Low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions

Abstract

The invention provides a low-complexity single-frame image reconstruction method applicable to extreme rain and fog conditions, and aims to solve the problems that an image is low in contrast ratio and fuzzy and a target is not easy to recognize under the extreme rain and fog weather conditions. An atmospheric light matrix and global atmospheric light parameters are obtained by using multiple Gaussian blur, preliminary defogging of the image is realized by combining an atmospheric defogging model considering a rain and fog retention factor, and finally brightness adjustment of the image is performed by using image normalization and gamma transformation to enhance weak target information so as to realize image defogging. The method only depends on the original degraded image, does not need any prior information and data set, has good scene adaptability, and can simply and quickly realize image reconstruction.

Description

Technical Field

[0001] The present invention belongs to the technical fields of computer vision, image sharpening, etc., and specifically relates to a low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions. Background Art

[0002] Under extremely adverse weather conditions such as heavy rain, dense fog, and strong haze, when shooting scene images for target identification, the scattering and absorption of light by scattering particles in the atmosphere (such as rain and fog) usually leads to a significant decrease in image contrast, color distortion, scene blur, and difficulty in clearly identifying targets. The image blur problem is particularly serious in extreme rainy and foggy weather. It not only affects the acquisition of visual information in the fields of security monitoring, traffic management, and maritime rescue, but also has a significant negative impact on professional applications such as optical remote sensing, military reconnaissance, and target identification.

[0003] Existing image defogging or haze removal technologies are mainly divided into three categories. Among them, image enhancement technology focuses on improving the contrast of the scene and suppressing noise through image processing technology. The applicable scenarios are very limited and cannot effectively improve the image reconstruction problem in extreme rainy and foggy weather. The invention patent with application number CN202311274222.8 and name “An image defogging method and device based on Retinex and variational model” performs image enhancement and defogging based on Retinex and variational models. This method is more effective for processing in misty scenes, but has limited processing effect under extreme meteorological conditions and cannot reconstruct the target information; the method based on physical models relies on prior information and has high computational complexity, which reduces the image reconstruction speed. The invention patent with application number CN201910720005.4 and name “Image defogging method based on prior information” The invention patent of "method" combines dark channel prior and color attenuation prior to realize the image defogging function, which has good effects for misty scenes, but it relies on multiple prior information, the processing process is relatively complicated, and the image reconstruction speed is greatly reduced; the image defogging technology based on deep learning depends on the scale of the training set and the type of scenes, and the generalization and real-time performance of the model need to be improved. The invention patents with application numbers CN202410974894.8, CN201811492894.5 and CN202010000331.0 are all image defogging methods based on different deep learning models. Due to the limitation of the training set, their versatility is still insufficient, especially in weak texture scenes under extreme rain and fog conditions. The effective feature information of the target is very limited. The deep learning method not only cannot accurately restore the target information, but also introduces artifacts. Therefore, the development of low-complexity single-frame image reconstruction methods suitable for extreme rain and fog conditions still faces huge challenges. Summary of the invention

[0004] In view of the problem that the existing image reconstruction technology cannot restore the target information under adverse weather conditions, the present invention proposes a low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions. By simulating the image degradation mechanism of rain and fog acting on a clear scene, the atmospheric light matrix is ​​obtained by Gaussian blur to achieve preliminary de-raining and defogging of the image, and the image brightness is adjusted to achieve the final image reconstruction. Compared with the prior art, the method proposed by the present invention does not involve complex traversal calculation processes and iterative loop algorithms, and the calculation process is simple. By combining image augmentation technology and atmospheric imaging physical models, low-complexity single-frame image reconstruction under extreme rain and fog weather conditions is achieved, and the detection distance is close to 2 times the visibility, which is suitable for target recognition and detection in severely degraded images.

[0005] The present invention aims at the problems of low image contrast, blur and difficult target identification under extreme rain and fog weather conditions. First, the contrast is improved by overall normalization of the image, and then multiple Gaussian blurs are used to obtain the atmospheric light matrix and global atmospheric light parameters. The atmospheric defogging model considering the rain and fog retention factor is combined to achieve preliminary image defogging. Finally, the image brightness is adjusted by image normalization and gamma transformation to enhance the weak target information and achieve image defogging. The method only relies on the original degraded image, does not require any prior information and data set, has good scene adaptability, and can simply and quickly achieve image reconstruction.

[0006] The technical solution specifically adopted by the present invention to solve the technical problem is:

[0007] A low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions: first, the rain and fog degraded original image is normalized as a whole to improve the contrast, and then multiple Gaussian blurs are used to obtain the atmospheric light matrix and global atmospheric light parameters. The atmospheric defogging model considering the rain and fog retention factor is combined to achieve preliminary image defogging. Finally, image normalization and gamma transform are used to adjust the image brightness to enhance weak target information and achieve image defogging.

[0008] Furthermore, the contrast is improved by overall normalization of the rain and fog degraded original image as follows: if it is a color image, the overall maximum value of the three channels R, G, and B of the image is extracted, and then the value is removed from the entire image so that the overall grayscale range of the image is within the range of 0-1; if it is a grayscale image, the maximum value of the grayscale image is obtained for image normalization.

[0009] Furthermore, the method of obtaining the atmospheric light matrix and the global atmospheric light parameters by using multiple Gaussian blurs specifically comprises obtaining the atmospheric light matrix A and obtaining the global atmospheric light A by using multiple Gaussian filtering processes. ∞ ;

[0010] The Gaussian filter is based on a two-dimensional Gaussian function, expressed as:

[0011]

[0012] Among them, i and j represent the pixel coordinates relative to the center of the filter, σ is the standard deviation of the Gaussian distribution, which is used to control the diffusion degree of the filter. By setting different σ for fog scenes with different concentrations, the atmospheric light matrix of different environments is inverted. The thicker the fog in the scene, the larger the value of σ.

[0013] Gaussian filtering is performed on the image. The process is obtained by convolving the Gaussian filter G(i,j,σ) with the image. The process is as follows:

[0014]

[0015] Among them, G(i,j,σ) represents a Gaussian filter with a standard deviation of σ, I(xi,yj) represents the pixel value at position (xi,yj) in image I, and GS(I(x,y),σ) represents the process of filtering I through a Gaussian filter with a standard deviation of σ.

[0016] Furthermore, the process of obtaining the atmospheric light matrix A by using multiple Gaussian filtering is specifically as follows:

[0017] First, perform a Gaussian filter on the input image, and the obtained image I1 is a Gaussian blurred image of the input image:

[0018] I1=GS(I,σ1)

[0019] Subtract the original input image I from the first blurred image I1 and take the absolute value. Perform a second Gaussian blur on the result to obtain the image I2 with image edge information. The parameter σ2 is set greater than σ1:

[0020] I2=GS(|i-I1|,σ2)

[0021] Then, the image I1 obtained by the first blur is subtracted from the image I2 obtained by the second blur, and the result is Gaussian blurred. The parameter σ3 is set greater than σ2, and the result is used as the atmospheric light matrix A:

[0022] A=GS(I1-I2,σ3).

[0023] Further, the global atmospheric light A is obtained ∞ The specific process is as follows: For color images, obtain A by channel ∞ , each channel has its own corresponding A ∞ , by obtaining the maximum values ​​of the atmospheric light matrix A in the three channels of R, G, and B, the maximum grayscale values ​​obtained in each channel are taken as the final A ∞ ; For grayscale, it is the global grayscale maximum value of the grayscale image.

[0024] Furthermore, the atmospheric defogging model that takes into account the rain and fog retention factor is used to achieve preliminary defogging of the image. Specifically, a rain and fog retention factor ω is set, and defogging and restoration are performed using the following formula:

[0025]

[0026] Among them, I is the image of the original rain and fog image after preprocessing, A is the obtained atmospheric light matrix, ω is the rain and fog retention factor, and A ∞ is the global atmospheric light value, and J is the preliminary dehazed image.

[0027] Furthermore, the brightness adjustment of the image by using image normalization and gamma transformation is specifically as follows: firstly, the preliminary defogging result is overall normalized, and then the brightness adjustment of the image is realized by using gamma correction, as shown below:

[0028] I deh aze =J' γ

[0029] Among them, J' is the dehazed image obtained after overall normalization, I deh aze is the final dehazed image; γ is the correction factor. When γ=1, the output image is the same as the input image; when γ<1, the brightness of the output image increases; when γ>1, the brightness of the output image decreases.

[0030] And, a low-complexity single-frame image reconstruction system suitable for extreme rain and fog conditions, including: a normalization module, which improves the contrast by overall normalization of the rain and fog degraded original image; a Gaussian filtering module, which uses multiple Gaussian blurs to obtain the atmospheric light matrix and global atmospheric light parameters; a preliminary defogging module, which realizes preliminary defogging of the image in combination with an atmospheric defogging model that considers rain and fog retention factors; and a brightness adjustment module, which uses image normalization and gamma transform to adjust the brightness of the image to enhance weak target information and realize image defogging.

[0031] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a low-complexity single-frame image reconstruction method applicable to extreme rain and fog conditions as described above are implemented.

[0032] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions as described above.

[0033] Compared with the prior art, the present invention and its preferred solution firstly use an imaging sensor to collect rain and fog scene images, and perform normalization preprocessing on the rain and fog images; then construct multiple different Gaussian filters to perform multiple Gaussian filtering on the preprocessed images to obtain an atmospheric light matrix, and use the atmospheric light matrix to obtain the global atmospheric light value; substitute the obtained parameters into the atmospheric scattering model, and select the best rain and fog retention factor to achieve preliminary image defogging; finally, use image normalization processing and gamma transformation to complete the brightness adjustment of the image, and finally obtain a clear defogging image. The method has a simple processing flow, can realize rapid processing of images, has a good defogging effect for extreme rain and fog images, and can effectively restore target information under severe meteorological conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0035] Figure 1 This is a flow chart of a method according to an embodiment of the present invention;

[0036] Figure 2 Grayscale images of various rainy and foggy scenes indoors and outdoors;

[0037] Figure 3 For using the method of the embodiment of the present invention Figure 2 The resulting image after image reconstruction processing. DETAILED DESCRIPTION

[0038] In order to make the features and advantages of the present invention more clearly understood, the following embodiments are specifically described in detail as follows:

[0039] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0040] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0041] like Figure 1 As shown, the defogging process of the embodiment of the present invention includes the following steps:

[0042] Step 1: Use an industrial camera or other imaging sensor to obtain a color image or grayscale image of a rainy and foggy scene; perform normalization preprocessing on the collected rainy and foggy image. If it is a color image, it is necessary to extract the maximum value of the three channels of the image R, G, and B as a whole, and then use the entire image to remove the value so that the overall grayscale range of the image is within the range of 0-1. If it is a grayscale image, it is only necessary to obtain the maximum value of the grayscale image for image normalization. The specific process is as follows:

[0043]

[0044] in, represents the original rain and fog image, and I represents the image after normalization preprocessing.

[0045] Step 2: Use multiple Gaussian filtering processes to obtain the atmospheric light matrix A and the global atmospheric light A ∞ The following is the filter construction work that needs to be done before Gaussian filtering of the image:

[0046] The construction of Gaussian filter kernel,Gaussian filter is based on two-dimensional Gaussian function, and its expression is:

[0047]

[0048] Among them, i and j represent the pixel coordinates relative to the center of the filter, σ is the standard deviation of the Gaussian distribution, which controls the diffusion degree of the filter. The larger the σ, the stronger the blurring effect of the image. In order to more accurately reflect the atmospheric light matrix of different environments, different σ will be set for fog scenes of different concentrations. The thicker the fog in the actual scene, the larger the value of σ will be.

[0049] Gaussian filtering is performed on the image. The process is obtained by convolving the Gaussian filter G(i,j,σ) with the image. The process is as follows:

[0050]

[0051] Among them, G(i,j,σ) represents a Gaussian filter with a standard deviation of σ, I(xi,yj) represents the pixel value at position (xi,yj) in image I, and GS(I(x,y),σ) represents the process of filtering I through a Gaussian filter with a standard deviation of σ.

[0052] Finally, the atmospheric light matrix A is obtained through multiple Gaussian filtering. The specific implementation process is as follows:

[0053] First, perform a Gaussian filter on the input image. The appropriate parameter σ1 set for this filter should be small. The obtained image I1 is a Gaussian blurred image of the input image. The process is as follows:

[0054] I1=GS(I,σ1) (4)

[0055] Obtain a blurred image I1, make a difference between the original input image I and the blurred image I1 and take the absolute value, perform a second Gaussian blur on the result, and obtain a darker image I2 with image edge information. The parameter σ2 set this time should be slightly larger than σ1. The process is as follows:

[0056] I2=GS(|I-I1|,σ2) (5)

[0057] Then, the image I1 obtained by the first blur is subtracted from the image I2 obtained by the second blur, and the result is Gaussian blurred. The parameter σ3 set for this blur should be greater than σ2. The image obtained by this operation is used as the final atmospheric light matrix A in this method. The calculation process is as follows:

[0058] A=GS(I1-I2,σ3) (6)

[0059] Obtain global atmospheric light A ∞ In fact, this method obtains A for color images and grayscale images ∞ There are some differences. For color images, it is necessary to obtain A by channel ∞ , each channel has its own corresponding A ∞ The process is to find the maximum value of the atmospheric light matrix A obtained in step 3 in the three channels of R, G, and B, and take the grayscale maximum value obtained in each channel as the final A ∞ , while the grayscale image only needs to take one A ∞ , is the global grayscale maximum value of the grayscale image.

[0060] Step 3: Substitute the obtained parameters into the atmospheric scattering model to obtain the preliminary defogging image. In the actual implementation process, if the obtained atmospheric light matrix A is completely substituted into the atmospheric scattering model for defogging operation, the final defogging image will appear dim in brightness, which is inconsistent with the real scene observed by the naked eye. This is because the atmospheric light must exist in the real field of view. Therefore, a rain and fog retention factor v needs to be set during the defogging process to retain a certain amount of rain and fog components for the final defogging image. Usually, v is set to 0.95. The specific defogging recovery is as follows:

[0061]

[0062] Step 4: The preliminary defogging image obtained by the above atmospheric scattering model has removed most of the rain and fog. However, due to the large proportion of rain and fog in extreme rain and fog scenes, when most of the rain and fog are removed, the overall image will show a darker brightness, so the brightness of the preliminary defogging result needs to be adjusted. First, the result needs to be normalized as a whole. The process is similar to step 1. The image is divided by the maximum value of the three channels of the image to achieve preliminary grayscale stretching and contrast enhancement. However, this operation cannot completely restore the brightness of the defogging image to a natural state. Therefore, this method will use gamma correction to achieve image brightness adjustment. The process is as follows:

[0063] I deh aze =J' γ (8)

[0064] Among them, J' is the dehazed image obtained after overall normalization, and the final dehazed image is obtained; γ is the correction factor. When γ=1, the output image is the same as the input image, that is, there is no change; when γ<1, the brightness of the output image increases, and the image becomes brighter; low-brightness details will be more prominent; when γ>1, the brightness of the output image decreases, and the image becomes darker; high-brightness details will be compressed; the parameters should be used in combination with actual conditions, and this method does not specifically limit their values.

[0065] After the above process, this method can successfully restore images under extreme rain and fog conditions to target visible images.

[0066] like Figure 3 As shown, the method provided by the above embodiment of the present invention is used for Figure 2 The results of image reconstruction processing on grayscale images of various indoor and outdoor rain and fog scenes prove the effectiveness of the above scheme.

[0067] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0068] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0069] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0070] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0071] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0072] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

[0073] The present invention is not limited to the above-mentioned optimal implementation mode. Anyone can derive various other forms of a low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions under the inspiration of the present invention. All equal changes and modifications made within the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions, characterized by: Firstly, the rain-fog degraded original image is normalized as a whole to improve the contrast. Then, multiple Gaussian blurs are used to obtain the atmospheric light matrix and global atmospheric light parameters. The atmospheric defogging model considering the rain-fog retention factor is combined to achieve preliminary image defogging. Finally, image normalization and gamma transform are used to adjust the image brightness to enhance the weak target information and achieve image defogging.

2. The low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions according to claim 1, characterized in that: The method of improving the contrast by overall normalization of the rain and fog degraded original image is as follows: if it is a color image, the overall maximum value of the three channels R, G, and B of the image is extracted, and then the value is removed from the entire image so that the overall grayscale range of the image is within the range of 0-1; if it is a grayscale image, the maximum value of the grayscale image is obtained to normalize the image.

3. The low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions according to claim 1, characterized in that: The method of using multiple Gaussian blurs to obtain the atmospheric light matrix and the global atmospheric light parameters is to use multiple Gaussian filtering processes to obtain the atmospheric light matrix A and the global atmospheric light A. ∞ ; The Gaussian filter is based on a two-dimensional Gaussian function, expressed as: Among them, i and j represent the pixel coordinates relative to the center of the filter, σ is the standard deviation of the Gaussian distribution, which is used to control the diffusion degree of the filter. By setting different σ for fog scenes with different concentrations, the atmospheric light matrix of different environments is inverted. The thicker the fog in the scene, the larger the value of σ. Gaussian filtering is performed on the image. The process is obtained by convolving the Gaussian filter G(i,j,σ) with the image. The process is as follows: Among them, G(i,j,σ) represents a Gaussian filter with a standard deviation of σ, I(xi,yj) represents the pixel value at position (xi,yj) in image I, and GS(I(x,y),σ) represents the process of filtering I through a Gaussian filter with a standard deviation of σ.

4. The low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions according to claim 3, characterized in that: The specific process of obtaining the atmospheric light matrix A using multiple Gaussian filtering is as follows: First, perform a Gaussian filter on the input image, and the obtained image I1 is a Gaussian blurred image of the input image: I1=GS(I,σ1) Subtract the original input image I from the first blurred image I1 and take the absolute value. Perform a second Gaussian blur on the result to obtain the image I2 with image edge information. The parameter σ2 is set greater than σ1: I2=GS(|I-I1|,σ2) Then, the image I1 obtained by the first blur is subtracted from the image I2 obtained by the second blur, and the result is Gaussian blurred. The parameter σ3 is set greater than σ2, and the result is used as the atmospheric light matrix A: A=GS(I1-I2,σ3).

5. The low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions according to claim 4, characterized in that: Obtain global atmospheric light A ∞ The specific process is as follows: For color images, obtain A by channel ∞ , each channel has its own corresponding A ∞ , by obtaining the maximum values ​​of the atmospheric light matrix A in the three channels of R, G, and B, the maximum grayscale values ​​obtained in each channel are taken as the final A ∞ ; For grayscale, it is the global grayscale maximum value of the grayscale image.

6. The low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions according to claim 5, characterized in that: The atmospheric defogging model combined with the rain and fog retention factor to achieve preliminary defogging of the image is specifically as follows: a rain and fog retention factor ω is set, and defogging and restoration are performed by the following formula: Among them, I is the image of the original rain and fog image after preprocessing, A is the obtained atmospheric light matrix, ω is the rain and fog retention factor, and A ∞ is the global atmospheric light value, and J is the preliminary dehazed image.

7. The low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions according to claim 1, characterized in that: The brightness adjustment of the image by using image normalization and gamma transformation is specifically as follows: firstly, the preliminary defogging result is overall normalized, and then the brightness adjustment of the image is realized by using gamma correction, as shown below: I dehaze =J' γ Among them, J' is the dehazed image obtained after overall normalization, I dehaze is the final dehazed image; γ is the correction factor. When γ=1, the output image is the same as the input image; when γ<1, the brightness of the output image increases; when γ>1, the brightness of the output image decreases.

8. A low-complexity single-frame image reconstruction system suitable for extreme rain and fog conditions, characterized in that: include: The normalization module improves the contrast by normalizing the rain-fog degraded original image as a whole. The Gaussian filter module uses multiple Gaussian blurs to obtain the atmospheric light matrix and global atmospheric light parameters. The preliminary defogging module combines the atmospheric defogging model that takes into account the rain and fog retention factor to achieve preliminary defogging of the image; And the brightness adjustment module uses image normalization and gamma transform to adjust the brightness of the image to enhance weak target information and achieve image defogging.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of a low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions as described in any one of claims 1-7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a low-complexity single-frame image reconstruction method suitable for extreme rain and fog conditions as described in any one of claims 1 to 7 are implemented.

Images