Image enhancement method and system based on adaptive filtering and polarization weighted fusion

Abstract

The present disclosure relates to an image enhancement method and system based on adaptive filtering and polarization weighting fusion, comprising the following steps: S10, obtaining an original polarization image of a polarization camera; S20, calculating maximum and minimum light intensity images and a linear polarization degree image according to the original polarization image using Stokes vectors; S30, performing low-pass filtering processing on the maximum and minimum light intensity images through an adaptive threshold low-pass filtering algorithm to obtain filtered images; S40, weighting and fusing the filtered images and the linear polarization degree image to obtain an image with clear texture. Through the adaptive threshold low-pass filtering algorithm, the present disclosure can restore a clear underwater scene without background prior information, and can dynamically set a threshold according to image content, thereby reducing noise and unnecessary details while retaining important structure and texture information of the image.

Description

Technical Field

[0001] The present disclosure relates to the field of computer vision, and in particular to an image enhancement method and system based on adaptive filtering and polarization weighted fusion. Background Art

[0002] In recent years, underwater optical imaging technology has been widely used in fields such as marine resource exploration, seabed topography mapping, and underwater ecological environment monitoring. However, due to the absorption and scattering of light as it propagates through water, the quality of underwater imaging often degrades significantly. These factors combine to cause a sharp drop in underwater image contrast and blur the details of the target. Therefore, how to effectively suppress scattering interference and extract key information about underwater targets in complex turbid water environments has become an important research topic. Underwater image enhancement methods based on adaptive threshold filtering and weighted fusion can automatically adjust the filter parameters based on local changes in the image through adaptive threshold filtering, thereby accurately removing noise and preserving image details. Weighted fusion technology, by combining the advantages of multiple enhancement methods, can better balance the relationship between enhancement effect and image quality, avoid over-enhancement or over-denoising, and obtain optimal underwater images.

[0003] Traditional underwater image enhancement algorithms rely primarily on various image processing techniques to improve image quality by optimizing and enhancing underwater image features. These methods typically include: histogram equalization-based image enhancement, which enhances image contrast by adjusting the image's grayscale distribution; Retinex-based image enhancement, which improves image color and contrast by estimating illumination and reflectance components; and filtering-based denoising methods, which smooth images to remove noise while preserving edges and details. However, these traditional methods often struggle to effectively address color distortion, blur, and low contrast in complex underwater environments, resulting in poor enhancement and limited detail recovery. Therefore, a method for underwater image enhancement based on adaptive threshold filtering is needed. This method automatically sets adaptive thresholds based on different image regions, making the filtering process more flexible and precise. Furthermore, it is necessary to enhance local details while improving the clarity, contrast, and information retention of underwater images, resulting in a more natural and detailed enhanced image. Summary of the Invention

[0004] To address the challenges of traditional robotic exploration methods in the face of dynamic changes, complex obstacles, and environments that are not known in advance, such as repeated exploration, incomplete environmental coverage, and susceptibility to falling into local optimality, this paper proposes an image enhancement method based on adaptive filtering and polarization-weighted fusion to address these issues.

[0005] According to one aspect of the present disclosure, an image enhancement method based on adaptive filtering and polarization weighted fusion is provided, comprising:

[0006] S10, obtaining an original polarization image from a polarization camera;

[0007] S20, obtaining maximum and minimum light intensity images and a linear polarization degree image based on the original polarization image using Stokes vector calculation;

[0008] S30, performing low-pass filtering on the maximum and minimum light intensity images using an adaptive threshold low-pass filtering algorithm to obtain filtered images;

[0009] S40 , performing weighted fusion on the filtered image and the linear polarization degree image to obtain an image with clear texture.

[0010] Preferably, the maximum and minimum light intensity images and the linear polarization degree image are obtained by Stokes vector calculation, which are expressed as:

[0011] ,

[0012] ,

[0013] in,

[0014] ,

[0015] Where, represents the position of the pixel, is the linear polarization image, 、 、 and Respectively represent the light intensity after polarization filtering at 0°, 90°, 45° and 135°, is the angle between the horizontal direction and the transmission axis of the polarizer. Angle, calculate the maximum and minimum light intensity images.

[0016] Preferably, the maximum and minimum light intensity images are subjected to low-pass filtering by an adaptive threshold low-pass filtering algorithm, including: optimizing the low-pass filtering according to the adaptive threshold, when When , the intensity of backscattered light is maintained after low-pass filtering; when When , the intensity of the original light intensity image remains unchanged, which can be expressed as:

[0017] ,

[0018] Where, represents an image containing only the backscattered light intensity distribution, is the adaptive threshold, To optimize the backscattered light intensity, represents the position of the pixel, is the original image.

[0019] Preferably, optimizing the low-pass filtering according to the adaptive threshold comprises: calculating the adaptive threshold using a normalized gradient, wherein the adaptive threshold is calculated as follows:

[0020] ,

[0021] Where, is the normalized gradient, is the size of the image, Represents the input image.

[0022] Preferably, optimizing the low-pass filtering according to the adaptive threshold further includes: optimizing the low-contrast area using an adaptive threshold low-pass filtering algorithm, and normalizing the gradient values ​​of different image areas, which is expressed as:

[0023] ,

[0024] in,

[0025] ,

[0026] ,

[0027] ,

[0028] Where, is the total gradient, Represents the maximum value of all gradient values ​​in the image, Represents the input image and * represents the convolution operation.

[0029] Preferably, weighted fusion of the filtered image and the linear polarization degree image includes: calculating an optimal weight of the filtered image and the linear polarization degree image based on a structural similarity index, visual information fidelity, and contrast weight, expressed as:

[0030] ,

[0031] ,

[0032] ,

[0033] ,

[0034] ,

[0035] Where, represents the image after filtering, represents the polarization degree image, represents the structural similarity index, represents the fidelity of visual information, is the contrast weight of the image, represents the adjustment factor, 、 Represents images respectively x and images y The mean of 、 Represents images respectively x and images y The variance of Representing an image x and images y The covariance of and is the stability constant, Representing an image x and images y The mutual information between Representing an image x Self-information, Indicates the image i The brightness value of each pixel, represents the average brightness value of the image, N Indicates the total number of pixels.

[0036] Preferably, performing weighted fusion of the filtered image and the linear polarization degree image includes: performing weighted fusion of the filtered image and the linear polarization degree image on an optimal weight value, which is expressed as:

[0037] ,

[0038] Where, Represents the weight of the image after filtering, represents the image after filtering, represents the weight of the polarization image, Represented as a polarization degree image.

[0039] According to one aspect of the present disclosure, an image enhancement system based on adaptive filtering and polarization weighted fusion is provided, comprising:

[0040] The original polarization image acquisition module acquires the original polarization image of the polarization camera;

[0041] The light intensity image and linear polarization image acquisition module uses Stokes vector calculation to obtain the maximum and minimum light intensity images and linear polarization images based on the original polarization image;

[0042] A low-pass filtering processing module performs low-pass filtering on the maximum and minimum light intensity images using an adaptive threshold low-pass filtering algorithm to obtain a filtered image;

[0043] The image weighted fusion module performs weighted fusion on the filtered image and the linear polarization image to obtain an image with clear texture.

[0044] According to one aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to: execute the above-mentioned image enhancement method based on adaptive filtering and polarization weighted fusion.

[0045] According to one aspect of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the image enhancement method based on adaptive filtering and polarization weighted fusion is implemented.

[0046] Compared with the prior art, the beneficial effects of the present disclosure are:

[0047] 1) The method disclosed in this paper uses Stokes vectors to generate optimal image pairs, which can overcome the limitations of traditional methods that rely on manual rotation of polarizers and reduce image acquisition time.

[0048] 2) The present invention uses an adaptive threshold low-pass filtering algorithm to restore clear underwater scenes without background prior information and can reduce noise and unnecessary details while retaining important image structure and texture information by dynamically setting the threshold according to the image content.

[0049] 3) This disclosure can significantly improve the detail clarity and global contrast of the fused underwater image.

[0050] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

[0051] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and, together with the specification, are used to explain the technical solutions of the present disclosure.

[0053] Figure 1 The flowchart of the image enhancement method based on adaptive filtering and polarization weighted fusion is shown;

[0054] Figure 2 Shows a block diagram of the structure of polarization weighted fusion in an example of the present disclosure;

[0055] Figure 3 A structure block diagram of an image enhancement system based on adaptive filtering and polarization weighted fusion in the embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0056] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numbers in the drawings represent functionally the same or similar elements. Although various aspects of the embodiments are illustrated in the drawings, the drawings are not necessarily drawn to scale unless specifically indicated.

[0057] The word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any implementation described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other implementations.

[0058] The term "and / or" herein is merely used to describe an associated relationship, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. In addition, the term "at least one" herein means any one of a plurality or any combination of at least two of a plurality, for example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0059] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following detailed description. Those skilled in the art should understand that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, elements and circuits well known to those skilled in the art are not described in detail in order to highlight the main idea of the present disclosure.

[0060] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0061] Embodiment 1

[0062] Based on the above idea, the present application proposes an image enhancement method based on adaptive filtering and polarization weighted fusion. Figure 1 A flowchart of an image enhancement method based on adaptive filtering and polarization weighted fusion is shown. The method comprises:

[0063] S10, obtaining an original polarization image of a polarization camera;

[0064] S20, obtaining maximum and minimum light intensity images and a linear polarization degree image based on the original polarization image using Stokes vector calculation;

[0065] S30, performing low-pass filtering on the maximum and minimum light intensity images using an adaptive threshold low-pass filtering algorithm to obtain filtered images;

[0066] S40 , performing weighted fusion on the filtered image and the linear polarization degree image to obtain an image with clear texture.

[0067] The present disclosure provides an image enhancement method based on adaptive filtering and polarization weighted fusion, which specifically includes the following steps:

[0068] S10. Obtain an original polarization image from a polarization camera.

[0069] In this embodiment, the raw polarization image is acquired using an area array polarization camera that utilizes the principle of focal plane segmentation polarization imaging. The raw polarization image contains four images with different polarization angles: 0°, 45°, 90°, and 135°. These polarization images are captured using a camera equipped with a focal plane segmentation area array sensor (such as the IMX250MYR model). The polarization sensor in a polarization camera utilizes a polarizer layer over the photodiode of a conventional CMOS image sensor. Four polarizers at different angles (90°, 45°, 135°, and 0°) are sequentially placed over individual pixels, with each four pixels forming a calculation unit. By analyzing the relationship between these polarizers, the degree and direction of polarization can be calculated. Directly processing the images acquired by the polarization camera avoids the tedious steps of segmentation, reassembly, and recalculation, thereby simplifying the system complexity and improving overall processing speed.

[0070] S20. According to the original polarization image, the maximum and minimum light intensity images and the linear polarization degree image are obtained by using Stokes vector calculation.

[0071] In this embodiment, the maximum and minimum light intensity images and the linear polarization degree image are obtained by Stokes vector calculation, which are expressed as:

[0072] ,

[0073] ,

[0074] in,

[0075] ,

[0076] Where, represents the position of the pixel, is the linear polarization image, , , and represent the intensity of light filtered by polarizer at 0°, 90°, 45° and 135° respectively, is the angle between the horizontal direction and the transmission axis of the polarizer, by setting a specific angle, the maximum and minimum intensity images are calculated.

[0077] S30, the maximum and minimum intensity images are low-pass filtered by an adaptive threshold low-pass filtering algorithm to obtain a filtered image.

[0078] In this embodiment, considering that the backscattered light mainly affects the low frequency part of the image, the background area without target is not selected for calculation, but the entire image is low-pass filtered in the frequency domain to accurately estimate the entire intensity B. The process is first to transform the intensity image into the frequency domain by Fourier function, then multiply it with the transfer function of the filter, and the result of multiplication is returned to the spatial image by inverse Fourier transform.

[0079] ,

[0080] In the formula, and are the abbreviations of Fourier transform and inverse Fourier transform respectively, represents an image containing only backscattered light intensity distribution, is a low-pass filter, represents the coordinates in the frequency domain.

[0081] However, this method has a certain impact on the low contrast area of the target, that is, there will be a certain error in the backscattered light intensity obtained by low-pass filtering in these low contrast areas. In order to further improve the estimation accuracy of the polarization degree of backscattered light, the adaptive threshold low-pass filtering algorithm is used in this example to optimize the processing of low contrast area, which can effectively reduce the error that may be generated in the low-pass filtering process.

[0082] Among them, the maximum and minimum intensity images are low-pass filtered by an adaptive threshold low-pass filtering algorithm, including: according to the adaptive threshold, the low-pass filtering is optimized, when , the intensity of the backscattered light is kept after low-pass filtering; when , the intensity of the original light intensity image remains unchanged, represented as:

[0083] ,

[0084] In the formula, represents an image containing only backscattered light intensity distribution, is an adaptive threshold, To optimize the backscattered light intensity, represents the position of the pixel, is the original image.

[0085] Optimizing the low-pass filter according to the adaptive threshold includes: calculating the adaptive threshold using the normalized gradient, and the calculation formula of the adaptive threshold is:

[0086] ,

[0087] Where, is the normalized gradient, is the size of the image, Represents the input image.

[0088] Furthermore, optimizing the low-pass filtering according to the adaptive threshold value further includes: optimizing the low-contrast area using the adaptive threshold low-pass filtering algorithm, and normalizing the gradient values ​​of different image areas, which is expressed as:

[0089] ,

[0090] in,

[0091] ,

[0092] ,

[0093] ,

[0094] Where, is the total gradient, Represents the maximum value of all gradient values ​​in the image, Represents the input image and * represents the convolution operation.

[0095] S40 , performing weighted fusion on the filtered image and the linear polarization degree image to obtain an image with clear texture.

[0096] In this embodiment, the structural block diagram of polarization weighted fusion is as follows: Figure 2 As shown, Figure 2 The working principle and process of the image fusion module in the underwater image enhancement method based on adaptive threshold filtering and polarization weighted fusion are demonstrated. Figure 2 The following steps are described:

[0097] Input image: First, input two images, which are the images processed by adaptive low-pass filtering (Target Image) and Degree of Linear Polarization Image (DoLP Image) .

[0098] Image Quality Evaluation Unit (IQEU): The input two images are evaluated by the image quality evaluation unit. This unit consists of three parts: Structural Similarity Index (SSIM) for evaluating the structural similarity of two images; Visual Information Fidelity (VIF) for evaluating the visual information fidelity of the image; Contrast Weight (CW) for evaluating the contrast of the image.

[0099] Weight Calculation: According to the results of the image quality evaluation unit, the weights of the two images are calculated.

[0100] Weighted Fusion: The calculated weights are applied to the input images for weighted fusion to generate the final fusion image.

[0101] Output Image: The final output fusion image has clear texture details and clear image quality.

[0102] Through this polarization weighted fusion method, the advantages of the image after adaptive low-pass filtering processing and the polarization degree image can be fully utilized to generate high-quality underwater images. The adaptive low-pass filtering processed image and the polarization degree image are weighted fused to fully utilize the advantages of both to achieve better image reconstruction.

[0103] In the process of weighted fusion, the image quality evaluation unit (IQEU) composed of structural similarity index (SSIM), visual information fidelity (VIF) and contrast weight (CW) is used to reasonably retain the advantage information of different source images, and the optimal weight value of the filtered image and the linear polarization degree image is calculated, represented as:

[0104] ,

[0105] ,

[0106] ,

[0107] ,

[0108] ,

[0109] In the formula, represents the filtered image, represents the polarization degree image, represents the structural similarity index, represents the visual information fidelity, is the contrast weight of the image, represents the adjustment factor, , respectively represent the mean value of image x and image y , , denote the variance of the image x and the image y , denote the covariance of the image x and the image y , and are stable constants, denote the mutual information between the image x and the image y , denote the self-information of the image x , denote the brightness value of the i-th pixel in the image i , denote the average brightness value of the image N denote the total number of pixels. By optimizing the weight of the image after adaptive low-pass filtering processing and the polarization degree image, the optimal weight map is obtained, and the final fused image is obtained by using the optimal weight map.

[0110] Further, the weighted fusion of the filtered image and the linear polarization degree image comprises: weighted fusion of the optimal weight of the filtered image and the linear polarization degree image, denoted as:

[0111] ,

[0112] In the formula, denote the weight of the filtered image, denote the filtered image, denote the weight of the polarization degree image, denote the polarization degree image.

[0113] The adaptive threshold low-pass filtering algorithm is used to perform low-pass filtering processing on the light intensity image to achieve accurate suppression of backscattering light, and the clarity of the underwater scene image can be effectively restored without prior knowledge of the background. The adaptive low-pass filtering processed image and the polarization degree image are fused by the weighting fusion algorithm, thereby generating a high-quality underwater image with high clarity and obvious texture details.

[0114] Embodiment 2

[0115] As another aspect of the embodiments of the present disclosure, an image enhancement system 100 based on adaptive filtering and polarization weighted fusion is also provided, as shown in Figure 3 , comprising:

[0116] An original polarization image acquisition module 1 acquires an original polarization image of a polarization camera;

[0117] Light intensity image and linear polarization image acquisition module 2, which uses Stokes vector calculation to obtain maximum and minimum light intensity images and linear polarization image based on the original polarization image;

[0118] A low-pass filtering processing module 3 performs low-pass filtering on the maximum and minimum light intensity images using an adaptive threshold low-pass filtering algorithm to obtain a filtered image;

[0119] The image weighted fusion module 4 performs weighted fusion on the filtered image and the linear polarization degree image to obtain an image with clear texture.

[0120] In the absence of any contradiction, the above modules in the system of the embodiment of the present disclosure can implement any implementation of the above method.

[0121] Based on the description of the above embodiments, it can be seen that the embodiments of the present disclosure can achieve the following technical effects:

[0122] 1) The method disclosed in this paper uses Stokes vectors to generate optimal image pairs, which can overcome the limitations of traditional methods that rely on manual rotation of polarizers and reduce image acquisition time.

[0123] 2) The present invention uses an adaptive threshold low-pass filtering algorithm to restore clear underwater scenes without background prior information and can reduce noise and unnecessary details while retaining important image structure and texture information by dynamically setting the threshold according to the image content.

[0124] 3) This disclosure can significantly improve the detail clarity and global contrast of the fused underwater image.

[0125] The present disclosure also provides an electronic device comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to implement the aforementioned image enhancement method based on adaptive filtering and polarization-weighted fusion. The electronic device can be provided as a terminal, server, or other device.

[0126] The present disclosure also provides a computer-readable storage medium having computer program instructions stored thereon. When executed by a processor, the computer program instructions implement the aforementioned image enhancement method based on adaptive filtering and polarization-weighted fusion. The computer-readable storage medium may be a non-volatile computer-readable storage medium.

[0127] Those skilled in the art will understand that in the above-mentioned image enhancement method and system based on adaptive filtering and polarization weighted fusion in the specific implementation manner, the writing order of each step does not mean a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0128] The computer program product of the present disclosure can be a computer program product, which is a machine-readable medium (media) having instances of the software embodied thereon, such as computer software, firmware, wireless application protocol (WAP), middleware or microcode. For example, a computer program product can be a floppy disk, a CD-ROM, a DVD, a Blu-ray Disc™, a flash drive, a memory stick, a magnetic tape, or a hard disk drive. The computer program product can also be a downloaded file, such as a file downloaded from the Internet or another computer network. Of course, many modifications can be made by those skilled in the art to the inventive embodiments and applications described herein without departing from the scope and spirit of the described embodiments. Any reference to claimed subject matter shall include any such modifications and variations in the scope of the disclosure.

[0129] The foregoing description of the exemplary embodiments has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the embodiments to the precise form disclosed. Many modifications and variations are possible in light of the above teaching. It is intended that the scope of the embodiments be limited not with this detailed description, but rather by the claims appended hereto.

Claims

1. An image enhancement method based on adaptive filtering and polarization weighted fusion, characterized in that: The steps include: S10, obtaining an original polarization image from a polarization camera; S20, obtaining maximum and minimum light intensity images and a linear polarization degree image based on the original polarization image using Stokes vector calculation; S30, performing low-pass filtering on the maximum and minimum light intensity images using an adaptive threshold low-pass filtering algorithm to obtain a filtered image; and calculating an adaptive threshold using a normalized gradient. The calculation formula for the adaptive threshold is: ， Where, is the normalized gradient, is the size of the image, represents the input image; S40, performing weighted fusion on the filtered image and the linear polarization degree image to obtain an image with clear texture; The optimal weights of the filtered image and the linear polarization image are calculated based on the structural similarity index, visual information fidelity, and contrast weight, which can be expressed as: ， ， ， ， ， Where, represents the structural similarity index, represents the fidelity of visual information, is the contrast weight of the image, represents the adjustment factor, 、 Represents images respectively x and images y The mean of 、 Represents images respectively x and images y The variance of Representing an image x and images y The covariance of and is the stability constant, Representing an image x and images y The mutual information between Representing an image x Self-information, Indicates the image i The brightness value of each pixel, represents the average brightness value of the image, N Indicates the total number of pixels; The optimal weighted fusion of the filtered image and the linear polarization image is expressed as: ， Where, Represents the weight of the image after filtering, represents the image after filtering, represents the weight of the polarization image, Indicates the light intensity value of the polarization image.

2. The method according to claim 1, characterized in that The maximum and minimum light intensity images and linear polarization degree images are obtained by Stokes vector calculation, which are expressed as: ， ， in, ， Where, represents the position of the pixel, is the linear polarization image, 、 、 and Respectively represent the light intensity after polarization filtering at 0°, 90°, 45° and 135°, is the angle between the horizontal direction and the transmission axis of the polarizer. Angle, calculate the maximum and minimum light intensity images.

3. The method according to claim 1, characterized in that The maximum and minimum light intensity images are subjected to low-pass filtering by an adaptive threshold low-pass filtering algorithm, including: optimizing the low-pass filtering according to the adaptive threshold, when When , the intensity of backscattered light is maintained after low-pass filtering; when When , the intensity of the original light intensity image remains unchanged, which can be expressed as: ， Where, represents an image containing only the backscattered light intensity distribution, is the adaptive threshold, To optimize the backscattered light intensity, represents the position of the pixel, is the original image.

4. The method according to claim 3, characterized in that Optimizing the low-pass filter according to the adaptive threshold also includes: optimizing the low-contrast area using the adaptive threshold low-pass filter algorithm, and normalizing the gradient values ​​of different image areas, which is expressed as: ， in, ， ， ， Where, is the total gradient, Represents the maximum value of all gradient values ​​in the image, Represents the input image and * represents the convolution operation.

5. Image enhancement system based on adaptive filtering and polarization weighted fusion, characterized in that: include: The original polarization image acquisition module acquires the original polarization image of the polarization camera; The light intensity image and linear polarization image acquisition module uses Stokes vector calculation to obtain the maximum and minimum light intensity images and linear polarization images based on the original polarization image; The low-pass filtering processing module performs low-pass filtering on the maximum and minimum light intensity images using an adaptive threshold low-pass filtering algorithm to obtain a filtered image; and calculates the adaptive threshold using the normalized gradient. The calculation formula of the adaptive threshold is: ， Where, is the normalized gradient, is the size of the image, represents the input image; The image weighted fusion module performs weighted fusion on the filtered image and the linear polarization image to obtain an image with clear texture; The optimal weights of the filtered image and the linear polarization image are calculated based on the structural similarity index, visual information fidelity, and contrast weight, which can be expressed as: ， ， ， ， ， Where, represents the structural similarity index, represents the fidelity of visual information, is the contrast weight of the image, represents the adjustment factor, 、 Represents images respectively x and images y The mean of 、 Represents images respectively x and images y The variance of Representing an image x and images y The covariance of and is the stability constant, Representing an image x and images y The mutual information between Representing an image x Self-information, Indicates the image i The brightness value of each pixel, represents the average brightness value of the image, N Indicates the total number of pixels; The optimal weighted fusion of the filtered image and the linear polarization image is expressed as: ， Where, Represents the weight of the image after filtering, represents the image after filtering, represents the weight of the polarization image, Indicates the light intensity value of the polarization image.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the image enhancement method based on adaptive filtering and polarization weighted fusion according to any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the image enhancement method based on adaptive filtering and polarization weighted fusion according to any one of claims 1 to 4 is implemented.

Images