An underwater non-uniformly illuminated image enhancement method, system, device and medium

Abstract

The present application relates to the technical field of computer vision, and particularly relates to an underwater non-uniform illumination image enhancement method, system, device and medium, comprising: correcting an obtained luminance channel standard image to obtain a luminance correction image by using an adaptive gamma correction algorithm and an illumination distribution map; synthesizing a color image according to the luminance correction image to obtain UV channel values of each pixel point of the color image in a YUV color space; determining a Y channel white point position in the YUV color space according to the UV channel values of all pixel points and a preset UV highlight threshold; performing white balance adjustment on each channel of the color image at the Y channel white point position to obtain an enhanced image; and performing wavelet denoising processing on the enhanced image to obtain a denoised image. The method effectively improves the quality of the underwater non-uniform illumination image, can be applied to various complex underwater scenes, and improves the reliability and stability of the enhancement effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer vision, and in particular to an underwater uneven illumination image enhancement method, system, device and medium. BACKGROUND

[0002] In the underwater shallow water environment, the light intensity and color will change due to the change of different depths, resulting in the phenomenon of uneven illumination of objects in the image, especially at a depth of several tens of meters from the water surface, due to the rapid attenuation of underwater light, there is almost no environmental lighting, which means that most optical imaging in the deep sea is completed through artificial lighting, therefore, optical imaging in the deep sea is usually accompanied by uneven lighting, which causes great difficulty in recognizing and detecting target objects, and therefore needs to be enhanced.

[0003] The traditional underwater uneven illumination image enhancement method mainly adopts image preprocessing and image enhancement algorithm. In the aspect of image preprocessing, the commonly used preprocessing methods include white balance correction, color correction and mean filtering, etc. In the aspect of image enhancement algorithm, the commonly used methods include histogram equalization, limited contrast adaptive histogram equalization (CLAHE), wavelet transform, etc. Although these methods can effectively reduce the noise and fuzziness in the underwater image, enhance the contrast and clarity of the target object, and improve the visibility and quality of the image, the complexity of the underwater environment makes it difficult to predict the distribution of underwater illumination, and the current underwater uneven illumination image enhancement method is difficult to handle different types of underwater environment, resulting in poor reliability and stability of the enhancement effect, and certain limitations in the naturalness of the enhancement effect. SUMMARY

[0004] The present application provides an underwater uneven illumination image enhancement method, system, device and medium, which solves the technical problem that the traditional underwater uneven illumination image enhancement method has poor reliability and stability of the enhancement effect when dealing with relatively complex underwater environment, and has certain limitations in the naturalness of the enhancement effect.

[0005] To solve the above technical problems, the present application provides an underwater uneven illumination image enhancement method, system, device and medium.

[0006] In a first aspect, the present application provides an underwater uneven illumination image enhancement method, which comprises the following steps:

[0007] According to the acquired original illumination image, an illumination distribution map and a brightness channel standard image are obtained;

[0008] The brightness channel standard image is corrected by using an adaptive gamma correction algorithm and the illumination distribution map to obtain a brightness corrected image.

[0009] According to the luminance correction image, a color image is acquired, the color image is converted to a YUV color space, and UV channel values at each pixel point in the YUV color space are acquired;

[0010] According to the UV channel values at all pixel points and a preset UV highlight threshold, a Y channel white point position in the YUV color space is determined;

[0011] Each channel of the color image at the Y channel white point position is subjected to white balance adjustment, and an enhanced image is obtained;

[0012] The enhanced image is subjected to wavelet denoising processing, and a denoised image is obtained.

[0013] In further embodiments, the step of acquiring the light distribution map and the luminance channel standard image according to the acquired original light image comprises:

[0014] An original light image is acquired;

[0015] The original light image is subjected to guided filtering, and a light distribution map is obtained;

[0016] The original light image is subjected to color space conversion and channel separation, and a single-channel image is obtained; the single-channel image comprises a hue single-channel image, a saturation single-channel image, and a luminance single-channel image;

[0017] The luminance single-channel image is subjected to standardization processing, and a luminance channel standard image is obtained.

[0018] In further embodiments, the calculation formula of the adaptive gamma correction algorithm is:

[0019] I out (x,y)=I(x,y) Y(x,y) ×255

[0020] wherein,

[0021]

[0022]

[0023]

[0024] In the formula, I out(x, y) represents a luminance correction image; I(x, y) represents a luminance channel standard image; Y(x, y) represents an adaptive gamma parameter; J(x, y) represents an illumination distribution map; β(x, y) represents a luminance image dark area threshold; α(x, y) represents a luminance channel standard image; V(x, y) represents a luminance value of the luminance channel standard image at (x, y); S(x, y) represents a saturation value of the luminance channel standard image at (x, y); represents an average value of luminance of the luminance channel standard image; represents an average value of saturation of the luminance channel standard image; θ(x, y) represents a gray value of the luminance channel standard image at (x, y); δ represents a median value of the gray value of the luminance channel standard image.

[0025] In a further embodiment, the step of determining the Y channel white point position in the YUV color space according to the UV channel values of all pixel points and a preset UV highlight threshold comprises:

[0026] determining the UV highlight threshold according to the UV channel values of all pixel points, wherein the UV highlight threshold comprises a U channel highlight threshold and a V channel highlight threshold;

[0027] screening out highlight pixel points according to the UV channel values of each pixel point and the UV highlight threshold;

[0028] counting a histogram of the highlight pixel points in the Y channel, and determining a highlight average value according to pixel values of highlight pixel points with a preset proportion in the histogram;

[0029] determining the Y channel white point position in the YUV color space according to the highlight average value.

[0030] In a further embodiment, the step of screening out highlight pixel points according to the UV channel values of each pixel point and the UV highlight threshold comprises:

[0031] traversing all pixel points in the YUV color space, and if a difference between a U channel value of a current pixel point and the U channel highlight threshold is not less than a first preset threshold, or a difference between a V channel value of the current pixel point and the V channel highlight threshold is not less than a second preset threshold, then the current pixel point is divided into a highlight pixel point;

[0032] wherein the U channel highlight threshold is determined according to a U channel mean value and a preset U channel dynamic threshold, and the V channel highlight threshold is determined according to a V channel mean value and a preset V channel dynamic threshold.

[0033] In a further embodiment, the calculation formula for performing white balance adjustment on each channel of the color image at the Y channel white point position to obtain an enhanced image is:

[0034]

[0035]

[0036]

[0037] In the formula, R(x, y) represents the R channel pixel value of the enhanced image; G(x, y) represents the G channel pixel value of the enhanced image; B(x, y) represents the B channel pixel value of the enhanced image; r(x, y) represents the R channel pixel value of the original light image; r aver represents the R channel pixel average value of the original light image; g(x, y) represents the G channel pixel value of the original light image; g aver represents the G channel pixel average value of the original light image; b(x, y) represents the B channel pixel value of the original light image; b aver represents the B channel pixel average value of the original light image; y max represents the Y channel pixel maximum value under the YUV color space.

[0038] In a further embodiment, the step of performing wavelet denoising processing on the enhanced image to obtain a denoised image comprises:

[0039] Converting the enhanced image into a YCrCb color space to obtain a Y luminance channel image;

[0040] Performing wavelet transform on the Y luminance channel image to obtain wavelet coefficients;

[0041] Setting a wavelet threshold, and performing wavelet inverse transform on the Y luminance channel image by using the wavelet threshold and the wavelet coefficients to obtain a denoised Y luminance channel image;

[0042] Merging the denoised Y luminance channel image with a hue channel and a saturation channel of the enhanced image to obtain a merged image;

[0043] Converting the merged image from the YCrCb color space into an RGB space to obtain a denoised image.

[0044] In a second aspect, the present application provides an underwater non-uniform light image enhancement system, which comprises:

[0045] An image acquisition module, configured to acquire a light distribution map and a luminance channel standard image according to an acquired original light image;

[0046] An image correction module, configured to correct the luminance channel standard image by using an adaptive gamma correction algorithm and the light distribution map to obtain a luminance correction image;

[0047] a space conversion module, configured to obtain a color image according to the brightness correction image, convert the color image to a YUV color space, and obtain UV channel values at each pixel point in the YUV color space, and determine a Y channel white point position in the YUV color space according to the UV channel values at all pixel points and a preset UV highlight threshold value;

[0048] an image enhancement module, configured to perform white balance adjustment on each channel of the color image at the Y channel white point position, to obtain an enhanced image;

[0049] an image denoising module, configured to perform wavelet denoising processing on the enhanced image, to obtain a denoised image.

[0050] In a third aspect, the present application further provides a computer device, comprising a processor and a memory, the processor being connected with the memory, the memory being used for storing a computer program, and the processor being used for executing the computer program stored in the memory, so that the computer device performs the steps of the above method.

[0051] In a fourth aspect, the present application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above method.

[0052] The present application provides an underwater non-uniform illumination image enhancement method, system, device and medium. The method corrects a brightness channel standard image through an adaptive gamma correction algorithm and an illumination distribution map, obtains UV channel values at each pixel point of the brightness correction image in a YUV color space, determines a Y channel white point position in the YUV color space according to the UV channel values at all pixel points and a preset UV highlight threshold value, performs white balance adjustment on each channel of a color image at the Y channel white point position, to obtain an enhanced image, and performs wavelet denoising processing on the enhanced image, to obtain a denoised image. Compared with a traditional method, the method effectively solves the color deviation problem of an underwater image, removes noise of the underwater image, improves the definition of a low-quality underwater shooting image, and achieves better visual effect. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 is a flowchart of an underwater non-uniform illumination image enhancement method provided by an embodiment of the present application;

[0054] Figure 2 is a schematic diagram of an underwater non-uniform illumination image enhancement process provided by an embodiment of the present application;

[0055] Figure 3 is a schematic diagram of an original illumination image provided by an embodiment of the present application;

[0056] Figure 4 is a schematic diagram of a denoised image provided by an embodiment of the present application;

[0057] Figure 5 is a block diagram of an underwater non-uniform illumination image enhancement system provided by an embodiment of the present application;

[0058] Figure 6 is a structural schematic diagram of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0059] The embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments are presented only for the purpose of illustration and should not be understood as limiting the present application. The accompanying drawings are only for reference and illustration and do not constitute a limitation on the scope of patent protection of the present application, because many changes can be made to the present application without departing from the spirit and scope thereof.

[0060] Reference Figure 1 , an underwater non-uniform illumination image enhancement method is provided by an embodiment of the present application, as shown in Figure 1 The method comprises the following steps:

[0061] S1. Obtain an illumination distribution map and a luminance channel standard image according to an acquired original illumination image.

[0062] In the embodiment, the step of obtaining an illumination distribution map and a luminance channel standard image according to an acquired original illumination image comprises:

[0063] Obtain an original illumination image;

[0064] Perform guided filtering on the original illumination image to obtain an illumination distribution map;

[0065] Perform color space conversion and channel separation on the original illumination image to obtain a single-channel image; the single-channel image comprises a hue single-channel image, a saturation single-channel image and a luminance single-channel image;

[0066] Perform standardization processing on the luminance single-channel image to obtain a luminance channel standard image.

[0067] Specifically, after obtaining the original illumination image, the embodiment performs guided filtering on the obtained original illumination image to obtain an illumination distribution map, performs Gaussian filtering on the gray image of the illumination distribution map as a guide image to obtain a blurred guide image, and calculates the mean vector of the blurred guide image and the original image. In the embodiment, the filtering radius is preferably set to 15, and the formula for calculating the mean vector of the blurred guide image and the original image is as follows:

[0068]

[0069]

[0070] wherein μ A denotes the mean vector of the original illumination image; ω denotes the domain of the convolution kernel; p denotes a pixel point in the original illumination image; A(p) denotes the original illumination image; μ G denotes the mean vector of the blurred guide image; G(p) denotes the blurred guide image.

[0071] After the mean vector of the blurred guide image and the original illumination image is calculated, the covariance matrix of the blurred guide image and the original illumination image in the corresponding domain of each pixel point is calculated, and the specific calculation formula is as follows:

[0072]

[0073] wherein C AG denotes the covariance matrix of the blurred guide image and the original illumination image.

[0074] After the covariance matrix of the blurred guide image and the original illumination image is calculated, the covariance matrix of the blurred guide image in the corresponding domain of each pixel point is calculated, and the specific calculation formula is as follows:

[0075]

[0076] wherein C G denotes the covariance matrix of the blurred guide image.

[0077] Then, the mean coefficient of the guide filter is calculated according to the covariance matrix of the blurred guide image and the original illumination image and the covariance matrix of the blurred guide image, the blurred guide image is taken as the guide image, the original illumination image is filtered by using the mean coefficient of the guide filter, and a filtered image, i.e., an illumination distribution map, is obtained. The calculation formula of filtering the original illumination image by using the mean coefficient of the guide filter is as follows:

[0078]

[0079] wherein,

[0080]

[0081] b = μ A -a x μ G

[0082] wherein a denotes the mean coefficient of the guide filter; b denotes the bias coefficient of the guide filter; ε denotes a very small positive number; and J(p) denotes the illumination distribution map.

[0083] Meanwhile, the embodiment carries out color segmentation and channel separation of the original light image in HSV color space, extracts three single-channel images, i.e. hue single-channel image, saturation single-channel image and luminance single-channel image, and carries out standardization processing on the luminance single-channel image to obtain a luminance channel standard image, and the calculation formula of the standardization processing is:

[0084]

[0085] In the formula, I(x, y) represents the luminance channel standard image; I in represents the luminance single-channel image; I min represents the minimum value of the pixel value in the luminance single-channel image; I max represents the maximum value of the pixel value in the luminance single-channel image.

[0086] The light distribution map obtained by the guided filtering method can well preserve the light details of the image, and provides a more accurate calculation model for the subsequent adjustment of the non-uniform light image.

[0087] S2. Correcting the luminance channel standard image by using an adaptive gamma correction algorithm and the light distribution map to obtain a luminance correction image, and in the embodiment, the calculation formula of the adaptive gamma correction algorithm is:

[0088] I out (x, y) = I(x, y) Y(x,y) × 255

[0089] In the formula,

[0090]

[0091]

[0092]

[0093] In the formula, I out(x, y) represents a luminance correction image; I(x, y) represents a luminance channel standard image; Y(x, y) represents an adaptive gamma parameter for controlling the increase or decrease of luminance, when Y(x, y) > 1, the luminance of the luminance channel standard image is increased, when Y(x, y) < 1, the luminance of the luminance channel standard image is decreased; J(x, y) represents an illumination distribution map; β(x, y) represents a luminance image dark area threshold, considering that when the luminance and saturation of the image are at a lower value, the image is in a darker area, therefore, the luminance image dark area threshold in the embodiment is determined by the luminance and saturation of the image; a(x, y) represents a luminance channel standard image; V(x, y) represents the luminance value of the luminance channel standard image at (x, y); S(x, y) represents the saturation value of the luminance channel standard image at (x, y); represents the average value of the luminance of the luminance channel standard image; represents the average value of the saturation of the luminance channel standard image; θ(x, y) represents the gray value of the luminance channel standard image at (x, y); δ represents the median value of the gray value of the luminance channel standard image.

[0094] Image gamma correction is an image processing technology for adjusting image brightness and contrast, the principle of which is to adjust the brightness and contrast of the image by performing nonlinear transformation on the image, so as to achieve the purpose of image enhancement. The embodiment performs adaptive gamma correction on the luminance channel standard image based on guided filtering, which can fully utilize the illumination component of the image to adaptively change the parameters of the gamma function, so that the luminance value of the high-luminance area in the image is reduced, the luminance value of the low-luminance area is increased, the color contrast of the bright-dark transition area of the original illumination image is enhanced, more details in the dark area are highlighted, and better visual effect is achieved.

[0095] S3. Obtain a color image according to the luminance correction image, convert the color image to YUV color space, and obtain the UV channel values at each pixel point in the YUV color space.

[0096] S4. Determine the Y channel white point position in the YUV color space according to the UV channel values at all pixel points and a preset UV highlight threshold.

[0097] Since the adaptive gamma correction algorithm can only improve the problem of uneven illumination of the image, but cannot effectively solve the color cast problem of the underwater image, therefore, the embodiment adopts a dynamic threshold white balance method to adjust the color temperature and color balance of the image, solves the color cast problem of the underwater image, and makes the image color more rich, and the specific process includes:

[0098] In this embodiment, the brightness-corrected image is recombined with the hue single-channel image and saturation single-channel image extracted in step S1 to form a color image. The synthesized color image is then converted to the YUV color space. The UV bright spot threshold is then determined based on the UV channel values ​​at all pixels. This UV bright spot threshold includes a U-channel bright spot threshold and a V-channel bright spot threshold. In this embodiment, the U-channel bright spot threshold is determined based on the U-channel mean and a preset U-channel dynamic threshold, and the V-channel bright spot threshold is determined based on the V-channel mean and a preset V-channel dynamic threshold. The specific process for obtaining the U-channel bright spot threshold and the V-channel bright spot threshold is as follows:

[0099] Calculate the mean of the U channel and the mean of the V channel based on the U channel and V channel values ​​of each pixel in the YUV color space.

[0100] The sum of the U-channel mean and the preset U-channel dynamic threshold is used as the U-channel highlight threshold;

[0101] The sum of the V channel mean and the preset V channel dynamic threshold is used as the V channel highlight threshold.

[0102] The calculation formulas for the U-channel dynamic threshold and the V-channel dynamic threshold are as follows:

[0103]

[0104]

[0105] In the formula, Indicates the dynamic threshold of the U channel; δ u (x,y) represents the U channel value of the color image at (x,y) in the YCbCr color space; M u This represents the mean value of the U channel in the YCbCr color space of the color image; N represents the total number of pixels in the color image. Indicates the dynamic threshold of the V channel; δ v (x,y) represents the V channel value of the color image at (x,y) in the YCbCr color space; M v This represents the mean value of the V channel in a color image within the YCbCr color space.

[0106] In this embodiment, after obtaining the U-channel bright spot threshold and the V-channel bright spot threshold, bright pixels are selected based on the UV channel values ​​at each pixel and the UV bright spot threshold. The specific process is as follows:

[0107] If a difference between a U channel value of a current pixel point and a U channel highlight point threshold is not less than a first preset threshold, or a difference between a V channel value of the current pixel point and a V channel highlight point threshold is not less than a second preset threshold, the current pixel point is divided into a highlight pixel point. In the embodiment, the first preset threshold is preferably set as The second preset threshold is preferably set as

[0108] After the highlight pixel points belonging to the near-white region are screened according to the above rules, a histogram of the highlight pixel points in the Y channel is counted, a highlight average value is determined according to pixel values of highlight pixel points with a preset proportion in the histogram, and a Y channel white point position in the YUV color space is determined according to the highlight average value. It should be noted that the Y channel white point position is preferably determined according to an average value of the first 1% of highlights in the histogram.

[0109] S5. The white balance of each channel of the color image at the Y channel white point position is adjusted to obtain an enhanced image.

[0110] The B, G and R channels of the color image at the Y channel white point position are respectively adjusted for white balance in the embodiment, the RGB channels are scaled so that the white point becomes the maximum brightness value, and then the RGB three channels are combined to obtain an enhanced image after dynamic threshold white balance processing. The calculation formula of the enhanced image obtained by white balance adjustment is as follows:

[0111]

[0112]

[0113]

[0114] In the formula, R(x, y) represents the R channel pixel value of the enhanced image; G(x, y) represents the G channel pixel value of the enhanced image; B(x, y) represents the B channel pixel value of the enhanced image; r(x, y) represents the R channel pixel value of the original illumination image; r aver represents the R channel pixel average value of the original illumination image; g(x, y) represents the G channel pixel value of the original illumination image; g aver represents the G channel pixel average value of the original illumination image; b(x, y) represents the B channel pixel value of the original illumination image; b aver represents the B channel pixel average value of the original illumination image; y max represents the maximum value of the Y channel pixel in the YUV color space.

[0115] S6. The enhanced image is subjected to wavelet denoising processing to obtain a denoised image.

[0116] In the embodiment, the step of performing wavelet denoising processing on the enhanced image to obtain a denoised image comprises:

[0117] Converting the enhanced image into YCrCb color space to obtain a Y luminance channel image;

[0118] Performing wavelet transform on the Y luminance channel image to obtain wavelet coefficients;

[0119] Setting a wavelet threshold, and performing wavelet inverse transform on the Y luminance channel image by using the wavelet threshold and the wavelet coefficients to obtain a denoised Y luminance channel image;

[0120] Merging the denoised Y luminance channel image with the hue channel and the saturation channel of the enhanced image to obtain a merged image;

[0121] Converting the merged image from YCrCb color space to RGB space to obtain a denoised image.

[0122] Although the uneven illumination and color cast problem of the image has been well improved through the above steps S1-S5, due to the influence of underwater water quality, turbidity and other factors, underwater images generally have noise, which has an adverse effect on the quality and clarity of underwater images. Therefore, the embodiment adopts wavelet denoising processing to effectively remove the noise in underwater images, and the specific process is as follows:

[0123] Converting the enhanced image into YCrCb color space to obtain a Y luminance channel image, and only performing wavelet transform and threshold processing on the Y luminance channel image, i.e., performing wavelet transform on the Y luminance channel image to obtain wavelet coefficients, and the formula of the wavelet transform is as follows:

[0124]

[0125] In the formula, W(e, λ) represents the wavelet coefficient; f(t) represents the Y luminance channel image; Ψ e,λ (t) represents the wavelet function; e represents the scale of the wavelet transform; and λ represents the translation parameter of the wavelet transform.

[0126] Performing threshold processing on the wavelet coefficients, and setting the high-frequency components lower than a certain threshold to zero, and the formula for calculating the wavelet threshold is as follows:

[0127]

[0128] In the formula, T thresh represents the wavelet threshold, σ represents the standard deviation of the noise, and n is the number of wavelet coefficients.

[0129] The wavelet inverse transform is performed on the wavelet coefficients after threshold processing, to obtain a Y luminance channel image after denoising, the Y luminance channel image after denoising and the chrominance channels of the enhanced image are combined, and the combined image is converted to an RGB space, to obtain a final denoised image.

[0130] The embodiment of the present application provides a kind of underwater uneven illumination image enhancement method, system, equipment and medium, the method utilizes guide filter to obtain illumination distribution chart, while the luminance single channel image extracted is carried out standardization processing, then using adaptive gamma correction algorithm and illumination distribution chart, luminance single channel image is adjusted to illumination, so as to highlight more detail information in dark area, reach better visual effect;The method is carried out dynamic threshold white balance and wavelet function denoising to the image obtained, to solve the color cast problem of underwater image, realize the removal of underwater image noise.Compared with prior art, the embodiment can be adaptively corrected according to image and the image is enhanced and denoised, and the detail information of underwater image after enhancement and denoising is more clear and complete, improve the visual quality of underwater uneven image.

[0131] It should be noted that the size of the serial number of the above processes does not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0132] In one embodiment, as shown in Figure 5 The embodiment of the present application provides an underwater uneven illumination image enhancement system, which comprises:

[0133] The image acquisition module 101 is configured to acquire an illumination distribution chart and a luminance channel standard image according to the original illumination image collected.

[0134] The image correction module 102 is configured to correct the luminance channel standard image using an adaptive gamma correction algorithm and the illumination distribution chart to obtain a luminance correction image.

[0135] The space conversion module 103 is configured to acquire a color image according to the luminance correction image, convert the color image to a YUV color space, acquire UV channel values of each pixel point in the YUV color space, and determine a Y channel white point position in the YUV color space according to the UV channel values of all pixel points and a preset UV highlight threshold.

[0136] The image enhancement module 104 is configured to perform white balance adjustment on each channel of the color image at the Y channel white point position to obtain an enhanced image.

[0137] The image denoising module 105 is configured to perform wavelet denoising processing on the enhanced image to obtain a denoised image.

[0138] The specific limitation of the underwater unevenly illuminated image enhancement system can refer to the limitation of the underwater unevenly illuminated image enhancement method, which will not be repeated here. Those skilled in the art can realize that the various modules and steps described in combination with the embodiments disclosed in the present application can be realized in hardware, software or combination of both. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0139] The embodiment of the present application provides an underwater unevenly illuminated image enhancement system, the system realizes adaptive gamma correction on the luminance single-channel image through the image correction module, so that more detail information in the dark area is highlighted, and better visual effect is achieved; the dynamic threshold white balance is realized on the obtained image through the space conversion module and the image enhancement module, so that the color deviation problem of the underwater image is solved; and the underwater image noise is removed through the image denoising module. Compared with the prior art, the present application can be applied to different complex underwater environments, the reliability and stability of the enhancement effect are improved, the overall color of the enhanced underwater image is more natural, and the visibility of the underwater image is significantly improved.

[0140] Figure 6 The embodiment of the present application provides a kind of computer equipment, including memory, processor and transceiver, they are connected through bus;Memory is used to store a group of computer program instructions and data, and can transmit the data stored to processor, processor can execute the program instruction stored in memory, to execute the steps of above-mentioned method.

[0141] Wherein, memory can include volatile memory or non-volatile memory, or can include both volatile and non-volatile memory;Processor can be central processing unit, microprocessor, application specific integrated circuit, programmable logic device or combination thereof. Through exemplary but not limiting description, above-mentioned programmable logic device can be complex programmable logic device, field programmable logic gate array, general array logic or any combination thereof.

[0142] In addition, memory can be physically independent unit, can also be integrated with processor.

[0143] Those skilled in the art can understand that, Figure 6 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer equipment to which the scheme of the present application is applied. The specific computer equipment can include more or less components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0144] In one embodiment, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above method.

[0145] The underwater uneven illumination image enhancement method, system, device and medium provided by the embodiment of the present application utilize adaptive gamma correction to adjust the illumination of the image, fully utilize the illumination component of the image to adaptively change the parameters of the gamma function, enhance the color contrast of the bright-dark transition area of the image, highlight more detail information in the dark area, achieve better visual effect, and at the same time, the obtained image is subjected to white balance adjustment and denoising, so that the underwater image is enhanced and the noise is removed, which can be applied to various complex underwater scenes and significantly improve the visibility of the underwater image.

[0146] In the above embodiments, all or part of the processes or functions can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the computer program product can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.). The computer readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (for example, floppy disk, hard disk, magnetic tape), optical media (for example, DVD), or semiconductor media (for example, SSD) and the like.

[0147] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments.

[0148] The above embodiments only express several preferred embodiments of the present application, which are described in a more specific and detailed manner, but cannot be understood as a limitation to the patent scope of the application. It should be pointed out that, for ordinary skilled in the art, several improvements and replacements can be made without departing from the technical principles of the present application, and these improvements and replacements should also be considered as the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the protection scope of the claims.

Claims

1. An underwater non-uniformly illuminated image enhancement method, characterized by, The method comprises the following steps: According to the original light image collected, the light distribution map and the brightness channel standard image are obtained; The brightness channel standard image is corrected by using the adaptive gamma correction algorithm and the light distribution map to obtain a brightness correction image; According to the brightness correction image, a color image is obtained, the color image is converted to YUV color space, and the UV channel values of each pixel point in the YUV color space are obtained; According to the UV channel values of all pixel points and the preset UV highlight threshold, the Y channel white point position in the YUV color space is determined; The white balance of each channel of the color image at the Y channel white point position is adjusted to obtain an enhanced image; The enhanced image is subjected to wavelet denoising processing to obtain a denoised image; The calculation formula of the adaptive gamma correction algorithm is: Wherein, wherein represents a luminance corrected image; represents a luminance channel standard image; represents an adaptive gamma parameter; represents an illumination profile; represents a luminance image dark region threshold; represents a luminance channel standard image; represents a luminance value of the luminance channel standard image at (x, y); represents a saturation value of the luminance channel standard image at (x, y); represents an average value of luminance of the luminance channel standard image; represents an average value of saturation of the luminance channel standard image; represents a gray value of the luminance channel standard image at (x, y); represents a median value of the gray value of the luminance channel standard image; The step of determining the Y channel white point position in the YUV color space according to the UV channel values of all pixel points and the preset UV highlight threshold comprises: The UV highlight threshold is determined according to the UV channel values of all pixel points, wherein the UV highlight threshold comprises a U channel highlight threshold and a V channel highlight threshold; Bright pixel points are screened out according to the UV channel values of each pixel point and the UV highlight threshold; The histogram of the bright pixel points in the Y channel is counted, and the average value of the highlights is determined according to the pixel value of the bright pixel points with a preset proportion in the histogram; The Y channel white point position in the YUV color space is determined according to the average value of the highlights.

2. The underwater non-uniform illumination image enhancement method of claim 1, wherein, The step of obtaining the light distribution map and the brightness channel standard image according to the original light image collected comprises: An original light image is obtained; The original light image is subjected to guided filtering to obtain a light distribution map; The original light image is subjected to color space conversion and channel separation to obtain a single channel image; the single channel image comprises a hue single channel image, a saturation single channel image and a brightness single channel image; The brightness single channel image is subjected to standardization processing to obtain a brightness channel standard image.

3. The underwater non-uniform illumination image enhancement method of claim 1, wherein, The step of screening out bright pixel points according to the UV channel values of each pixel point and the UV highlight threshold comprises: All pixel points in the YUV color space are traversed, and if the difference between the U channel value of the current pixel point and the U channel highlight threshold is not less than a first preset threshold, or the difference between the V channel value of the current pixel point and the V channel highlight threshold is not less than a second preset threshold, the current pixel point is divided into a bright pixel point; Wherein, the U channel highlight threshold is determined according to the U channel mean value and a preset U channel dynamic threshold, and the V channel highlight threshold is determined according to the V channel mean value and a preset V channel dynamic threshold.

4. The underwater non-uniform illumination image enhancement method of claim 1, wherein, The calculation formula for adjusting the white balance of each channel of the color image at the Y channel white point position to obtain an enhanced image is: wherein represents the R channel pixel value of the enhanced image; represents the G channel pixel value of the enhanced image; represents the B channel pixel value of the enhanced image; represents the R channel pixel value of the original lighting image; represents the R channel pixel average value of the original lighting image; represents the G channel pixel value of the original lighting image; represents the G channel pixel average value of the original lighting image; represents the B channel pixel value of the original lighting image; represents the B channel pixel average value of the original lighting image; represents the Y channel pixel maximum value under the YUV color space.

5. The underwater non-uniform illumination image enhancement method of claim 1, wherein, The step of subjecting the enhanced image to wavelet denoising processing to obtain a denoised image comprises: The enhanced image is converted to YCrCb color space to obtain a Y brightness channel image; The Y brightness channel image is subjected to wavelet transform to obtain wavelet coefficients; A wavelet threshold is set, and a wavelet inverse transform is performed on the Y luminance channel image by using the wavelet threshold and wavelet coefficients to obtain a denoised Y luminance channel image; The denoised Y luminance channel image is combined with the hue channel and the saturation channel of the enhanced image to obtain a combined image; The combined image is converted from the YCrCb color space to the RGB space to obtain a denoised image.

6. An underwater non-uniformly illuminated image enhancement system characterized by, The system comprises: An image acquisition module configured to acquire a light distribution map and a luminance channel standard image according to an original light image collected; An image correction module configured to correct the luminance channel standard image by using an adaptive gamma correction algorithm and the light distribution map to obtain a luminance correction image; A space conversion module configured to acquire a color image according to the luminance correction image, convert the color image to the YUV color space, acquire UV channel values at each pixel point in the YUV color space, and determine a Y channel white point position in the YUV color space according to the UV channel values at all pixel points and a preset UV highlight threshold value; An image enhancement module configured to perform white balance adjustment on each channel of the color image at the Y channel white point position to obtain an enhanced image; An image denoising module configured to perform wavelet denoising processing on the enhanced image to obtain a denoised image; The calculation formula of the adaptive gamma correction algorithm is as follows: wherein, wherein represents a luminance correction image; represents a luminance channel standard image; represents an adaptive gamma parameter; represents an illumination profile; represents a luminance image dark region threshold; represents a luminance channel standard image; represents a luminance value of the luminance channel standard image at (x, y); represents a saturation value of the luminance channel standard image at (x, y); represents an average value of luminance of the luminance channel standard image; represents an average value of saturation of the luminance channel standard image; represents a gray value of the luminance channel standard image at (x, y); represents a median value of the gray value of the luminance channel standard image; The determination of the Y channel white point position in the YUV color space according to the UV channel values at all pixel points and the preset UV highlight threshold value specifically comprises: determining the UV highlight threshold value according to the UV channel values at all pixel points, wherein the UV highlight threshold value comprises a U channel highlight threshold value and a V channel highlight threshold value; screening out bright pixel points according to the UV channel values at each pixel point and the UV highlight threshold value; statistically analyzing a histogram of the bright pixel points in the Y channel, and determining a highlight average value according to pixel values of bright pixel points with a preset proportion in the histogram; determining the Y channel white point position in the YUV color space according to the highlight average value.

7. A computer device, characterized by: The computer device comprises a processor and a memory, the processor is connected with the memory, the memory is used for storing a computer program, and the processor is used for executing the computer program stored in the memory to enable the computer device to execute the method in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer readable storage medium stores a computer program, and when the computer program is executed, the method in any one of claims 1 to 5 is implemented.

Images