A low-illumination image enhancement method based on color space transformation

Abstract

The application discloses a low-illumination image enhancement method based on color space transformation. The method comprises the following steps: acquiring an initial low-illumination image, pre-processing the initial low-illumination image to obtain an intensity graph of the initial low-illumination image; generating a saturation graph according to the intensity graph of the initial low-illumination image; generating a hue graph based on the saturation graph, the intensity graph of the initial low-illumination image and the initial low-illumination image; processing the hue graph to obtain a horizontal axis set and a vertical axis set of the hue graph; and setting an adaptive intensity collapse function of the intensity graph of the initial low-illumination image. The application solves the technical problem of low conversion accuracy of the prior art in converting the low-illumination image into a normal-illumination image.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more specifically, to a low-light image enhancement method based on color space transformation. Background Technology

[0002] In low-light conditions, imaging sensors often capture weak light signals with severe noise, resulting in poor visual quality of low-light images. To obtain high-quality images from such poor images, low-light image enhancement (LLIE) is necessary. Its purpose is to improve image brightness while reducing the effects of noise and color deviation. Most existing LLIE methods focus on finding suitable image brightness, typically by employing deep neural networks to learn the mapping relationship between low-brightness and normal-brightness images within the standard RGB (sRGB) color space. However, in these LLIE methods, image brightness exhibits a strong coupling with the colors from the three sRGB channels, i.e., high color sensitivity (mentioned in the book "Color Computer Vision: Fundamentals and Applications" by Theo Gaywells et al.), resulting in significant color distortion in the restored image. Inspired by the Kubelka-Munk theory, some recent methods attempt to convert images from the sRGB color space to the Hue, Saturation, and Value (HSV) color space. These methods help to achieve brightness enhancement more accurately, but amplify local color space noise, thus introducing severe artifacts in the result. Specifically, the conversion from sRGB to HSV disrupts the continuity of red (red discontinuous noise) and black (black plane noise), leading to an increase in the Euclidean distance between similar colors and introducing artifacts in the final image. Both types of noise cause severe artifacts when enhancing images that are predominantly red or extremely dark. Summary of the Invention

[0003] This invention provides a low-light image enhancement method based on color space transformation, which at least solves the technical problem of low accuracy in converting low-light images to normal-light images in existing technologies.

[0004] According to one aspect of the present invention, a low-light image enhancement method based on color space transformation is provided. The method may include: acquiring an initial low-light image; preprocessing the initial low-light image to obtain an intensity map of the initial low-light image; generating a saturation map based on the intensity map of the initial low-light image; generating a hue map based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image; processing the hue map to obtain a horizontal axis set and a vertical axis set of the hue map; setting an adaptive intensity collapse function for the intensity map of the initial low-light image; obtaining a horizontal map and a vertical map based on the adaptive intensity collapse function, the saturation map, the horizontal axis set, and the vertical axis set; concatenating the horizontal map and the vertical map to obtain an HV map; and constructing a CIDNet convolutional neural network, wherein the CIDNet convolutional neural network includes I- The process involves three branches: the I-branch, the HV-branch, and the feature fusion layer. The intensity map of the initial low-light image is input to the I-branch to obtain the first feature map. The HV map and the intensity map of the initial low-light image are input to the HV-branch to obtain the second feature map. The first feature map, the second feature map, the HV map, and the intensity map of the initial low-light image are input to the feature fusion layer to obtain the first low-light image. The first low-light image undergoes an inverse HVI transform to obtain an enhanced map of the initial low-light image. Based on the enhanced map of the initial low-light image and the adaptive intensity collapse function, a first intermediate variable and a second intermediate variable are set. Based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhanced map of the initial low-light image, a normal illumination image of the target is obtained.

[0005] Optionally, generating a saturation map based on the intensity map of the initial low-light image includes: determining the pixel value of each pixel in the saturation map based on the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image; and concatenating the pixel values ​​of each pixel in the saturation map to obtain the saturation map.

[0006] Optionally, generating a tone map based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image includes: obtaining the pixel value of each pixel in the tone map based on the pixel value of each pixel in the saturation map, the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image, and the R, G, and B values ​​of each pixel in the initial low-light image; and concatenating the pixel values ​​of each pixel in the tone map to obtain the tone map.

[0007] Optionally, the process of processing the tone map to obtain the horizontal axis set and the vertical axis set of the tone map includes: polarizing the hue axis of each pixel in the tone map to obtain the horizontal axis and the vertical axis of each pixel in the tone map; concatenating the horizontal axes of each pixel to obtain the horizontal axis set of the tone map; and concatenating the vertical axes of each pixel to obtain the vertical axis set of the tone map.

[0008] Optionally, obtaining the target normal illumination image based on the enhanced image of the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the initial low-light image includes: converting the enhanced image of the initial low-light image into a target HSV image based on the enhanced image of the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the initial low-light image; and performing an inverse transform on the target HSV image to obtain the target normal illumination image.

[0009] The beneficial effects of this invention are:

[0010] (1) Enhance the initial low-light image to obtain the intensity map of the initial low-light image; convert the intensity map of the initial low-light image into a saturation map and a hue map, so that the low-light image can be converted into a normal-light image without being constrained by the light intensity.

[0011] (2) When the low-light image is converted into a normal-light image, and the red color space is discontinuous and the black plane is noisy, the horizontal and vertical images of the color image are obtained by processing the hue axis of each pixel of the color image. The adaptive intensity collapse function of the intensity image of the initial low-light image is introduced to obtain the horizontal and vertical images.

[0012] (3) When a low-light image is converted into a normal-light image, and the chromaticity and brightness in the HVI space are low, a CIDNet convolutional neural network is constructed. The intensity map and HV map of the initial low-light image are input into the CIDNet convolutional neural network to obtain the enhanced map of the initial low-light image. This makes more effective use of the chromaticity and brightness information in the HVI space and reduces global color shift and color noise. Attached Figure Description

[0013] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0014] Figure 1 This is a flowchart of a low-light image enhancement method based on color space transformation according to an embodiment of the present invention;

[0015] Figure 2 This is a schematic diagram of an initial low-light image according to an embodiment of the present invention;

[0016] Figure 3 This is a schematic diagram of the HVI color space based on sRGB transformation according to an embodiment of the present invention;

[0017] Figure 4 This is a schematic diagram of the CIDNet convolutional neural network according to an embodiment of the present invention;

[0018] Figure 5 This is a schematic diagram of a target under normal lighting conditions according to an embodiment of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0020] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] Example 1

[0022] According to an embodiment of the present invention, a low-light image enhancement method based on color space transformation is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system containing at least one set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0023] Figure 1 This is a flowchart of a low-light image enhancement method based on color space transformation according to an embodiment of the present invention, as follows: Figure 1 As shown, the method may include the following steps:

[0024] Step S101: Obtain the initial low-light image, preprocess the initial low-light image to obtain the intensity map of the initial low-light image.

[0025] In the technical solution provided by step S101 of the present invention, Figure 2This is a schematic diagram of an initial low-light image according to an embodiment of the present invention. The pixel value of each pixel in the initial low-light image is processed to obtain the expression for the pixel value of each pixel in the intensity map of the initial low-light image:

[0026]

[0027] in, For each pixel in the initial low-light image, For each pixel of the initial low-light image, the R, G, and B values ​​are: The maximum pixel value for each pixel in the intensity map of the initial low-light image. This is the maximum value among the R, G, and B values ​​for each pixel in the initial low-light image.

[0028] Step S102: Generate a saturation map based on the intensity map of the initial low-light image.

[0029] In the technical solution provided by step S102 of the present invention, Figure 3 This is a schematic diagram of the HVI color space based on sRGB transformation according to an embodiment of the present invention, as shown below. Figure 3 In the image, (a) is the intensity map of the initial low-light image, such as... Figure 3 In the diagram (b), the intensity map of the initial low-light image corresponds to the HSV color space map. Each pixel of the intensity map of the initial low-light image is processed to generate a saturation map.

[0030] Step S103: Generate a tone map based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image.

[0031] In the technical solution provided by step S103 of the present invention, each pixel of the saturation map, each pixel of the intensity map of the initial low-light image, and each pixel of the initial low-light image are processed to generate a tone map.

[0032] Step S104: Process the tone map to obtain the horizontal axis set and the vertical axis set of the tone map.

[0033] In the technical solution provided by step S104 of the present invention, such as Figure 3 As shown, the hue axis of each pixel in the tone map is polarized to obtain the set of horizontal and vertical axes of the tone map.

[0034] Step S105: Set the adaptive intensity collapse function for the intensity map of the initial low-light image.

[0035] In the technical solution provided by step S105 of the present invention, such as Figure 3As shown, for the black plane noise problem, the goal is to fold low-intensity regions while preserving high-intensity regions. However, the optimal degree of collapse varies across different datasets and networks. Therefore, it is crucial to adaptively collapse this region through a learning process. To this end, an adaptive intensity collapse function for the intensity map of the initial low-light image is introduced, resulting in the HVI color space. The expression for the adaptive intensity collapse function of the intensity map of the initial low-light image is:

[0036]

[0037] in, For each pixel of the intensity map of the initial low-light image, there is an adaptive intensity collapse function. For rational numbers, .

[0038] Step S106: Based on the adaptive intensity collapse function, saturation map, horizontal axis set, and vertical axis set, obtain the horizontal map and the vertical map.

[0039] In the technical solution provided by step S106 of the present invention, the adaptive intensity collapse function of each pixel in the intensity map of the initial low-light image, the pixel value of each pixel in the saturation map, and the horizontal axis of each pixel in the hue map are calculated to obtain the expression for the pixel value of each pixel in the horizontal map as follows:

[0040]

[0041] in, Represents element-wise multiplication. This represents the pixel value of each pixel in the horizontal image.

[0042] The horizontal image is obtained by concatenating the pixel values ​​of each pixel in the horizontal image.

[0043] The expression for the pixel value of each pixel in the vertical image is obtained by calculating the adaptive intensity collapse function of the intensity map, the pixel value of each pixel in the saturation map, and the vertical axis of each pixel in the tone map of the initial low-light image:

[0044]

[0045] in, This represents the pixel value of each pixel in the vertical image.

[0046] The vertical image is obtained by concatenating the pixel values ​​of each pixel in the vertical image.

[0047] Step S107: Combine the horizontal and vertical graphs to obtain the HV graph.

[0048] In the technical solution provided by step S107 of the present invention, the horizontal graph and the vertical graph are connected by a channel to obtain the HV graph.

[0049] Step S108: Construct the CIDNet convolutional neural network, which includes an I-branch, an HV-branch, and a feature fusion layer.

[0050] In the technical solution provided by step S108 of the present invention, Figure 4 This is a schematic diagram of the CIDNet convolutional neural network according to an embodiment of the present invention. Figure 4 The first row in the diagram represents the I-branch, the second row represents the HV-branch, and the last plus sign indicates the feature fusion layer.

[0051] Step S109: Input the intensity map of the initial low-light image into the I-branch to obtain the first feature map.

[0052] In the technical solution provided by step S109 of the present invention, such as Figure 4 As shown, the intensity map of the initial low-light image ( Figure 4 The illumination intensity map in the image is input into the I-branch to obtain the first feature map. Figure 4 (I in the middle).

[0053] Step S110: Input the intensity map of the HV map and the initial low-light image into the HV-branch to obtain the second feature map.

[0054] In the technical solution provided by step S110 of the present invention, such as Figure 4 As shown, the HV diagram ( Figure 4 The HV color map and the intensity map of the initial low-light image are input into the HV-branch to obtain the second feature map ( Figure 4 HV in (the text is incomplete and cannot be translated).

[0055] Step S111: Input the first feature map, the second feature map, the HV map and the intensity map of the initial low-light image into the feature fusion layer to obtain the first low-light image.

[0056] In the technical solution provided by step S111 of the present invention, the first feature map, the second feature map, the HV map, and the intensity map of the initial low-light image are input to the feature fusion layer ( Figure 4 The plus sign in the image yields the first low-light image.

[0057] Step S112: Perform an inverse HVI transform on the first low-light image to obtain an enhanced image of the initial low-light image.

[0058] In the technical solution provided by step S112 of the present invention, such as Figure 4As shown, the first low-light image is subjected to inverse HVI transform to obtain an enhanced image of the initial low-light image. Figure 4 (Enhanced image in the image).

[0059] Step S113: Based on the enhanced image of the initial low-light image and the adaptive intensity collapse function, set the first intermediate variable and the second intermediate variable.

[0060] In the technical solution provided in step S113 of the present invention, based on the enhancement map of the initial low-light image and the adaptive intensity collapse function, a first intermediate variable and a second intermediate variable are set, and the expressions for the first intermediate variable and the second intermediate variable are as follows:

[0061]

[0062]

[0063] in, As the first intermediate variable, As the second intermediate variable, The h value for each pixel in the enhancement map of the initial low-light image. The v value for each pixel in the enhancement map of the initial low-light image. .

[0064] Step S114: Based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhanced image of the initial low-light image, obtain the target normal lighting image.

[0065] In the technical solution provided by step S114 of the present invention, the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhanced image of the initial low-light image are processed to obtain the target normal lighting image. Figure 5 This is a schematic diagram of a target under normal lighting conditions according to an embodiment of the present invention.

[0066] The method described in this embodiment will be further described below.

[0067] As an optional embodiment, step S102, generating a saturation map based on the intensity map of the initial low-light image, includes: determining the pixel value of each pixel in the saturation map based on the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image; and concatenating the pixel values ​​of each pixel in the saturation map to obtain the saturation map.

[0068] In this embodiment, the expression for determining the pixel value of each pixel in the saturation map based on the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image is as follows:

[0069]

[0070] in, , For each pixel in the saturation map, This represents the minimum pixel value for each pixel in the intensity map of the initial low-light image.

[0071] As an optional embodiment, step S103, generating a tone map based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image, includes: obtaining the pixel value of each pixel in the tone map based on the pixel value of each pixel in the saturation map, the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image, and the R, G, and B values ​​of each pixel in the initial low-light image; and concatenating the pixel values ​​of each pixel in the tone map to obtain the tone map.

[0072] In this embodiment, based on the pixel value of each pixel in the saturation map, the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image, and the R, G, and B values ​​of each pixel in the initial low-light image, the expression for the pixel value of each pixel in the hue map is obtained as follows:

[0073]

[0074] in, This represents the pixel value of each pixel in the tone map. R value for each pixel in the initial low-light image. The G value is the value of each pixel in the initial low-light image. The B value of each pixel in the initial low-light image. To find the remainder, we need to find the remainder 6.

[0075] As an optional embodiment, step S104, processing the tone map to obtain the horizontal axis set and vertical axis set of the tone map, includes: polarizing the hue axis of each pixel in the tone map to obtain the horizontal axis and vertical axis of each pixel in the tone map; concatenating the horizontal axes of each pixel to obtain the horizontal axis set of the tone map; and concatenating the vertical axes of each pixel to obtain the vertical axis set of the tone map.

[0076] In this embodiment, the hue axis of each pixel in the tone map is... Polar coordinate conversion is performed to obtain the horizontal axis of each pixel in the tone map. and vertical axis ,in, , .

[0077] As an optional embodiment, step S104, obtaining the target normal illumination image based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhancement map of the initial low-light image, includes: converting the enhancement map of the initial low-light image into a target HSV image based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhancement map of the initial low-light image; and performing an inverse transformation on the target HSV image to obtain the target normal illumination image.

[0078] In this embodiment, based on the h value of each pixel in the enhanced map of the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the initial low-light image, the H value of each pixel in the target HSV image is obtained. The expressions for the S and V values ​​are as follows:

[0079]

[0080] in, The H value for each pixel in the target HSV image. The S value for each pixel in the target HSV image. The V value for each pixel in the target HSV image. and These are custom parameters used to adjust the saturation and brightness of the target HSV image. To obtain a remainder of 1, the H, S, and V values ​​of each pixel in the target HSV image are concatenated to obtain the target HSV image. Then, the target HSV image is inversely transformed to obtain the target's normal illumination image, which is... Figure 5 .

[0081] In this embodiment of the invention, an initial low-light image is acquired, preprocessed to obtain an intensity map of the initial low-light image; a saturation map is generated based on the intensity map of the initial low-light image; a hue map is generated based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image; the hue map is processed to obtain a horizontal axis set and a vertical axis set; an adaptive intensity collapse function is set for the intensity map of the initial low-light image; a horizontal map and a vertical map are obtained based on the adaptive intensity collapse function, the saturation map, the horizontal axis set, and the vertical axis set; the horizontal map and the vertical map are concatenated to obtain an HV map; a CIDNet convolutional neural network is constructed, wherein the CIDNet convolutional neural network includes an I-branch, an HV-branch, and a feature fusion layer; the intensity map of the initial low-light image is input into the I-branch. - The HV-branch is used to obtain the first feature map; the HV map and the intensity map of the initial low-light image are input into the HV-branch to obtain the second feature map; the first feature map, the second feature map, the HV map, and the intensity map of the initial low-light image are input into the feature fusion layer to obtain the first low-light image; the first low-light image is subjected to HVI inverse transformation to obtain the enhanced map of the initial low-light image; based on the enhanced map of the initial low-light image and the adaptive intensity collapse function, the first intermediate variable and the second intermediate variable are set; based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhanced map of the initial low-light image, the target normal illumination image is obtained. This solves the technical problem of low conversion accuracy of low-light image to normal illumination image in the prior art, and achieves the technical effect of improving the conversion accuracy of low-light image to normal illumination image.

[0082] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0083] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0084] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0085] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0086] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a first processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0087] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A low-light image enhancement method based on color space transformation, characterized in that, include: Acquire the initial low-light image, preprocess the initial low-light image to obtain the intensity map of the initial low-light image; Generate a saturation map based on the intensity map of the initial low-light image; A tone map is generated based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image. The tone map is processed to obtain the horizontal axis set and the vertical axis set of the tone map; Set an adaptive intensity collapse function for the intensity map of the initial low-light image; the expression for the adaptive intensity collapse function of the intensity map of the initial low-light image is: in, For each pixel of the intensity map of the initial low-light image, there is an adaptive intensity collapse function. For rational numbers, ; Based on the adaptive intensity collapse function, saturation map, horizontal axis set, and vertical axis set, the horizontal and vertical plots are obtained. By stitching together the horizontal and vertical graphs, we obtain the HV graph; Construct a CIDNet convolutional neural network, which includes an I-branch, an HV-branch, and a feature fusion layer; The intensity map of the initial low-light image is input into the I-branch to obtain the first feature map; The intensity maps of the HV map and the initial low-light image are input into the HV-branch to obtain the second feature map; The first feature map, the second feature map, the HV map, and the intensity map of the initial low-light image are input into the feature fusion layer to obtain the first low-light image; The first low-light image is subjected to inverse HVI transform to obtain an enhanced image of the initial low-light image; Based on the enhanced map of the initial low-light image and the adaptive intensity collapse function, a first intermediate variable and a second intermediate variable are set; Based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhanced image of the initial low-light image, the normal lighting image of the target is obtained.

2. The method according to claim 1, characterized in that, The step of generating a saturation map based on the intensity map of the initial low-light image includes: Based on the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image, determine the pixel value of each pixel in the saturation map. The saturation map is obtained by concatenating the pixel values ​​of each pixel in the saturation map.

3. The method according to claim 2, characterized in that, The process of generating a tone map based on the saturation map, the intensity map of the initial low-light image, and the initial low-light image includes: The pixel value of each pixel in the tone map is obtained based on the pixel value of each pixel in the saturation map, the maximum and minimum pixel values ​​of each pixel in the intensity map of the initial low-light image, and the R, G and B values ​​of each pixel in the initial low-light image. The color tone image is obtained by concatenating the pixel values ​​of each pixel in the color tone image.

4. The method according to claim 3, characterized in that, The process of processing the tone map to obtain the horizontal axis set and the vertical axis set of the tone map includes: By polarizing the hue axis of each pixel in the tone map, we can obtain the horizontal and vertical axes of each pixel in the tone map. By stitching together the horizontal axes of each pixel, we obtain the set of horizontal axes of the tone map; By stitching together the vertical axes of each pixel, we obtain the set of vertical axes of the tone map.

5. The method according to claim 4, characterized in that, The enhancement image based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the initial low-light image is used to obtain the target normal lighting image, including: Based on the first intermediate variable, the second intermediate variable, the first parameter, the second parameter, and the enhancement map of the initial low-light image, the enhancement map of the initial low-light image is converted into the target HSV image; The target HSV image is inversely transformed to obtain the target image under normal illumination.

6. A computer system, characterized in that... include: One or more processors, a computer-readable storage medium for storing one or more programs, wherein, when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method of claim 1.

7. A computer-readable storage medium, characterized in that... The device stores computer-executable instructions that, when executed by a processor, are used to implement the method of claim 1.

8. A computer program product, characterized in that... It includes computer-executable instructions that, when executed by a processor, are used to implement the method of claim 1.

Images