradiated onto the CCD detector 1 for recording and analysis.

[0058] The total light intensity received by CCD detector 1 has the following relationship with the target light and backscattered light:

[0059] I=T+B (1)

[0060] Among them, I is the total light intensity of the underwater image, T is the target light, and B is the backscattered light. I, T, and B can be decomposed in mutually orthogonal directions.

[0061] The polarization direction of the second polarizer 2 is rotated to 0° to obtain the colinear polarization image I ∥ Rotate the polarization direction of the second polarizer 2 to 90° to obtain a cross-polarization image I ⊥ , the total light intensity of the underwater image I and the collinear polarization image I ∥ and the cross-polarization image I ⊥ With the following relationship:

[0062] I=I ∥ +I ⊥ (2)

[0063] From the above two formulas, we can know that the collinear polarization image I ∥ , cross polarization image I ⊥ It can be composed of the respective target light and backscattered light, namely:

[0064] I ∥ =T ∥ +B ∥ (3)

[0065] I ⊥ =T ⊥ +B ⊥ (4)

[0066] Among them, T ∥ ,B ∥ are the collinear target light intensity and the collinear backscattered light intensity, T ⊥ ,B ⊥ are the cross target light intensity and the cross backscattered light intensity, respectively.

[0067] According to the collinear target light intensity T ∥ With the cross target light intensity T ⊥ Calculate the target light polarization degree P obj ,Right now:

[0068]

[0069] According to the collinear backscattered light intensity B ∥ and the cross-backscattered light intensity B ⊥ Calculate the polarization degree P of the backscattered light scat ,Right now:

[0070]

[0071] The target information light intensity model is established by formula (1) to formula (6), namely:

[0072]

[0073] The target light has high-frequency characteristics in the spectrum, while the backscattered light has low-frequency characteristics in the spectrum. The backscattered light imaging model B obtained by low-frequency filtering the total light intensity image can be expressed as:

[0074]

[0075] Among them, F -1 , F is the inverse Fourier transform and Fourier transform, D(u,v) represents the distance between two points in the spectrum, D 0is the cut-off frequency.

[0076] Step 2: Establish the target light polarization boundary model based on the global backscattered light polarization intensity. Specifically include:

[0077] According to the physical definition of polarization degree, the ratio of the intensity of partially polarized light to the intensity of the entire light, it can be known that the polarization degree of backscattered light and the polarization degree of target light range between 0 and 1, that is:

[0078] 0≤P scat ≤1 (9)

[0079] 0≤P obj ≤1 (10)

[0080] The target light polarization boundary model is established according to formulas (7), (9) and (10), namely:

[0081]

[0082] in,(·) min To find the minimum value.

[0083] Step 3: With the boundary-constrained target light polarization degree and the cutoff frequency in the backscattered light imaging model as independent variables, and the minimum mutual information between the target light imaging image and the backscattered light imaging image and the maximum underwater image contrast function of the target light imaging image as evaluation indicators, a dual-objective optimization algorithm is used to establish the optimal solution set of the target light polarization degree and the cutoff frequency. Specifically, it includes:

[0084] According to the gray value corresponding to the target intensity imaging model, the first edge probability distribution function is obtained, which is expressed as:

[0085]

[0086] Wherein, t(i) represents the data obtained by dividing the number of fixed-value pixels of the target light imaging model by the total number of pixels.

[0087] The second edge probability distribution function is obtained according to the gray value corresponding to the backscattered light imaging model, which is expressed as:

[0088]

[0089] Wherein, b(i) represents the data obtained by dividing the fixed value pixel number of the backscattered light imaging model by the total pixel number.

[0090] According to the grayscale value corresponding to the target light imaging model and the grayscale value corresponding to the backscattered light imaging model, the joint probability distribution function prob(t,b) is obtained, which is expressed as:

[0091]

[0092] The mutual information is obtained by combining the probability distribution function and the marginal distribution function, which is expressed as:

[0093]

[0094] Among them, MI(T,B) is the mutual information, which is used to characterize the correlation between the target light imaging image and the backscattered light imaging image, and is greater than 0. When the mutual information is smaller, it means that the target light and the backscattered light are more thoroughly separated, and the imaging effect is better.

[0095] Calculate the underwater image contrast evaluation function of the target light imaging image, that is:

[0096]

[0097] In the formula, the output target light image is divided into k 1 ,k 2 There are blocks with serial numbers (k, l), where k 1 ,k 2 are both 10, and Θ is the Kronecker operator, and L max,k,l ,L min,k,l are the maximum and minimum values ​​in the block with sequence number (k, l) in the image.

[0098] Taking mutual information and underwater image contrast evaluation function as objective functions, a dual-objective optimization algorithm is used to find the optimal solution set that satisfies the small mutual information and the maximum underwater image contrast evaluation function, namely:

[0099]

[0100] Among them (P obj ,D 0 ) optimal is the optimal solution set; Dual-objective-Optimal dual-objective optimization algorithm, MI(T,B) represents the mutual information between the target light imaging image and the backscattered light imaging image; UIconm(T) represents the underwater image contrast function of the target light imaging image.

[0101] In this embodiment, a dual-objective genetic optimization algorithm is used, and the optimal front-end individual coefficient is set to 0.2, the population size is 200, the maximum genetic generation is 300, the stop generation is 300, and the fitness function deviation is 0.01, and finally 30 sets of optimal solutions are given. The optimal solution set refers to the value that satisfies: if the value reduces the mutual information, the contrast will be reduced, or if the value increases the contrast, the mutual information will be increased.

[0102] Step 4: Establish an image quality evaluation function, use the image quality evaluation function as the evaluation index, select the optimal solution that can maximize the calculation result of the image quality evaluation function from the optimal solution set; bring the optimal solution into the target light imaging model to obtain the final imaging result. The image quality evaluation function is calculated as follows:

[0103]

[0104] The image is divided into N×M blocks in the (k, l) dimensions and marked with serial numbers ω, where N and M are 10. and They represent the maximum and minimum values ​​of the light intensity in the ωth block of the image, respectively. q is a very small constant used to correct the denominator to zero, which is 0.1×10 -4 , the larger the function value, the clearer the image details and the higher the image quality.

[0105] Select the value that can maximize the result of the image quality evaluation function calculation of the target light imaging image from the optimal solution set, that is:

[0106]

[0107] Among them, EME(T) represents the image detail quality evaluation function of the target light imaging image. (P obj optimal ,D 0 optimal ) represents the maximum satisfactory value of the image detail quality evaluation function of the target optical imaging image in the optimal solution set. For low turbidity water bodies, it is (0.35, 2.4), and for high turbidity water bodies, it is (0.29, 2.26).

[0108] The optimal target light polarization degree is obtained and cutoff frequency Substitute the target light intensity model to obtain the final imaging result, namely:

[0109]

[0110]

[0111] Among them, T opt Represents the desired final imaging result.

[0112] To verify the effectiveness of this embodiment, experiments with high and low concentrations were performed, such as Figure 3 (a) and Figure 3 (c) in the figure are the total light intensity images of high and low concentration turbid water, respectively. Subjectively, the images are 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 (b) and Figure 3 (d) in the figure, and is not restricted by the background area and human-computer interaction. It can realize automatic restoration of underwater images of targets with complex polarization characteristics and is suitable for underwater environments with different turbidity.

[0113] Underwater image contrast (UIconM), image average gradient (AG), and information entropy (Entorpy) 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 the following table:

[0114] Table 1 Quality evaluation of restored images of complex polarization targets in low turbidity water

[0115] contrast UIcon AG Entorpy Total light intensity diagram 0.11 0.0026 3.97 This embodiment 0.32 0.1318 6.44

[0116] Table 2 Quality evaluation of restored images of complex polarization targets in highly turbid water

[0117] contrast UIcon AG Entorpy Total light intensity diagram 0.10 0.0025 2.97 This embodiment 0.28 0.0879 4.76

[0118] It can be seen from the table that, compared with the original total light intensity image, the restored image of this embodiment has significantly improved various objective evaluation indicators, which verifies the effectiveness of the present invention in restoring complex polarization characteristics.

[0119] 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.

[0120] 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. An underwater polarization imaging method based on target light polarization degree restriction condition, characterized by: The following steps are involved: Step 1: Obtain collinear polarization image I ∥ and cross-polarized images The total light intensity of the underwater image is the collinear polarization image I ∥ With cross polarization image The collinear polarization image I ∥ By the collinear target light intensity T ∥ and the collinear backscattered light intensity B ∥ Composition, cross-polarized image By cross target light intensity and the cross backscatter intensity Composition; According to the collinear target light intensity T ∥ With cross target light intensity Calculate the target light polarization degree P obj , according to the collinear backscattered light intensity B ∥ Cross-backscatter intensity Calculate the polarization degree P of the backscattered light scat , in order to establish the target light imaging model and the backscattered light imaging model; according to the collinear polarization image I ∥ , cross polarization image Target light polarization degree P obj and the backscattered light polarization P scat Establishing target information light intensity model; Step 2: Based on the polarization degree range of the backscattered light and the polarization degree range of the target light, a target light polarization degree boundary model is established; Step 3: The mutual information function is used to characterize the correlation between the target light imaging image and the backscattered light imaging image, and an underwater image contrast evaluation function of the target light imaging image is established. The boundary-constrained target light polarization degree and the cutoff frequency in the backscattered light imaging model are used as independent variables, and the mutual information function and the underwater image contrast evaluation function are used as objective functions. A dual-objective optimization algorithm is used to find the optimal solution set that satisfies the minimum mutual information and the maximum underwater image contrast evaluation function of the target light imaging image; Step 4: Establish an image quality evaluation function, and select the optimal solution that can maximize the calculation result of the image quality evaluation function from the optimal solution set; The optimal solution is introduced into the target light imaging model to obtain the final imaging result; Among them, based on the target information light intensity model, the polarization degree range of the backscattered light and the target light polarization degree range, the target light polarization degree boundary model is established, which is expressed as: Where P obj is the target light polarization degree, I is the total light intensity of the underwater image, B is the backscattered light imaging model, (·) min To find the minimum operation; Wherein, the acquisition of the collinear polarization image I ∥ and cross-polarized images Specific operations include: An underwater active polarization imaging simulation system is built, which includes: an LED red light source, a first polarizer, a transparent water tank, a target, suspended particles, a second polarizer and a CCD detector; the LED red light source emits a starting light beam, the starting light beam passes through the first polarizer, the polarization direction of the first polarizer is set to the horizontal direction, then passes through the transparent water tank, and finally irradiates the target and the suspended particles; the target is formed by bonding a first target with a high polarization characteristic and a second target with a low polarization characteristic together to form a target with complex polarization characteristics; when the starting light beam irradiates the target, the target reflects the target light, the suspended particles scatter the backscattered light, and the target light and the backscattered light are processed by the second polarizer and then irradiated to the CCD detector for recording and analysis; The total light intensity received by the CCD detector has the following relationship with the target light and backscattered light: I=T+B Among them, I is the total light intensity of the underwater image, T is the target light; B is the backscattered light, I, T, B can be decomposed in mutually orthogonal directions; Rotate the polarization direction of the second polarizer to 0° and obtain the colinear polarization image I ∥ ; Rotate the polarization direction of the second polarizer to 90° to obtain a cross-polarization image 2. The underwater polarization imaging method based on the target light polarization degree restriction condition according to claim 1, characterized in that: In step 1, the collinear target light intensity T ∥ With cross target light intensity Calculate the target light polarization degree P obj , according to the collinear backscattered light intensity B ∥ Cross-backscatter intensity Calculate the polarization degree P of the backscattered light scat , in order to establish the target light imaging model and the backscattered light imaging model, specifically including: According to the collinear target light intensity T ∥ With cross target light intensity Calculate the target light polarization degree P obj , expressed as: According to the collinear backscattered light intensity B ∥ Cross-backscatter intensity Calculate the polarization degree P of the backscattered light scat , expressed as: The target light imaging model is expressed as: The backscattered light imaging model is expressed as: Among them, F -1 , F is the inverse Fourier transform and Fourier transform, D(u,v) represents the distance between two points in the spectrum, and D0 is the cutoff frequency.

3. The underwater polarization imaging method based on the target light polarization degree restriction condition according to claim 2, characterized in that: The method uses the boundary-constrained target light polarization degree and the cutoff frequency in the backscattered light imaging model as independent variables, the mutual information function and the underwater image contrast evaluation function as objective functions, and uses a dual-objective optimization algorithm to find the optimal solution set that satisfies the minimum mutual information and the maximum underwater image contrast evaluation function of the target light imaging image, specifically including: Taking the mutual information function and the underwater image contrast evaluation function as the objective function, a dual-objective genetic optimization algorithm is used to solve the problem, and the following is obtained: Among them, (P obj ,D0) optimal is the optimal solution set; Dual-objective-Optimal dual-objective optimization algorithm; MI(T,B) represents the mutual information between target light imaging and backscattered light imaging; UIconm(T) represents the underwater image contrast function of the target light imaging image.

4. The underwater polarization imaging method based on the target light polarization degree restriction condition according to claim 3, characterized in that: The MI(T,B) is expressed as: Where MI(T,B) is the mutual information, prob(t,b) is the joint probability distribution function, prob(t) is the first edge probability distribution function obtained according to the grayscale value corresponding to the target light imaging model; prob(b) is the second edge probability distribution function obtained according to the grayscale value corresponding to the backscattered light imaging model; in: Wherein, t(i) represents the data obtained by dividing the number of fixed-value pixels i of the target light imaging model by the total number of pixels, and b(i) represents the data obtained by dividing the number of fixed-value pixels i of the backscattered light imaging model by the total number of pixels.

5. The underwater polarization imaging method based on the target light polarization degree restriction condition according to claim 4, characterized in that: The UIconm(T) is represented as: In the formula, the target light image is divided into k1×k2 blocks with serial numbers (k, l) in both horizontal and vertical dimensions. Θ is the Kronecker operator, T max,k,l , T min,k,l are the maximum and minimum values ​​in the block with sequence number (k, l) in the target light image.

6. The underwater polarization imaging method based on the target light polarization degree restriction condition according to claim 5, characterized in that: The image quality evaluation function is expressed as: The underwater restored image is divided into N×M blocks numbered (k, l) in both horizontal and vertical dimensions and marked with the serial number ω. and They represent the maximum and minimum values ​​of the light intensity in the ωth block of the image respectively, and q is a constant used to correct the denominator to a positive value.

7. The underwater polarization imaging method based on the target light polarization degree restriction condition according to claim 6, characterized in that: The step of selecting the optimal solution that can maximize the calculation result of the image quality evaluation function from the optimal solution set; The optimal solution is introduced into the target light imaging model to obtain the final imaging result, which includes: From the optimal solution set (P obj ,D0) optimal The optimal solution that can maximize the value of the image quality evaluation function calculation result is selected, which is expressed as: Among them, P obj optimal Denotes the optimal target light polarization degree, D0 optimal represents the optimal cutoff frequency; The optimal solution Bring it into the target light imaging model to get the final imaging result, namely: Among them, T opt Represents the final imaging result.

Images