A method for removing fog and noise from a polarized image

Abstract

The application provides a defogging and denoising method applied to a polarized image, and the method comprises the following steps: obtaining an original image to be defogged; obtaining a high-noise high-defogging image by using a first amplification coefficient; obtaining a low-noise low-defogging image by using a second amplification coefficient, wherein the first amplification coefficient is greater than the second amplification coefficient; obtaining a noise image according to the high-noise high-defogging image and the low-noise low-defogging image; calculating a transmission coefficient by using a first local atmospheric light intensity, and performing normalization to obtain a normalized transmission coefficient; obtaining a distance-processed noise image according to the normalized transmission coefficient and the noise image; and obtaining a low-noise high-defogging image according to the high-noise high-defogging image and the distance-processed noise image. The defogging and denoising method provided by the application can effectively remove the noise in the image without using filtering under the premise of ensuring the defogging effect, can maximize the retention of image details, and can flexibly control the degree of image defogging and denoising, thereby improving the recovery quality of the image.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a dehazing and denoising method applied to polarized images. BACKGROUND

[0002] Images captured in adverse weather conditions are often affected by fog, because the light is disturbed by particles in the air, so the quality of the captured images is often significantly degraded, such as poor contrast, color fidelity is not true, and loss of scene details. This decline in image quality is a common obstacle to a wide range of computer vision tasks. For example, computer vision tasks such as urban traffic monitoring, outdoor video surveillance, and autonomous driving. Therefore, it is necessary to improve the quality of images captured in adverse weather conditions through dehazing processing.

[0003] Polarization dehazing is a commonly used dehazing method, which is a method based on atmospheric scattering model, which attributes the degradation of foggy images to the attenuation of scene transmission light and the interference caused by atmospheric scattering light participating in imaging. Because the two have different polarization characteristics, image restoration can be achieved through polarization analysis means. This method does not require images and scene prior knowledge under different gas phase conditions, has low algorithm complexity, good image restoration quality, and is easy to implement.

[0004] However, the greater the dehazing strength of the traditional polarization image-based dehazing algorithm, the higher the noise introduced by the dehazed image, thereby seriously affecting the quality of image restoration. SUMMARY

[0005] The purpose of the present application is to provide a dehazing and denoising method applied to polarized images, which realizes high dehazing while introducing only low noise, thereby improving the quality of image restoration.

[0006] To solve the above technical problems, the present application provides a dehazing and denoising method applied to polarized images, which comprises:

[0007] Obtaining an original image to be dehazed;

[0008] Using a first amplification coefficient to calculate the first local atmospheric light intensity of the original image to restore a high-noise high-dehazing image;

[0009] Using a second amplification coefficient to calculate the second local atmospheric light intensity of the original image to restore a low-noise low-dehazing image, the first amplification coefficient being greater than the second amplification coefficient;

[0010] Obtaining a noise image according to the high-noise high-dehazing image and the low-noise low-dehazing image;

[0011] Using the first local atmospheric light intensity to calculate the transmission coefficient, and performing normalization to obtain a normalized transmission coefficient;

[0012] obtaining a distance-processed noise image according to the normalized transmission coefficient and the noise image; and

[0013] obtaining a low-noise high-fog-removed image according to the high-noise high-fog-removed image and the distance-processed noise image.

[0014] Optionally, the calculation formula of the local atmospheric light intensity is:

[0015]

[0016] wherein, θ is the atmospheric light polarization angle, P is the atmospheric light polarization degree, S0 is the Stokes parameter, I is the original image, and w is a magnification factor. A A ∞

[0017] Optionally, the method for obtaining a noise image according to the high-noise high-fog-removed image and the low-noise low-fog-removed image comprises: subtracting the low-noise low-fog-removed image from the high-noise high-fog-removed image to obtain the noise image.

[0018] Optionally, the calculation formula of the transmission coefficient is:

[0019]

[0020] wherein, A is the local atmospheric light intensity, A is the atmospheric light intensity at infinity. ∞

[0021] Optionally, the method for normalizing the transmission coefficient comprises: dividing each transmission coefficient by the maximum transmission coefficient of all the transmission coefficients.

[0022] Optionally, the method for obtaining a distance-processed noise image according to the normalized transmission coefficient and the noise image comprises: multiplying the normalized transmission coefficient and the noise image to obtain the distance-processed noise image.

[0023] Optionally, the method for obtaining a low-noise high-fog-removed image according to the high-noise high-fog-removed image and the distance-processed noise image comprises: subtracting the distance-processed noise image from the high-noise high-fog-removed image to obtain the low-noise high-fog-removed image.

[0024] Optionally, before subtracting the distance-processed noise image from the high-noise high-fog-removed image, the method further comprises: multiplying the distance-processed noise image by a coefficient greater than 0 and less than 1 to determine the denoising degree.

[0025] ​​Optionally, after obtaining the low-noise high-fog-removed image, the method further comprises: using a three-dimensional block matching algorithm or a bilateral filtering method to further denoise the low-noise high-fog-removed image.

[0026] Optionally, after obtaining the low-noise high-fog-removed image, before further denoising the low-noise high-fog-removed image, the method further comprises: using a Gaussian Laplacian operator to sharpen the low-noise high-fog-removed image to enhance the details of the low-noise high-fog-removed image.

[0027] In summary, in the de-fogging and de-noising method for polarized images provided by the present application, first, an original image to be de-fogged is obtained, then a first local atmospheric light intensity of the original image is calculated using a first amplification coefficient to obtain a high-noise high-fog-removed image, a second local atmospheric light intensity of the original image is calculated using a second amplification coefficient to obtain a low-noise low-fog-removed image, the first amplification coefficient is greater than the second amplification coefficient, then a noise image is obtained according to the high-noise high-fog-removed image and the low-noise low-fog-removed image, then a transmission coefficient is calculated using the first local atmospheric light intensity and normalized to obtain a normalized transmission coefficient, then a distance-processed noise image is obtained according to the normalized transmission coefficient and the noise image, and a low-noise high-fog-removed image is obtained according to the high-noise high-fog-removed image and the distance-processed noise image. The de-fogging and de-noising method provided by the present application can effectively remove the noise in the image without using filtering under the premise of ensuring the de-fogging effect, can maximize the retention of image details, and can flexibly control the degree of image de-fogging and de-noising, thereby improving the recovery quality of the image.

[0028] Further, after obtaining the low-noise high-fog-removed image using the difference method, the low-noise high-fog-removed image can be further denoised using a three-dimensional block matching algorithm or a bilateral filtering method, and before further denoising, the low-noise high-fog-removed image is sharpened using a Gaussian Laplacian operator to enhance the details of the low-noise high-fog-removed image, thereby retaining the details of the image while further de-noising and further improving the recovery quality of the image. BRIEF DESCRIPTION OF DRAWINGS

[0029] Those of ordinary skill in the art will understand that the provided drawings are for better understanding of the present application and do not constitute any limitation on the scope of the present application. Among them:

[0030] Figure 1 is a flowchart of the de-fogging and de-noising method for polarized images provided by an embodiment of the present application.

[0031] Figure 2 is a flowchart of the de-fogging and de-noising method for polarized images provided by an embodiment of the present application.

[0032] Figure 3a is a schematic diagram of an original image provided by an embodiment of the present application.

[0033] Figure 3b is a schematic diagram of a high-noise high-fog-removed image provided by an embodiment of the present application.

[0034] Figure 3c is a schematic diagram of a low-noise high-fog-removed image provided by an embodiment of the present application.

[0035] Figure 3d is a schematic diagram of an image after further denoising provided by an embodiment of the present application. DETAILED DESCRIPTION

[0036] The cause of foggy image degradation is the attenuation of scene transmitted light and the interference caused by atmospheric scattered light participating in imaging. The polarization fog-removing algorithm is based on the atmospheric scattering model: I = D + A = Lt(z) + A ∞ [1 - t(z)], where D = Lt(z) is the scene transmitted light intensity, and A = A ∞ [1 - t(z)] is the atmospheric scattered light intensity. The polarization fog-removing algorithm separates the atmospheric scattered light by polarization analysis means, thereby realizing image restoration, by using the different polarization characteristics of the scene transmitted light and the atmospheric scattered light.

[0037] The image fog-removing model is:

[0038]

[0039] wherein L is the fog-removed restored image, i.e., the final restored fog-free image, I is the original image to be fog-removed, i.e., the foggy image obtained by the detection system, specifically, the polarization image obtained by using the polarization camera, and t is the transmission coefficient, i.e., the proportion of the light that can reach the detection system after particle attenuation, which reflects the relative distance of the scene in the image.

[0040] According to the fog-removing model, the noise of the image after fog-removing will be amplified to different degrees according to the size of the transmission coefficient t (i.e., the relative distance of the scene). The farther the distance, i.e., the smaller t, the more serious the noise amplification.

[0041] In the fog-removing process, the larger the estimated local atmospheric light intensity A, the more the interference of the atmospheric scattered light subtracted when restoring the transmitted light, and the stronger the fog-removing effect. According to the calculation formula of the transmission coefficient t = 1 - A / A ∞ , it can be known that the transmission coefficient t is correspondingly smaller, and the noise in the transmitted light is also amplified more. In other words, when the estimated local atmospheric intensity A is small, the fog-removing effect in the restored image is poor, but the noise is also small.

[0042] The estimation of the local atmospheric light intensity A can be controlled by the amplification coefficient w of the maximum polarization degree, and the calculation formula of the local atmospheric light intensity A is:

[0043]

[0044] wherein, θ A is the atmospheric light polarization angle, P A is the atmospheric light polarization degree, S0 is the Stokes parameter, I is the original image, and w is the amplification coefficient.

[0045] Based on the above analysis, the inventors propose a method for denoising the dehazed image, which first obtains a noise image through image difference, then processes the noise image based on relative distance, and finally obtains a low-noise dehazed image by subtracting the noise image, thereby achieving high dehazing while introducing only low noise, thus improving the recovery quality of the image.

[0046] Specifically, the application provides a dehazing and denoising method applied to a polarized image, which comprises:

[0047] obtaining an original image to be dehazed;

[0048] calculating a first local atmospheric light intensity of the original image using a first amplification coefficient to obtain a high-noise high-dehazing image;

[0049] calculating a second local atmospheric light intensity of the original image using a second amplification coefficient to obtain a low-noise low-dehazing image, wherein the first amplification coefficient is greater than the second amplification coefficient;

[0050] obtaining a noise image according to the high-noise high-dehazing image and the low-noise low-dehazing image;

[0051] calculating a transmission coefficient using the first local atmospheric light intensity and performing normalization to obtain a normalized transmission coefficient;

[0052] obtaining a distance-processed noise image according to the normalized transmission coefficient and the noise image; and

[0053] obtaining a low-noise high-dehazing image according to the high-noise high-dehazing image and the distance-processed noise image.

[0054] The dehazing and denoising method provided by the application can effectively remove the noise in the image without using filtering under the premise of ensuring the dehazing effect, thereby maximizing the preservation of image details and flexibly regulating the degree of image dehazing and denoising, thus improving the recovery quality of the image.

[0055] In order to make the purposes, advantages and features of the present application clearer, the following further describes the present application in conjunction with the drawings and specific embodiments. It should be noted that the drawings are all very simplified and not drawn to scale, and are only used to facilitate and clarify the purpose of describing the embodiments of the present application. In addition, the structures shown in the drawings are often part of the actual structures. In particular, the emphasis shown in each drawing is different, and sometimes different scales are used.

[0056] As used in the present application, the singular forms "a", "an" and "the" include plural referents, the term "or" is generally used in the sense of "and / or", the term "several" is generally used in the sense of "at least one", the term "at least two" is generally used in the sense of "two or more", and in addition, the terms "first", "second", "third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second", "third" can explicitly or implicitly include one or at least two of the features, the term "proximal" generally refers to the end closer to the operator, the term "distal" generally refers to the end closer to the patient, "one end" and "the other end" and "proximal" and "distal" generally refer to two parts corresponding to each other, which not only includes the end point, the terms "mounting", "connecting", "connecting" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements.

[0057] In addition, as used in the present application, a component is provided in another component, which generally only indicates that there is a connection, coupling, cooperation or transmission relationship between the two components, and the connection, coupling, cooperation or transmission between the two components can be direct or indirect through an intermediate component, and cannot be understood as indicating or implying the spatial positional relationship between the two components, i.e. one component can be in any orientation inside, outside, above, below or one side of another component, unless the content is otherwise explicitly indicated. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0058] Figure 1 is a flowchart of a dehazing and denoising method for polarized images provided by an embodiment of the present application. Referring to FIG. 1, the present application provides a dehazing and denoising method for polarized images, which comprises: Figure 1

[0059] S1: obtaining an original image to be dehazed;

[0060] ​S2: calculating a first local atmospheric light intensity of the original image using a first amplification coefficient to obtain a high-noise high-fog-removed image;

[0061] S3: calculating a second local atmospheric light intensity of the original image using a second amplification coefficient to obtain a low-noise low-fog-removed image, the first amplification coefficient being greater than the second amplification coefficient;

[0062] S4: obtaining a noise image according to the high-noise high-fog-removed image and the low-noise low-fog-removed image;

[0063] S5: calculating a transmission coefficient using the first local atmospheric light intensity and performing normalization to obtain a normalized transmission coefficient;

[0064] S6: obtaining a distance-processed noise image according to the normalized transmission coefficient and the noise image;

[0065] S7: obtaining a low-noise high-fog-removed image according to the high-noise high-fog-removed image and the distance-processed noise image.

[0066] Figure 2 is a flow chart of a de-fogging and de-noising method applied to a polarized image provided by an embodiment of the present application, and the following will be described in combination with Figure 1 and Figure 2 The de-fogging and de-noising method applied to a polarized image provided by an embodiment of the present application will be described in detail.

[0067] In step S1, an original image to be de-fogged is obtained. That is, a foggy image is obtained by a detection system, and the image is taken as the original image to be de-fogged. The original image can be represented by I. In the embodiment, the original image I is a polarized image obtained by using a polarized camera.

[0068] In step S2, a first local atmospheric light intensity of the original image is calculated using a first amplification coefficient to obtain a high-noise high-fog-removed image.

[0069] In step S3, a second local atmospheric light intensity of the original image is calculated using a second amplification coefficient to obtain a low-noise low-fog-removed image, the first amplification coefficient being greater than the second amplification coefficient.

[0070] The calculation formula of the local atmospheric light intensity is as follows:

[0071]

[0072] wherein θ A is an atmospheric light polarization angle, P A is an atmospheric light polarization degree, S0 is a Stokes parameter, I is the original image, and w is an amplification coefficient. Different local atmospheric light intensities A corresponding to different amplification coefficients w can be obtained in the above formula.

[0073] In step S2, the first local atmospheric light intensity A1 is calculated using a first amplification coefficient w1, which is a relatively large amplification coefficient, so as to obtain a high-noise high-fog-removed image. In step S3, the second local atmospheric light intensity A2 is calculated using a second amplification coefficient w2, which is a relatively small amplification coefficient, so as to obtain a low-noise low-fog-removed image. Here, relatively large and relatively small are relative, that is, the first amplification coefficient w1 is larger than the second amplification coefficient w2, and the specific values can be selected according to actual conditions.

[0074] After the first local atmospheric light intensity A1 and the second local atmospheric light intensity A2 are obtained, a high-noise high-fog-removed image and a low-noise low-fog-removed image can be obtained respectively according to the calculation formula of the fog-removed recovery image (i.e. according to the image fog-removed model). The image fog-removed model is:

[0075]

[0076] Wherein, L is the fog-removed recovery image, that is, the final recovered fog-free image, I is the original image to be fog-removed, and t is the transmission coefficient. Different local atmospheric light intensities are substituted into the above formula to obtain different fog-removed recovery images. The first local atmospheric light intensity A1 is substituted to obtain a high-noise high-fog-removed image (i.e. a high-noise high-fog-removed fog-removed recovery image), and the second local atmospheric light intensity A2 is substituted to obtain a low-noise low-fog-removed image (i.e. a low-noise low-fog-removed fog-removed recovery image).

[0077] In this embodiment, the degree of fog removal in the finally obtained low-noise high-fog-removed image is related to the degree of fog removal of the high-noise high-fog-removed image obtained in step S2, so the degree of fog removal in the final low-noise high-fog-removed image can be regulated by the selection of the first amplification coefficient w1. The degree of denoising in the finally obtained low-noise high-fog-removed image is related to the high-noise high-fog-removed image obtained in step S2 and the low-noise low-fog-removed image obtained in step S3 (because the two images will be subtracted to obtain a noise image and then a distance-processed noise image), so the degree of fog removal in the final low-noise high-fog-removed image can be regulated by the selection of the first amplification coefficient w1 and the second amplification coefficient w2. That is, the degrees of image fog removal and denoising can be flexibly regulated by the selection of the first amplification coefficient w1 and the second amplification coefficient w2.

[0078] In step S4, a noise image is obtained according to the high-noise high-fog-removed image and the low-noise low-fog-removed image.

[0079] For example, the high-noise high-fog-removed image can be subtracted from the low-noise low-fog-removed image to obtain the noise image, which can reflect the noise distribution.

[0080] In step S5, the transmission coefficient is calculated using the first local atmospheric light intensity, and normalization is performed to obtain a normalized transmission coefficient.

[0081] The formula for calculating the transmission coefficient is:

[0082]

[0083] Where A is the local atmospheric light intensity, A ∞ is the atmospheric light intensity at infinity.

[0084] In this embodiment, first, the transmission coefficient t is calculated using the first local atmospheric light intensity, that is, a transmission coefficient image is obtained, that is, a depth image of the high-noise high-fog-removed image is obtained. Then, the transmission coefficient t is normalized. Exemplarily, the method for normalizing the transmission coefficient includes: selecting the maximum value of all the transmission coefficients, that is, the maximum transmission coefficient, and dividing each transmission coefficient by the maximum transmission coefficient, thereby completing the normalization of the transmission coefficient.

[0085] In step S6, a distance-processed noise image is obtained according to the normalized transmission coefficient and the noise image.

[0086] Exemplarily, the distance-processed noise image is obtained by multiplying the normalized transmission coefficient and the noise image.

[0087] In step S7, a low-noise high-fog-removed image is obtained according to the high-noise high-fog-removed image and the distance-processed noise image.

[0088] Exemplarily, the low-noise high-fog-removed image is obtained by subtracting the distance-processed noise image from the high-noise high-fog-removed image.

[0089] Before subtracting the distance-processed noise image from the high-noise high-fog-removed image, the distance-processed noise image can also be multiplied by a coefficient greater than 0 and less than 1 to determine the degree of noise removal. The coefficient can be determined according to actual needs. The distance-processed noise image multiplied by the coefficient is used to regulate the appropriate degree of noise removal to prevent loss of details.

[0090] The dehazing and denoising method for polarized images provided by this invention first acquires the original image to be dehazed. Then, a first local atmospheric light intensity of the original image is calculated using a first magnification factor to obtain a high-noise, high-dehazed image. A second local atmospheric light intensity of the original image is calculated using a second magnification factor to obtain a low-noise, low-dehazed image. The first magnification factor is greater than the second magnification factor. Next, a noise image is obtained based on the high-noise, high-dehazed image and the low-noise, low-dehazed image. Then, a transmission coefficient is calculated using the first local atmospheric light intensity and normalized to obtain a normalized transmission coefficient. Finally, a distance-processed noise image is obtained based on the normalized transmission coefficient and the noise image, and a low-noise, high-dehazed image is obtained based on the high-noise, high-dehazed image and the distance-processed noise image. The dehazing and denoising method provided by this invention effectively removes noise from images without filtering while ensuring dehazing performance. It maximizes the preservation of image details and allows for flexible control of the degree of dehazing and denoising, thereby improving the image restoration quality.

[0091] Understandably, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows; however, these steps are not necessarily executed in the order indicated by the arrows, unless explicitly stated otherwise. There is no strict order constraint for the execution of these steps; they can be performed in other orders. Furthermore, Figure 1 At least some steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but may be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but may be executed alternately or in turn with other steps or at least some of the sub-steps or stages in other steps. For example, please refer to... Figure 1 and Figure 2 As shown, step S2 can be executed before step S3, or step S3 can be executed before step S2, or steps S2 and S3 can be executed simultaneously. After completing part of steps S2 and S3 (e.g., calculating the local atmospheric light intensity), the remaining parts of steps S2 and S3 (e.g., image restoration) and step S4 can be executed first, followed by step S5. Alternatively, step S5 can be executed first, followed by the remaining parts of steps S2 and S3 and step S4, but this is not limited to these steps.

[0092] In the embodiment, after obtaining the low-noise high-fog-removed image, further denoising operation can be performed on the low-noise high-fog-removed image, for example, three-dimensional block matching algorithm (BM3D) or bilateral filtering method can be used to further denoise the low-noise high-fog-removed image. Considering that such operation will still lose part of the image details while denoising, therefore, after obtaining the low-noise high-fog-removed image, before performing further denoising operation on the low-noise high-fog-removed image, the low-noise high-fog-removed image is sharpened by using a Gaussian Laplacian (LoG) operator to strengthen the details of the low-noise high-fog-removed image, so as to retain the details of the image while further denoising, and further improve the recovery quality of the image.

[0093] It can be understood that the subtraction and multiplication involved in the above steps refer to matrix operations, that is, the above-mentioned images, transmission coefficients and noise maps are represented in the form of matrices, and the operations between the images, transmission coefficients and noise maps refer to the operations between the matrices.

[0094] Figure 3a is a schematic diagram of an original image provided by an embodiment of the present application, Figure 3b is a schematic diagram of a high-noise high-fog-removed image provided by an embodiment of the present application, Figure 3c is a schematic diagram of a low-noise high-fog-removed image provided by an embodiment of the present application, Figure 3d is a schematic diagram of an image after further denoising processing provided by an embodiment of the present application. As shown in Figures 3a to 3c , by using the fog-removing and denoising method applied to polarized images in the embodiment, the noise in the image can be effectively removed without using filtering under the premise of ensuring the fog-removing effect, which can maximize the retention of image details and flexibly control the degree of image fog-removing and denoising, thereby improving the recovery quality of the image. As shown in Figure 3c , and Figure 3d , after further denoising processing of the low-noise high-fog-removed image shown in Figure 3c , the remaining noise can be further removed, and the recovery quality of the image is further improved.

[0095] ​In summary, the application provides a polarized image defogging and denoising method, which first acquires an original image to be defogged, then calculates a first local atmospheric light intensity of the original image using a first amplification coefficient to obtain a high-noise high-defogging image, calculates a second local atmospheric light intensity of the original image using a second amplification coefficient to obtain a low-noise low-defogging image, the first amplification coefficient is greater than the second amplification coefficient, then obtains a noise image according to the high-noise high-defogging image and the low-noise low-defogging image, then calculates a transmission coefficient using the first local atmospheric light intensity and performs normalization to obtain a normalized transmission coefficient, then obtains a distance-processed noise image according to the normalized transmission coefficient and the noise image, and obtains a low-noise high-defogging image according to the high-noise high-defogging image and the distance-processed noise image. The polarized defogging and denoising method provided by the application can effectively remove the noise in the image without using filtering under the premise of ensuring the defogging effect, can maximize the retention of image details, and can flexibly control the degree of image defogging and denoising, thereby improving the recovery quality of the image.

[0096] Further, after obtaining the low-noise high-defogging image using the differential method, the low-noise high-defogging image can be further denoised using a three-dimensional block matching algorithm or a bilateral filtering method, and before further denoising, the low-noise high-defogging image is sharpened using a Gaussian Laplacian operator to enhance the details of the low-noise high-defogging image, so that the details of the image are retained while further denoising, and the recovery quality of the image is further improved.

[0097] The above description is only a description of the preferred embodiments of the application, and does not limit the scope of the application in any way. Any person skilled in the art can make possible changes and modifications to the technical solutions of the application using the disclosed methods and technical contents without departing from the spirit and scope of the application. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the application, without departing from the technical solutions of the application, all belong to the protection scope of the application.

Claims

1. A method for dehazing and denoising polarized images, characterized in that, The method includes: Obtain the original image to be dehazed; The first local atmospheric light intensity of the original image is calculated using the first magnification factor, and a high-noise, high-dehazing image is obtained according to the image dehazing model. Calculate the second local atmospheric light intensity of the original image using the second amplification factor, and obtain a low-noise and low-dehazed image according to the image dehazing model, where the first amplification factor is greater than the second amplification factor; wherein, the image dehazing model is: 0 < t < 1, L is the dehazed and restored image, I is the original image to be dehazed, t is the transmission coefficient, and A is the local atmospheric light intensity; A noise image is obtained based on the high-noise, high-dehaze image and the low-noise, low-dehaze image; The transmission coefficient is calculated using the first local atmospheric light intensity and then normalized to obtain the normalized transmission coefficient. A distance-processed noise image is obtained based on the normalized transmission coefficient and the noise image; and A low-noise, high-dehaze image is obtained by comparing the high-noise, high-dehaze image with the distance-processed noise image.

2. The method for dehazing and denoising polarized images according to claim 1, characterized in that, The formula for calculating the local atmospheric light intensity is: Where, θ A It is the atmospheric light polarization angle, P A S is the atmospheric polarization degree, S0 is the Stokes parameter, I is the original image, and w is the magnification factor.

3. The method for dehazing and denoising polarized images according to claim 1, characterized in that, The method for obtaining a noise image based on the high-noise, high-dehaze image and the low-noise, low-dehaze image includes: subtracting the low-noise, low-dehaze image from the high-noise, high-dehaze image to obtain the noise image.

4. The method for dehazing and denoising polarized images according to claim 1, characterized in that, The formula for calculating the transmission coefficient is: Where A is the local atmospheric light intensity, A ∞ It is the atmospheric light intensity at infinity.

5. The method for dehazing and denoising polarized images according to claim 4, characterized in that, The method for normalizing the transmission coefficients includes dividing each of the transmission coefficients by the largest of all the transmission coefficients.

6. The method for dehazing and denoising polarized images according to claim 1, characterized in that, A method for obtaining a distance-processed noise image based on the normalized transmission coefficient and the noise image includes: multiplying the normalized transmission coefficient and the noise image to obtain the distance-processed noise image.

7. The method for dehazing and denoising polarized images according to claim 1, characterized in that, The method for obtaining a low-noise, high-dehaze image based on the high-noise, high-dehaze image and the distance-processed noise image includes: subtracting the distance-processed noise image from the high-noise, high-dehaze image to obtain the low-noise, high-dehaze image.

8. The method for dehazing and denoising polarized images according to claim 7, characterized in that, Before subtracting the distance-processed noisy image from the high-noise, high-dehaze image, the method further includes: multiplying the distance-processed noisy image by a coefficient greater than 0 and less than 1 to determine the degree of denoising.

9. The method for dehazing and denoising polarized images according to claim 1, characterized in that, After obtaining the low-noise, high-dehaze image, the method further includes: using a three-dimensional block matching algorithm or a bilateral filtering method to further denoise the low-noise, high-dehaze image.

10. The method for dehazing and denoising polarized images according to claim 9, characterized in that, After obtaining the low-noise, high-dehaze image, and before performing further denoising processing on the low-noise, high-dehaze image, the method further includes: performing a sharpening operation on the low-noise, high-dehaze image using the Laplacian Gaussian operator to enhance the details of the low-noise, high-dehaze image.

Images