Low-illumination image enhancement method and system, storage medium and equipment

Abstract

The invention discloses a low-illumination image enhancement method and system, a storage medium and equipment, and relates to the technical field of image processing, the noise in a low-illumination image is divided into signal-related noise and signal-independent noise according to the relationship between the noise and an image signal through the method, and the noise in the low-illumination image is enhanced through the subdivision. And different types of noise can be processed more sufficiently, so that the problem of de-noising amplification is avoided. In order to avoid interference of noise on Retinex decomposition, different from processing flows of other methods, the method disclosed by the invention proposes a layer-by-layer decomposition type network based on a constructed low-illumination image model for mixed correlation noise separation and a Retinex theory. According to the network, an input image is decomposed into a noise image, a clean image, a reflection component and an illumination component layer by layer, and the accuracy of denoising and Retinex decomposition can be improved by the network.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a low-illumination image enhancement method, system, storage medium and device. Background Art

[0002] Low-light images usually have problems such as low illumination, high noise, and loss of detail information, which will affect the visual effect of the image and subsequent applications. Specifically, low illumination and high noise will interfere with the visual perception of the human eye. After improving the illumination and denoising, the image will be smoother and more natural, which will help improve the user's visual experience, especially in scenes where careful observation of image details is required. Second, in applications such as image analysis, processing or recognition, low illumination and the presence of noise will interfere with the performance of the algorithm. Image enhancement can significantly improve the accuracy and efficiency of these algorithms.

[0003] Among them, supervised low-light image enhancement methods require a large amount of low-light / normal-light image pair data to train the model. It is costly and difficult to obtain such data. Moreover, the quality of the data will affect the training effect and accuracy of the model. In addition, the generalization of the model may also be limited. When processing data that is significantly different from the training data, the results usually have the following problems: 1. Flexibility and local under / overexposure: The histogram represents the pixel intensity distribution of the image through a graph. The histogram equalization-based method achieves the enhancement effect by stretching the dynamic range of the image. However, the histogram equalization-based method lacks flexibility and is difficult to adjust the local part of the image, which will lead to over / underexposure and noise amplification of the enhanced image. The traditional Retinex-based method relies on artificially designed priors and therefore cannot adapt to multiple different scenarios at the same time.

[0004] 2. Noise amplification and color shift: With the development of deep learning, many low-light enhancement methods based on deep learning have emerged in recent years. In addition, many researchers have combined Retinex theory with deep learning methods. These methods usually first perform Retinex decomposition on the input image, process the reflection component and the illumination component separately, and then reconstruct the enhanced image. This type of method has a good enhancement effect, but does not clearly distinguish and deal with the defects in the reflection component, resulting in noise and color shift in the enhanced image.

[0005] 3. Data dependency and dataset construction: Supervised low-light image enhancement methods rely on low-light / normal-light image pairs, which poses certain difficulties in data collection. Usually, it is necessary to collect low-light and normal-light image pairs under different lighting conditions (night, rainy days, indoor environments, etc.). The collection process is not only time-consuming and labor-intensive, but may also be affected by multiple factors such as weather, equipment, and shooting angles, resulting in uneven data quality. Summary of the invention

[0006] In view of the deficiencies in the prior art, an object of the present invention is to provide a low-illumination image enhancement method, system, storage medium and device, aiming to solve the above-mentioned problems recorded in the background technology.

[0007] A first aspect of the present invention is to provide a low-illumination image enhancement method, the method comprising: Acquire a low-light image that has been subjected to denoising, convert a color space corresponding to the low-light image from an RGB color space to an HSV color space, and extract brightness information of the low-light image; According to the brightness information, dividing the low-light image into a dark area and a bright area according to a preset brightness threshold; Calculating the area ratios of the dark area and the bright area in the low-light image respectively, estimating the overall brightness of the low-light image, and outputting the illumination component corresponding to the dark area; The illumination component corresponding to the dark area is adjusted by gamma correction, and the gamma value is calculated according to the overall brightness to achieve adaptive correction and image enhancement of the low-illuminance image and obtain a target image.

[0008] According to one aspect of the above technical solution, the steps of respectively calculating the area ratios of the dark area and the bright area in the low-light image, estimating the overall brightness of the low-light image, and outputting the illumination component corresponding to the dark area include: Acquiring an image area of ​​the low-light image, and identifying a dark area of ​​the dark area and a bright area of ​​the bright area in the low-light image; Calculating area ratios of the dark area and the bright area in the low-illuminance image, respectively, according to the image area of ​​the low-illuminance image, the dark area of ​​the dark area, and the bright area of ​​the bright area; According to the area ratio of the dark area to the bright area in the low-illuminance image, the overall brightness of the low-illuminance image is estimated, and the illumination component corresponding to the dark area is output.

[0009] According to one aspect of the above technical solution, the expression for estimating the overall brightness of the low-light image is: ; In the formula, B is the overall brightness, is the dark area of ​​the dark region, is the bright area of ​​the bright area, θ is the brightness threshold, is the average brightness of the dark area, is the average brightness of the bright area, L is the illumination component corresponding to the dark area, H is the illumination component corresponding to the bright area.

[0010] According to one aspect of the above technical solution, the illumination component corresponding to the dark area is adjusted by using a gamma correction method, and the Gamma value is calculated according to the overall brightness to achieve the step of adaptive correction and image enhancement of the low-illuminance image. The expression for adjusting the illumination component by using a gamma correction method is: ; In the formula, is the compensation coefficient, is the gamma coefficient, and ; Among them, when The smaller the value, the higher the image brightness. The larger the value, the smaller the boost in image brightness.

[0011] According to one aspect of the above technical solution, the method further includes: Based on the preset Retinex model, the independent illumination components obtained by decomposition are reversely reconstructed, and the gamma-corrected target illumination components are multiplied pixel by pixel with the preset reflection components to obtain the enhanced target image.

[0012] According to one aspect of the above technical solution, the target illumination component after gamma correction is multiplied pixel by pixel with the preset reflection component to obtain the expression of the enhanced target image: ; In the formula, is the target image, R is the reflection component, is the target illumination component.

[0013] A second aspect of the present invention is to provide a low-light image enhancement system, which is applied to the method described in the above technical solution, and the system comprises: A color space conversion module is used to obtain a low-light image that has been subjected to denoising, convert the color space corresponding to the low-light image from an RGB color space to an HSV color space, and extract brightness information of the low-light image; A brightness region division module, configured to divide the low-illuminance image into a dark region and a bright region according to the brightness information and a preset brightness threshold; An image parameter calculation module, used to calculate the area ratio of the dark area and the bright area to the low-light image, estimate the overall brightness of the low-light image, and output the illumination component corresponding to the dark area; The image correction and enhancement module is used to adjust the illumination component corresponding to the dark area by adopting a gamma correction method, and calculate the gamma value according to the overall brightness and darkness, so as to realize adaptive correction and image enhancement of the low-illuminance image.

[0014] According to one aspect of the above technical solution, the brightness area division module is specifically used for: Acquiring an image area of ​​the low-light image, and identifying a dark area of ​​the dark area and a bright area of ​​the bright area in the low-light image; Calculating area ratios of the dark area and the bright area in the low-illuminance image, respectively, according to the image area of ​​the low-illuminance image, the dark area of ​​the dark area, and the bright area of ​​the bright area; According to the area ratio of the dark area to the bright area in the low-illuminance image, the overall brightness of the low-illuminance image is estimated, and the illumination component corresponding to the dark area is output.

[0015] A third aspect of the present invention is to provide a readable storage medium having a computer program stored thereon, which implements the method described in the above technical solution when executed by a processor.

[0016] The fourth aspect of the present invention is to provide an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the above technical solution when executing the computer program.

[0017] Compared with the prior art, the low-light image enhancement method, system, storage medium and device shown in the present invention have the following beneficial effects: The method shown in the present invention divides the noise in the low-light image more finely, and divides it into signal-dependent noise and signal-independent noise according to the relationship between the noise and the image signal. Through this subdivision, different types of noise can be more fully processed, thereby avoiding the problem of denoising amplification. And in order to avoid the interference of noise on Retinex decomposition, different from the processing flow of other methods, the method shown in the present invention proposes a layer-by-layer decomposition network based on the constructed low-light image model of mixed correlated noise separation and Retinex theory. The network decomposes the input image layer by layer into a noise image, a clean image, a reflection component and an illumination component, which helps the network improve the accuracy of denoising and Retinex decomposition. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which: Figure 1 is a flow chart of a low-illumination image enhancement method in one embodiment of the present invention; Figure 2 FIG. 4 is a structural block diagram of a low-illumination image enhancement system in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings. Several embodiments of the present invention are shown in the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0020] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0022] Embodiment 1 See also Figure 1 The first embodiment of the present invention provides a low-light image enhancement method, the method comprising steps S10 to S40: Step S10, obtaining a low-light image that has been subjected to denoising, converting a color space corresponding to the low-light image from an RGB color space to an HSV color space, and extracting brightness information of the low-light image.

[0023] First of all, it should be noted that RGB color space and HSV color space are two commonly used color representation methods. RGB color space is based on the three basic colors of red, green, and blue, and superimposed to varying degrees to produce rich and wide colors. In the RGB color space, the value range of red, green, and blue is 0-255. Among them, 0 represents the minimum intensity of the color component, and 255 represents the maximum intensity. The HSV color space uses three parameters: hue, saturation, and value to describe the color. The hue H parameter represents the color information, that is, the position of the spectral color. This parameter is expressed as an angle, and the value range is 0°-360°. The value range of saturation S and brightness V are both 0.0-1.0.

[0024] In this embodiment, the denoised low-light image is converted from the RGB color space to the HSV color space, and the V channel is extracted to obtain the brightness information of the image.

[0025] Step S20: dividing the low-illuminance image into a dark area and a bright area according to the brightness information and a preset brightness threshold.

[0026] In this embodiment, after extracting the V channel, that is, obtaining the brightness information of the image, a brightness threshold is first set. , used to divide the dark area and the bright area, and the brightness value below the threshold The area with brightness higher than the threshold is classified as the dark area. The area is divided into the bright area.

[0027] Step S30 , respectively calculating the area ratios of the dark area and the bright area in the low-illuminance image, estimating the overall brightness of the low-illuminance image, and outputting the illumination component corresponding to the dark area.

[0028] In this embodiment, the steps of respectively calculating the area ratios of the dark area and the bright area in the low-light image, estimating the overall brightness of the low-light image, and outputting the illumination component corresponding to the dark area include: Acquiring an image area of ​​the low-light image, and identifying a dark area of ​​the dark area and a bright area of ​​the bright area in the low-light image; Calculating area ratios of the dark area and the bright area in the low-illuminance image, respectively, according to the image area of ​​the low-illuminance image, the dark area of ​​the dark area, and the bright area of ​​the bright area; According to the area ratio of the dark area to the bright area in the low-illuminance image, the overall brightness of the low-illuminance image is estimated, and the illumination component corresponding to the dark area is output.

[0029] Among them, the proportion of the dark area and the bright area in the whole image is calculated respectively, and the size of the dark area is recorded as , the size of the bright area is . Estimate the overall brightness of the image based on the division of bright and dark areas , the expression is: ; in, is the overall brightness, is the average brightness of the dark area, is the average brightness of the bright area, is the brightness threshold, is the average brightness of the dark area, is the average brightness of the bright area, is the illumination component corresponding to the dark area, is the illumination component corresponding to the bright area.

[0030] Step S40, using a gamma correction method to adjust the illumination component corresponding to the dark area, and calculating a gamma value according to the overall brightness, so as to achieve adaptive correction and image enhancement of the low illumination image and obtain a target image.

[0031] In this embodiment, after denoising and Retinex decomposition by layer-by-layer decomposition network, the illumination component corresponding to the dark area is Make further adjustments to brighten the image.

[0032] Among them, the gamma correction is used to correct the luminance component Adjustment is performed, and the gamma value is calculated based on the overall image brightness obtained based on the regional weighted estimation, so as to realize adaptive correction of image brightness to meet the enhancement of images with different brightness. The expression is: ; In the formula, is the compensation coefficient, which is usually 0. is the gamma coefficient, and .

[0033] Specifically, during gamma correction, the gamma value The size of has a direct impact on the brightness of the image. The smaller the value, the higher the image brightness will be. The larger the value, the lower the image brightness will be. By adjusting the illumination component to normal illumination, the overall brightness of the image can reach a more appropriate level. In this way, the processed image not only removes noise and retains rich details and texture information, but also significantly improves the brightness, providing users with a clearer and more natural visual experience.

[0034] Finally, based on the Retinex model, the independent components obtained by decomposition are reversely reconstructed to obtain the target illumination component after gamma correction. With the reflection component By multiplying pixel by pixel, the enhanced target image can be obtained. , the expression is: ; In the formula, is the target image, is the reflection component, represents element-wise multiplication, is the target illumination component.

[0035] The low-light image enhancement method shown in this embodiment can be applied to sub-fields such as public safety, consumer electronics, medical imaging, and autonomous driving. Specifically: Public safety: At night or in low light, the images captured by surveillance cameras are often difficult to read. By enhancing the images captured by the camera, the clarity and brightness of the images can be improved, which helps to improve the level of security monitoring in public areas.

[0036] Consumer electronics: With the popularity of mobile devices, consumers are increasingly pursuing high-quality images. Low-light image enhancement can improve the shooting effect of mobile phones in dark environments, allowing users to take clear and vivid photos even in low-light conditions, which helps to improve users' photography experience and satisfaction.

[0037] Medical imaging: In endoscopy, due to limited lighting conditions, the images captured often have poor visibility and are difficult to judge the condition. Low-light image enhancement can improve the brightness, contrast and details of these images, allowing doctors to more accurately judge the condition and formulate treatment plans.

[0038] Autonomous driving: At night, in low light and in bad weather, autonomous driving systems rely on cameras and sensors to perceive their surroundings. By enhancing the images captured by the camera, the safety and accuracy of the autonomous driving system can be improved.

[0039] Compared with the prior art, the low-illumination image enhancement method shown in this embodiment has the following beneficial effects: The method shown in this embodiment divides the noise in the low-light image more finely, and divides it into signal-dependent noise and signal-independent noise according to the relationship between the noise and the image signal. Through this subdivision, different types of noise can be more fully processed, thereby avoiding the problem of denoising amplification. And in order to avoid the interference of noise on Retinex decomposition, different from the processing flow of other methods, the method shown in this embodiment proposes a layer-by-layer decomposition network based on the constructed low-light image model of mixed correlated noise separation and Retinex theory. The network decomposes the input image layer by layer into a noise image, a clean image, a reflection component, and an illumination component, which helps the network improve the accuracy of denoising and Retinex decomposition.

[0040] Embodiment 2 The second embodiment of the present invention also provides a low-light image enhancement method. In the method shown in this embodiment: Due to the influence of lighting conditions and the limitations of shooting equipment, low-light images contain a lot of noise, especially in dark areas. Existing methods usually use low-light images as the Considered as a clean image and noise However, low-light images The noise in the image can be divided into signal-dependent noise and signal-independent noise according to its relationship with the signal. To achieve better enhancement effect, the noise is subdivided according to the above classification, and the corresponding low-light image model for mixed correlated noise separation is constructed. The expression is: ; In the formula, is the signal-dependent noise, is signal-independent noise.

[0041] Low-light images often have problems such as insufficient brightness, low contrast, and high noise. Traditional low-light image enhancement methods often find it difficult to simultaneously meet the requirements of denoising, enhancing brightness and contrast, retaining details, and other aspects. The neural network shown in this embodiment is used to learn and extract deep, nonlinear features in the data in a data-driven manner, which has better flexibility and generalization. Therefore, in this embodiment, it is necessary to construct a layer-by-layer decomposition-type low-light image enhancement network model to gradually decompose the input image into a signal-independent noise map, a signal-dependent noise map, a reflection component, and an illumination component. The low-light image enhancement network system shown in this embodiment includes: First layer: low-light image model using mixed correlated noise separation. The first layer network removes signal-independent noise from the input image, i.e., the low-light image. By designing the self-supervised image reconstruction loss and the signal-independent noise consistency loss, the first layer of the layer-by-layer decomposition network is jointly constrained.

[0042] Among them, based on the constructed low-light image model, first pass through the first layer network For the input image, i.e. the low-light image This layer of network consists of 2 parallel convolution branches, each branch contains 5 layers of convolution. The two convolution branches include and , respectively, we get images containing signal-correlated noise Signal-independent noise . This achieves signal-independent noise removal.

[0043] The second layer: Based on the above image model, a self-supervised image consistency loss constraint is designed, and the second layer of the network with the above image reconstruction loss constraint is combined. This layer of network further separates the signal-related noise in the image, thereby obtaining a completely denoised low-light image.

[0044] Among them, the second layer network of the layer-by-layer decomposition network , further decomposes the image containing signal-correlated noise. and The network structure is the same as that of , which is also composed of two parallel convolution branches. and These two convolution branches obtain clean images respectively and signal-dependent noise , thereby removing signal-dependent noise , get a clean image .

[0045] The third layer: Based on the standard Retinex model, the Retinex loss and the reflection component consistency loss are designed, and the Retinex decomposition network of the last layer is jointly constrained to complete the construction of the layer-by-layer decomposition low-light image enhancement network model. This layer of network further decomposes the clean low-light image into the reflection component representing the attributes of the object in the image and the illumination component representing the ambient illumination information.

[0046] Among them, the Retinex model is implemented based on the Retinex theory. Retinex is a compound word composed of the two words retina and cortex. According to the Retinex theory, the image depends on the incident light and the reflection of the incident light by the surface of the object. Therefore, low-light images The reflection component (reflectance) and illumination components (Illumination) is composed of: ; In the formula, the reflection component Represents the physical properties of objects in the image, the illumination component Indicates the light intensity of the shooting environment. Represents element-wise multiplication.

[0047] Specifically, the Retinex theory has good physical interpretability and is therefore widely used in various low-light image enhancement methods. After the first and second layers of denoising, a clean image can be obtained. Then, based on the standard Retinex model, For clean images Perform Retinex decomposition. Like the two sub-networks in the first and second layers, it consists of two parallel convolution branches. and The two convolution branches obtain the reflection components respectively and illumination component .

[0048] In addition, supervised methods rely on low / normal illumination image pairs, and there are certain difficulties in data collection. Unsupervised low-illumination enhancement methods do not rely on paired image pairs of low / normal illumination, but the information that can be obtained from a single image is limited. Therefore, in this embodiment, a low-illumination image pair dataset is used as a training dataset. The dataset contains 324 scenes, each scene has 2 to 6 low-illumination images of different illuminations, for a total of 1,000 images. When training the above-mentioned layer-by-layer decomposition low-illumination image enhancement network model, the input of the network is a low-illumination image pair, that is, two low-illumination images with different image brightness in the same scene. and For scenes containing more than two images, images with different illuminations are paired and used as the input of the model.

[0049] In this embodiment, the layer-by-layer decomposition network is constrained by self-supervised image reconstruction loss, signal-independent noise consistency loss, image consistency loss, reflection component consistency loss and Retinex loss. Specifically, the self-supervised image reconstruction loss and signal-independent noise consistency loss are designed to jointly constrain the first layer of the layer-by-layer decomposition network; the self-supervised image consistency loss is designed to jointly constrain the second layer of the layer-by-layer decomposition network with the image reconstruction loss; the Retinex loss and reflection component consistency loss are designed to jointly constrain the Retinex decomposition network of the last layer. Among them: a. Image reconstruction loss, expressed as: ; According to the constructed image model, the three components obtained by decomposition , and , the input image should be reconstructed , i.e. low-light image In addition, according to the image model, it is also possible to input the image , clean image Signal-independent noise Calculate the signal-dependent noise .

[0050] The second item is to make the second layer The decomposition of is more in line with the proposed low-light image model for mixed correlated noise separation, that is, it is mainly aimed at At this point, it can be considered that The first term of the image reconstruction loss and the constraints of the following image consistency loss have been accurately decomposed. The main function of the second term of the image reconstruction loss is to further accurately decompose .therefore, Some of the relevant network parameters should not be updated with the calculation of the second term of the reconstruction loss, so stopgrad is used to decompose That part of the network ( ) does not participate in the gradient calculation, so as to avoid affecting the final decomposition effect and make the decomposition result more accurate.

[0051] b. Signal-independent noise consistency loss, expressed as: ; Among them, signal-independent noise is irrelevant to the image signal and is mainly generated when the camera scans and processes the image during shooting. Therefore, the signal-independent noise in the image pair of the same scene shot by the same camera should be consistent.

[0052] c. Image consistency loss, expressed as: ; The image content should not change before and after denoising. To ensure the consistency of the image content before and after denoising, the clean image To constrain.

[0053] d. Consistency loss of reflection component, expressed as: ; The input image is a pair of low-light images with the same scene content and different illumination. Therefore, the decomposed reflection component , It represents the physical characteristics of the same object, and its signal strength and gradient should remain consistent.

[0054] e.Retinex loss, the expression is: ; Among them, according to the Retinex theory, the decomposition obtained and The clean image of the input can be reconstructed . And, from the clean image and It can also be calculated To ensure the accuracy of Retinex decomposition, stopgrad is also used here. Illumination map is the maximum value of the three channels of the input image, the illumination component In addition, the illumination component It represents the illumination information of the environment and does not contain the texture features of the object, so it should be as smooth as possible.

[0055] Embodiment 3 See also Figure 2 The third embodiment of the present invention provides a low-light image enhancement system, which is applied to the method described in the above embodiment, and the system includes: The color space conversion module 10 is used to obtain a low-light image that has been subjected to denoising, convert the color space corresponding to the low-light image from the RGB color space to the HSV color space, and extract the brightness information of the low-light image; A brightness region division module 20, configured to divide the low-light image into a dark region and a bright region according to the brightness information and a preset brightness threshold; An image parameter calculation module 30, for respectively calculating the area ratios of the dark area and the bright area in the low-illuminance image, estimating the overall brightness of the low-illuminance image, and outputting the illumination component corresponding to the dark area; The image correction and enhancement module 40 is used to adjust the illumination component corresponding to the dark area by adopting a gamma correction method, and calculate the gamma value according to the overall brightness to achieve adaptive correction and image enhancement of the low-illuminance image.

[0056] In this embodiment, the brightness area division module is specifically used for: Acquiring an image area of ​​the low-light image, and identifying a dark area of ​​the dark area and a bright area of ​​the bright area in the low-light image; Calculating area ratios of the dark area and the bright area in the low-illuminance image, respectively, according to the image area of ​​the low-illuminance image, the dark area of ​​the dark area, and the bright area of ​​the bright area; According to the area ratio of the dark area to the bright area in the low-illuminance image, the overall brightness of the low-illuminance image is estimated, and the illumination component corresponding to the dark area is output.

[0057] Compared with the prior art, the low-light image enhancement system shown in this embodiment has the following beneficial effects: The system shown in this embodiment divides the noise in the low-light image more finely, and divides it into signal-dependent noise and signal-independent noise according to the relationship between the noise and the image signal. Through this subdivision, different types of noise can be more fully processed, thereby avoiding the problem of denoising amplification. And in order to avoid the interference of noise on Retinex decomposition, different from the processing flow of other methods, the system shown in this embodiment proposes a layer-by-layer decomposition network based on the constructed low-light image model of mixed correlated noise separation and Retinex theory. The network decomposes the input image layer by layer into a noise image, a clean image, a reflection component, and an illumination component, which helps the network improve the accuracy of denoising and Retinex decomposition.

[0058] Embodiment 4 A fourth embodiment of the present invention provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method described in the above embodiment is implemented.

[0059] Embodiment 5 A fifth embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the above embodiments when executing the computer program.

[0060] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0061] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A low-light image enhancement method, characterized in that: The method comprises: Acquire a low-light image that has been subjected to denoising, convert a color space corresponding to the low-light image from an RGB color space to an HSV color space, and extract brightness information of the low-light image; According to the brightness information, dividing the low-light image into a dark area and a bright area according to a preset brightness threshold; Respectively calculating the area ratios of the dark area and the bright area in the low-light image, estimating the overall brightness of the low-light image, and outputting the illumination component corresponding to the dark area; The illumination component corresponding to the dark area is adjusted by gamma correction, and the gamma value is calculated according to the overall brightness to achieve adaptive correction and image enhancement of the low-illuminance image and obtain a target image.

2. The low-light image enhancement method according to claim 1, characterized in that: The steps of respectively calculating the area ratios of the dark area and the bright area to the low-light image, estimating the overall brightness of the low-light image, and outputting the illumination component corresponding to the dark area include: Acquiring an image area of ​​the low-light image, and identifying a dark area of ​​the dark area and a bright area of ​​the bright area in the low-light image; Calculating area ratios of the dark area and the bright area in the low-illuminance image, respectively, according to the image area of ​​the low-illuminance image, the dark area of ​​the dark area, and the bright area of ​​the bright area; According to the area ratio of the dark area to the bright area in the low-illuminance image, the overall brightness of the low-illuminance image is estimated, and the illumination component corresponding to the dark area is output.

3. The low-light image enhancement method according to claim 2, characterized in that: The expression for estimating the overall brightness of the low-light image is: ； In the formula, B is the overall brightness, is the dark area of ​​the dark region, is the bright area of ​​the bright area, θ is the brightness threshold, is the average brightness of the dark area, is the average brightness of the bright area, L is the illumination component corresponding to the dark area, H is the illumination component corresponding to the bright area.

4. The low-light image enhancement method according to claim 3, characterized in that: In the step of adjusting the illumination component corresponding to the dark area by adopting a gamma correction method, and calculating the Gamma value according to the overall brightness and darkness to realize the adaptive correction and image enhancement of the low illumination image, the expression for adjusting the illumination component by adopting a gamma correction method is: ； In the formula, is the compensation coefficient, is the gamma coefficient, and ; Among them, when The smaller the value, the higher the image brightness. The larger the value, the smaller the boost in image brightness.

5. The low-light image enhancement method according to claim 1, characterized in that: The method further comprises: Based on the preset Retinex model, the independent illumination components obtained by decomposition are reversely reconstructed, and the gamma-corrected target illumination components are multiplied pixel by pixel with the preset reflection components to obtain the enhanced target image.

6. The low-light image enhancement method according to claim 5, characterized in that: The gamma-corrected target illumination component is multiplied pixel by pixel with the preset reflection component to obtain the expression of the enhanced target image: ； In the formula, is the target image, R is the reflection component, is the target illumination component.

7. A low-light image enhancement system, characterized in that: The method applied to any one of claims 1 to 6, wherein the system comprises: A color space conversion module is used to obtain a low-light image that has been subjected to denoising, convert the color space corresponding to the low-light image from an RGB color space to an HSV color space, and extract brightness information of the low-light image; A brightness region division module, configured to divide the low-illuminance image into a dark region and a bright region according to the brightness information and a preset brightness threshold; An image parameter calculation module, used to calculate the area ratio of the dark area and the bright area to the low-light image, estimate the overall brightness of the low-light image, and output the illumination component corresponding to the dark area; The image correction and enhancement module is used to adjust the illumination component corresponding to the dark area by adopting a gamma correction method, and calculate the gamma value according to the overall brightness to achieve adaptive correction and image enhancement of the low-illuminance image.

8. The low-light image enhancement system according to claim 7, characterized in that: The brightness area division module is specifically used for: Acquiring an image area of ​​the low-light image, and identifying a dark area of ​​the dark area and a bright area of ​​the bright area in the low-light image; Calculating area ratios of the dark area and the bright area in the low-illuminance image, respectively, according to the image area of ​​the low-illuminance image, the dark area of ​​the dark area, and the bright area of ​​the bright area; According to the area ratio of the dark area to the bright area in the low-illuminance image, the overall brightness of the low-illuminance image is estimated, and the illumination component corresponding to the dark area is output.

9. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. 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 computer program, the method according to any one of claims 1 to 6 is implemented.

Images