1. An image enhancement processing method, characterized in that the method comprises:

if a first image input by a user is received, processing the first image according to a preset channelized processing rule to obtain a corresponding multi-channel characteristic image;

extracting channel image information corresponding to the channel characteristic image of each channel from the multi-channel characteristic image according to a preset image information extraction rule;

respectively whitening the channel characteristic images corresponding to the channel image information according to the channel image information of each channel to obtain whitening characteristic images corresponding to each channel;

performing cross-channel extraction on the whitening characteristic image to obtain corresponding cross-channel image information;

performing pixel-by-pixel optimization processing on the whitening characteristic image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimization characteristic image corresponding to each channel;

performing deconvolution processing on the optimized characteristic image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image;

and carrying out size adjustment on the deconvolution image according to the image size of the first image, and taking the obtained enhanced image as a target optimization image corresponding to the first image.

2. The image enhancement processing method according to claim 1, wherein the channelizing rule includes size information and a multi-channel convolution kernel, and the processing the first image according to the preset channelizing rule to obtain a corresponding multi-channel feature image includes:

carrying out size adjustment on the first image according to the size information to obtain a second image matched with the size information;

and performing convolution processing on the second image according to the multi-channel convolution kernel to obtain a corresponding multi-channel characteristic image.

3. The method according to claim 1, wherein the image information extraction rule includes a mean calculation formula and a matrix calculation formula, the channel image information includes a mean vector and a covariance matrix, and the extracting the channel image information corresponding to the channel feature image of each channel from the multi-channel feature image according to a preset image information extraction rule includes:

calculating a mean vector of each channel feature image according to the mean calculation formula;

and calculating the covariance matrix of each channel characteristic image according to the matrix calculation formula and the mean vector.

4. The method according to claim 3, wherein the whitening processing is performed on the channel feature images corresponding to the channel image information according to the channel image information of each channel, so as to obtain whitened feature images corresponding to each channel, and the method includes:

calculating a pixel difference value between each pixel value in each channel characteristic image and the mean vector of the channel characteristic image respectively to obtain difference value information corresponding to each channel characteristic image;

and performing inverse transformation on each covariance matrix and multiplying each pixel difference value in the difference value information of the corresponding channel characteristic image respectively to obtain a whitening characteristic image corresponding to each channel characteristic image.

5. The image enhancement processing method according to claim 1, wherein the cross-channel image information includes a cross-channel mean value and a cross-channel standard deviation, and the cross-channel extraction of the whitened feature image to obtain corresponding cross-channel image information includes:

performing cross-channel average calculation on a plurality of pixel values corresponding to each pixel in the whitening characteristic image to obtain a cross-channel average value corresponding to each pixel;

and calculating the cross-channel standard deviation of each pixel according to a preset standard deviation calculation formula and the cross-channel mean value.

6. The image enhancement processing method according to claim 5, wherein the optimization model includes a standardized calculation formula and a convolution operator, and the performing pixel-by-pixel optimization processing on the whitened feature image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimized feature image corresponding to each channel includes:

respectively carrying out standardization calculation on each pixel value of each whitening characteristic image according to the standardization calculation formula and the cross-channel image information to obtain a standard pixel value of each pixel in each whitening characteristic image;

performing convolution dimensionality reduction calculation on the cross-channel image information according to the convolution operator to obtain cross-channel image information subjected to dimensionality reduction;

and performing superposition calculation on the standard pixel value of each pixel in each whitening characteristic image and the cross-channel image information after dimensionality reduction to obtain an optimized characteristic image corresponding to each whitening characteristic image.

7. The image enhancement processing method according to claim 1, wherein the resizing the deconvolution image according to an image size of the first image, and taking a resulting enhanced image as a target optimized image corresponding to the first image, comprises:

acquiring corresponding size proportion information according to the image size of the first image and the image size of the deconvolution image;

and performing upsampling processing on pixels contained in the deconvolution image according to the size proportion information so as to perform size adjustment on the deconvolution image to obtain the target optimization image.

8. An image enhancement processing apparatus, characterized in that the apparatus comprises:

the multichannel characteristic image acquisition unit is used for processing a first image input by a user according to a preset channelized processing rule to obtain a corresponding multichannel characteristic image if the first image is received;

the channel image information acquisition unit is used for extracting channel image information corresponding to the channel characteristic image of each channel from the multi-channel characteristic image according to a preset image information extraction rule;

a whitening characteristic image obtaining unit, configured to perform whitening processing on channel characteristic images corresponding to the channel image information according to the channel image information of each channel, to obtain whitening characteristic images corresponding to each channel;

a cross-channel extraction unit, configured to perform cross-channel extraction on the whitening feature image to obtain corresponding cross-channel image information;

the optimization feature image acquisition unit is used for respectively carrying out pixel-by-pixel optimization processing on the whitening feature image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimization feature image corresponding to each channel;

the deconvolution image acquisition unit is used for performing deconvolution processing on the optimized characteristic image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image;

and the target optimization image acquisition unit is used for carrying out size adjustment on the deconvolution image according to the image size of the first image, and taking the obtained enhanced image as a target optimization image corresponding to the first image.

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the image enhancement processing method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, implements the image enhancement processing method according to any one of claims 1 to 7.

Background

Underwater image recognition plays an important role in the fields of ocean research, underwater robots and the like. For example, the proper functioning of an underwater surveillance system or an underwater drone relies heavily on efficient image recognition. The research on marine ecology by marine biologists also requires clear images as support. However, the existing underwater image technology has many problems, an accurate and clear image cannot be obtained by shooting in a visible spectrum, and in order to obtain more accurate information from the image shot underwater, the image shot underwater can be subjected to image enhancement processing so as to obtain more accurate information from the image subjected to enhancement processing. However, the processing method for performing enhancement processing on an image in the prior art is limited, and the prior art methods perform uniform enhancement processing on the whole image, so that enhancement processing cannot be performed on a specific region in the image, which results in low quality of the enhanced image obtained after enhancement processing and poor enhancement processing effect. Therefore, the prior art method has the problem of poor processing effect of enhancing the image.

Disclosure of Invention

The embodiment of the invention provides an image enhancement processing method, device, equipment and medium, aiming at solving the problem of poor processing effect in the prior art.

In a first aspect, an embodiment of the present invention provides an image enhancement processing method, which includes:

if a first image input by a user is received, processing the first image according to a preset channelized processing rule to obtain a corresponding multi-channel characteristic image;

extracting channel image information corresponding to the channel characteristic image of each channel from the multi-channel characteristic image according to a preset image information extraction rule;

respectively whitening the channel characteristic images corresponding to the channel image information according to the channel image information of each channel to obtain whitening characteristic images corresponding to each channel;

performing cross-channel extraction on the whitening characteristic image to obtain corresponding cross-channel image information;

performing pixel-by-pixel optimization processing on the whitening characteristic image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimization characteristic image corresponding to each channel;

performing deconvolution processing on the optimized characteristic image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image;

and carrying out size adjustment on the deconvolution image according to the image size of the first image, and taking the obtained enhanced image as a target optimization image corresponding to the first image.

In a second aspect, an embodiment of the present invention provides an image enhancement processing apparatus, including:

the multichannel characteristic image acquisition unit is used for processing a first image input by a user according to a preset channelized processing rule to obtain a corresponding multichannel characteristic image if the first image is received;

the channel image information acquisition unit is used for extracting channel image information corresponding to the channel characteristic image of each channel from the multi-channel characteristic image according to a preset image information extraction rule;

a whitening characteristic image obtaining unit, configured to perform whitening processing on channel characteristic images corresponding to the channel image information according to the channel image information of each channel, to obtain whitening characteristic images corresponding to each channel;

a cross-channel extraction unit, configured to perform cross-channel extraction on the whitening feature image to obtain corresponding cross-channel image information;

the optimization feature image acquisition unit is used for respectively carrying out pixel-by-pixel optimization processing on the whitening feature image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimization feature image corresponding to each channel;

the deconvolution image acquisition unit is used for performing deconvolution processing on the optimized characteristic image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image;

and the target optimization image acquisition unit is used for carrying out size adjustment on the deconvolution image according to the image size of the first image, and taking the obtained enhanced image as a target optimization image corresponding to the first image.

In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the processor implements the image enhancement processing method according to the first aspect.

In a fourth aspect, the present invention further provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and the computer program, when executed by a processor, causes the processor to execute the image enhancement processing method according to the first aspect.

The embodiment of the invention provides an image enhancement processing method, an image enhancement processing device and a computer readable storage medium. Processing a first image input by a user according to a channelization processing rule to obtain a multi-channel characteristic image, extracting channel image information of each channel from the multi-channel characteristic image, respectively whitening the channel characteristic image of the corresponding channel according to the channel image information to obtain a whitened characteristic image, performing cross-channel extraction to obtain cross-channel image information, respectively optimizing the whitened characteristic image of each channel to obtain a corresponding optimized characteristic image, performing deconvolution processing to obtain a deconvolution image, and performing size adjustment on the deconvolution image to obtain a target optimized image. By the method, each channel characteristic image is whitened according to the channel image information acquired from the first image information, and then the whitening characteristic image of each channel is optimized pixel by pixel through the cross-channel image information, so that each pixel in the image can be enhanced in a targeted manner, the effect of image enhancement is greatly enhanced, the intelligent enhancement of the image quality is realized, and the image quality is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic flowchart of an image enhancement processing method according to an embodiment of the present invention;

FIG. 2 is a schematic sub-flow chart of an image enhancement processing method according to an embodiment of the present invention;

FIG. 3 is a schematic view of another sub-flow of an image enhancement processing method according to an embodiment of the present invention;

FIG. 4 is a schematic view of another sub-flow of an image enhancement processing method according to an embodiment of the present invention;

FIG. 5 is a schematic view of another sub-flow of an image enhancement processing method according to an embodiment of the present invention;

FIG. 6 is a schematic view of another sub-flow of an image enhancement processing method according to an embodiment of the present invention;

FIG. 7 is a schematic view of another sub-flow of an image enhancement processing method according to an embodiment of the present invention;

FIG. 8 is a schematic block diagram of an image enhancement processing apparatus provided by an embodiment of the present invention;

FIG. 9 is a schematic block diagram of a computer device provided by an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

Referring to fig. 1, fig. 1 is a schematic flow chart of an image enhancement processing method according to an embodiment of the present invention; the image enhancement processing method is applied to a user terminal or a management server, the image enhancement processing method is executed through application software installed in the user terminal or the management server, the user terminal is a terminal device which can receive a first image input by a user and perform image enhancement processing, such as a desktop computer, a notebook computer, a tablet computer or a mobile phone, and the management server is a server end which can receive the first image sent by the user through the terminal and perform image enhancement processing, such as a server constructed by an enterprise, a medical institution or a government department. As shown in fig. 1, the method includes steps S110 to S170.

S110, if a first image input by a user is received, processing the first image according to a preset channelized processing rule to obtain a corresponding multi-channel characteristic image.

And if a first image input by a user is received, processing the first image according to a preset channelized processing rule to obtain a corresponding multi-channel characteristic image. The first image input by the user may be an image with poor image quality, such as an image taken in a poor light environment, an image taken underwater, and the like. The channelizing processing rule is a specific rule for processing the first image to obtain channel characteristic images corresponding to a plurality of channels, wherein the channelizing processing rule comprises size information and a multichannel convolution kernel, the size of the first image can be adjusted through the size information to obtain a second image matched with the size information, and the second image is convolved through the multichannel convolution kernel to obtain the corresponding multichannel characteristic images.

In one embodiment, as shown in FIG. 2, step S110 includes sub-steps S111 and S112.

And S111, carrying out size adjustment on the first image according to the size information to obtain a second image matched with the size information.

Specifically, the first image may be downsampled based on the size information, and if downsampling is performed to reduce the first image, the first image may be reduced to the second image based on the size information, and the image size of the second image may be matched to the size information. The size information may include information on an image length size, i.e., information on the number of pixels included in the image in the length direction, and an image width size, i.e., information on the number of pixels included in the image in the width direction. For example, the size information of the first image may be represented as N × M, and the size information in the channelization rule is N × M (N < N and M < M), then the first image may be downsampled to obtain the second image with the size of N × M.

And S112, performing convolution processing on the second image according to the multi-channel convolution kernel to obtain a corresponding multi-channel characteristic image.

The multi-channel convolution kernel comprises a plurality of channels, each channel correspondingly comprises a convolution kernel or a plurality of convolution kernels, the convolution kernel of any channel can perform convolution processing on the second image to obtain a channel characteristic image corresponding to the channel, the convolution kernels of the channels respectively perform convolution processing on the second image to obtain a plurality of channel characteristic images corresponding to the channels, and the plurality of channel characteristic images form the multi-channel characteristic image corresponding to the second image. For example, if a certain channel includes two convolution kernels of 3 × 3, the first convolution kernel may perform convolution processing with step size 1 on the second image to obtain a convolution image, and then the second convolution kernel may perform further convolution processing on the convolution image to obtain one channel feature image corresponding to the second image, where the sizes of the obtained multiple channel feature images are the same.

And S120, extracting channel image information corresponding to the channel characteristic image of each channel from the multi-channel characteristic image according to a preset image information extraction rule.

And extracting channel image information corresponding to the channel characteristic image of each channel from the multi-channel characteristic image according to a preset image information extraction rule. The multi-channel characteristic image comprises a channel characteristic image corresponding to each channel, channel image information can be correspondingly extracted from each channel characteristic image according to an image information extraction rule, the channel image information can represent the whole information of the channel characteristic image, the image information extraction rule is a specific rule for extracting the channel image information from each channel characteristic image, the image information extraction rule comprises a mean value calculation formula and a matrix calculation formula, and the channel image information comprises a mean value vector and a covariance matrix. The mean vector of each channel characteristic image can be obtained through calculation of a mean calculation formula, the covariance matrix of each channel characteristic image is further obtained through calculation of a matrix calculation formula and the mean vector, and the mean vector and the covariance matrix of the channel characteristic images are combined to serve as channel image information of the channel characteristic images.

In an embodiment, as shown in fig. 3, step S120 includes substeps S121 and S122.

And S121, calculating a mean vector of each channel characteristic image according to the mean calculation formula.

Specifically, the multi-channel feature image obtained from the second image may be represented as R^C×H×WThen any pixel x in the multi-channel feature image can be represented as x ∈ R^C×H×WAnd C is the total number of channels, H is the length of the channel characteristic image, and W is the width of the channel characteristic image. And calculating one channel characteristic image according to a mean value calculation formula to obtain a mean value vector corresponding to the channel characteristic image, and calculating a plurality of channel characteristic images to obtain a plurality of corresponding mean value vectors.

The average calculation formula can be expressed by formula (1):

wherein x is_hwThat is, the image is represented as the pixel value corresponding to the pixel point with coordinate position (H, w) in one channel characteristic image, and H can be [1, H]An arbitrary integer within, W may be [1, W ]]And any integer in the vector is mu, namely the mean vector obtained by calculation.

And S122, calculating a covariance matrix of each channel characteristic image according to the matrix calculation formula and the mean vector.

The covariance matrix of each channel feature image can be further calculated by combining the calculated mean vector through a matrix calculation formula. Specifically, the matrix calculation formula can be represented by formula (2):

wherein x is_hwThe method comprises the steps of representing a pixel value corresponding to a pixel point with a coordinate position (h, w) in a channel characteristic image, wherein mu is a mean vector corresponding to the channel characteristic image, T is matrix transfer calculation, alpha is a parameter value preset in a formula, alpha can be a positive integer, I is a unit matrix in the formula, and psi is a covariance matrix corresponding to the channel characteristic image.

And S130, respectively whitening the channel characteristic images corresponding to the channel image information according to the channel image information of each channel to obtain whitened characteristic images corresponding to each channel.

And respectively whitening the channel characteristic images corresponding to the channel image information according to the channel image information of each channel to obtain whitening characteristic images corresponding to each channel. Each channel corresponds to one channel characteristic image, channel image information can be correspondingly acquired from each channel characteristic image, one channel corresponds to one channel image information, the channel characteristic image uniquely corresponding to the channel image information can be whitened according to each channel image information, a whitening characteristic image corresponding to each channel characteristic image is obtained, and one channel corresponds to one whitening characteristic image.

In one embodiment, as shown in fig. 4, step S130 includes sub-steps S131 and S132.

S131, calculating a pixel difference value between each pixel value in each channel characteristic image and the mean value vector of the channel characteristic image respectively to obtain difference value information corresponding to each channel characteristic image.

Specifically, for any one channel feature image, a pixel difference between a pixel value of each pixel point in the channel feature image and a mean vector of the channel feature image can be calculated, the pixel differences of each pixel in the channel feature image are combined to obtain corresponding difference information, and then one channel feature image can correspondingly obtain one difference information.

S132, performing inverse transformation on each covariance matrix, and multiplying each pixel difference value in the difference value information of the corresponding channel characteristic image respectively to obtain a whitening characteristic image corresponding to each channel characteristic image.

And performing inverse transformation on each covariance matrix, and multiplying the matrix obtained by the inverse transformation by each pixel difference in the difference information of the corresponding channel characteristic image respectively, so that each pixel of the corresponding channel characteristic image can obtain a corresponding product result, and the product result corresponding to each pixel is obtained to be used as the whitening characteristic image of the corresponding channel characteristic image.

Specifically, the above calculation process can be represented by formula (3):

Γ(x_hw)＝Ψ^-1/2×(x_hw-μ) (3)；

wherein H is ∈ [1, H]And is an integer, W is [1, W ]]And is an integer, Ψ^-1/2I.e. a calculation process for inverse transformation of the covariance matrix Ψ of a certain channel feature image, x_hwThat is, the pixel value corresponding to the pixel point with the coordinate position (h, w) in the difference information of the channel feature image, μ is the mean vector corresponding to the channel feature image, Γ (x)_hw) I.e. obtained and x_hwA corresponding one of the product results.

And S140, performing cross-channel extraction on the whitening characteristic image to obtain corresponding cross-channel image information.

And performing cross-channel extraction on the whitening characteristic image to obtain corresponding cross-channel image information. Each channel corresponds to one whitening characteristic image, in order to obtain the cross-channel information of a plurality of whitening characteristic images corresponding to the channels, the cross-channel extraction can be carried out on the plurality of whitening characteristic images to obtain the corresponding cross-channel image information, and the cross-channel image information integrates the information of each whitening characteristic image, so that the cross-channel image information can carry out integral representation on the cross-channel information of the plurality of whitening characteristic images. Wherein the cross-channel image information comprises a cross-channel mean and a cross-channel standard deviation.

In an embodiment, as shown in fig. 5, step S140 includes sub-steps S141 and S142.

And S141, performing cross-channel average calculation on a plurality of pixel values corresponding to each pixel in the whitening characteristic image to obtain a cross-channel average value corresponding to each pixel.

Any pixel corresponds to a pixel value in each whitening characteristic image, one pixel corresponds to a plurality of pixel values in a plurality of whitening characteristic images, a plurality of pixel values respectively corresponding to each pixel in the plurality of whitening characteristic images can be obtained for cross-channel average calculation, and a cross-channel mean value corresponding to each pixel is obtained, namely, one pixel is uniquely corresponding to one cross-channel mean value. Specifically, the calculation formula of the cross-channel average calculation can be represented by formula (4):

wherein H is ∈ [1, H]And is an integer, W is [1, W ]]And is an integer, C is [1, C ]]And is an integer, y_chwThat is, the pixel value corresponding to the pixel point with the coordinate position (h, w) in the whitening characteristic image corresponding to the C-th channel, wherein C represents the total number of the channels, and ξ_hwNamely, the calculated cross-channel mean value corresponding to the pixel point with the coordinate position (h, w).

And S142, calculating the cross-channel standard deviation of each pixel according to a preset standard deviation calculation formula and the cross-channel mean value.

The cross-channel standard deviation of each pixel can be calculated according to a standard deviation calculation formula and the obtained cross-channel mean value, and then one cross-channel standard deviation can be calculated by combining one pixel with one cross-channel mean value corresponding to the pixel. Specifically, the standard deviation calculation formula can be represented by formula (5):

wherein H is ∈ [1, H]And is an integer, W is [1, W ]]And is an integer, C is [1, C ]]And is an integer, y_chwThat is, the pixel value corresponding to the pixel point with the coordinate position (h, w) in the whitening characteristic image corresponding to the C-th channel, wherein C represents the total number of the channels, and ξ_hwI.e. the cross-channel mean value, sigma, corresponding to the pixel point with the coordinate position (h, w) obtained by calculation_hwThe cross-channel standard deviation corresponding to the pixel point with the coordinate position (h, w) is obtained through calculation, alpha 'is a preset parameter value in a formula, the alpha' can be a positive integer, and I is an identity matrix in the formula.

S150, performing pixel-by-pixel optimization processing on the whitening characteristic image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimization characteristic image corresponding to each channel.

And respectively carrying out pixel-by-pixel optimization processing on the whitening characteristic image of each channel according to a preset optimization model and the cross-channel image information to obtain an optimization characteristic image corresponding to each channel. The whitening characteristic images corresponding to each channel are respectively subjected to pixel-by-pixel optimization processing through the optimization model and the obtained cross-channel image information, namely, each pixel in the whitening characteristic images is respectively subjected to targeted optimization processing, one whitening characteristic image is subjected to pixel-by-pixel optimization processing to obtain a corresponding optimized characteristic image, and then multiple kinds of whitening characteristic images are respectively subjected to pixel-by-pixel optimization processing to obtain multiple corresponding optimized characteristic images, wherein the optimization model comprises a standardized calculation formula and a convolution operator.

In one embodiment, as shown in fig. 6, step S150 includes sub-steps S151, S152, and S153.

And S151, respectively carrying out standardization calculation on each pixel value of each whitening characteristic image according to the standardization calculation formula and the cross-channel image information to obtain a standard pixel value of each pixel in each whitening characteristic image.

The method comprises the steps of firstly, respectively carrying out standardized calculation on the pixel value of each pixel in each whitening characteristic image through a standardized calculation formula and a cross-channel mean value of each pixel in cross-channel image information, and obtaining a standard pixel value corresponding to each pixel in each whitening characteristic image. Specifically, the normalized calculation formula can be expressed by formula (6):

wherein H is ∈ [1, H]And is an integer, W is [1, W ]]And is an integer, C is [1, C ]]And is an integer, y_chwThat is, the pixel value xi corresponding to the pixel point with the coordinate position (h, w) in the whitening characteristic image corresponding to the c-th channel_hwNamely, the cross-channel images corresponding to the pixel points with the coordinate positions (h, w) in the cross-channel image informationValue σ_hwNamely the cross-channel standard deviation omega (y) corresponding to the pixel point with the coordinate position of (h, w) in the cross-channel image information_chw) I.e. calculated and y_chwThe corresponding standard pixel value.

S152, carrying out convolution dimensionality reduction calculation on the cross-channel image information according to the convolution operator to obtain the cross-channel image information after dimensionality reduction.

The cross-channel mean value and the cross-channel standard deviation contained in the cross-channel image information can be respectively subjected to convolution dimensionality reduction calculation according to a convolution operator, for example, the cross-channel mean value and the cross-channel standard deviation can be subjected to convolution dimensionality reduction calculation through a 1-by-1 convolution kernel, and two different convolution kernels are used for respectively carrying out the convolution dimensionality reduction calculation on the cross-channel mean value and the cross-channel standard deviation.

S153, performing superposition calculation on the standard pixel value of each pixel in each whitening characteristic image and the cross-channel image information after dimensionality reduction to obtain an optimized characteristic image corresponding to each whitening characteristic image.

Specifically, the process of the superposition calculation can be expressed by the following formula (7):

z_chw＝F₁(σ_hw)×Ω(y_chw)+F₂(ξ_hw) (7)；

wherein H is ∈ [1, H]And is an integer, W is [1, W ]]And is an integer, C is [1, C ]]And is an integer, omega (y)_chw) That is, the standard pixel value corresponding to the pixel point with the coordinate position (h, w) in the whitening characteristic image corresponding to the c-th channel, F₁(σ_hw) That is, a calculated value obtained by performing dimensionality reduction calculation on the cross-channel standard deviation corresponding to the pixel point with the coordinate position (h, w), F₂(ξ_hw) Namely, the calculated value obtained by carrying out dimensionality reduction calculation on the cross-channel mean value corresponding to the pixel point with the coordinate position of (h, w), z_chwI.e. with pixel point y_chwThe corresponding calculated value of the superposition.

And acquiring a superposition calculation value corresponding to each pixel point in any whitening characteristic image to be combined into an optimized characteristic image corresponding to the whitening characteristic image, and acquiring the optimized characteristic image corresponding to each whitening characteristic image according to the method.

And S160, carrying out deconvolution processing on the optimized characteristic image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image.

And carrying out deconvolution processing on the optimized characteristic image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image. Specifically, the deconvolution processing is performed on the obtained multiple optimized feature images, that is, the inverse operation corresponding to the convolution processing performed on the second image, and the multiple optimized feature images corresponding to the multiple channels can be integrated into one deconvolution image through the deconvolution operation, so that the deconvolution image includes image information obtained by combining the optimized feature images of each channel.

If the size of the optimized feature images is H × W and the number of the optimized feature images is C, the size of one deconvolution image obtained by deconvolving the plurality of optimized feature images may be n × m.

S170, carrying out size adjustment on the deconvolution image according to the image size of the first image, and taking the obtained enhanced image as a target optimization image corresponding to the first image.

And carrying out size adjustment on the deconvolution image according to the image size of the first image, and taking the obtained enhanced image as a target optimization image corresponding to the first image. After the deconvolution image is obtained, the size of the deconvolution image can be adjusted according to the image size of the first image, the enhanced image obtained after the size adjustment is used as a corresponding target optimization image, the image size of the obtained target optimization image can be the same as the image size of the first image, that is, the image size of the target optimization image is N × M.

In one embodiment, as shown in FIG. 7, step S170 includes sub-steps S171 and S172.

And S171, acquiring corresponding size proportion information according to the image size of the first image and the image size of the deconvolution image.

Specifically, the size ratio information may be obtained by correspondingly calculating the image size of the first image and the image size of the deconvolution image, and the size ratio information includes a width ratio of the first image to the deconvolution image and a length ratio of the first image to the deconvolution image.

And S172, performing up-sampling processing on pixels contained in the deconvolution image according to the size proportion information so as to perform size adjustment on the deconvolution image to obtain the target optimization image.

The pixels contained in the deconvolution image can be subjected to upsampling processing according to the size proportion information, the processing process of the upsampling processing is opposite to that of the downsampling processing, namely, the images can be subjected to up-sampling processing to perform amplification adjustment, the times of the amplification adjustment correspond to the corresponding ratios in the size proportion information, and then the enhanced images obtained after the size adjustment of the deconvolution images can be used as target optimized images corresponding to the first images.

The technical method can be applied to application scenes including intelligent enhancement processing on image quality, such as intelligent government affairs, intelligent city management, intelligent community, intelligent security protection, intelligent logistics, intelligent medical treatment, intelligent education, intelligent environmental protection and intelligent traffic, and the like, so that the construction of a smart city is promoted.

In the image enhancement processing method provided by the embodiment of the invention, a first image input by a user is processed according to a channelization processing rule to obtain a multi-channel characteristic image, channel image information of each channel is extracted from the multi-channel characteristic image, the channel characteristic images of the corresponding channels are respectively whitened according to the channel image information to obtain a whitened characteristic image, cross-channel extraction is carried out to obtain cross-channel image information, the whitened characteristic images of each channel are respectively subjected to pixel-by-pixel optimization processing to obtain a corresponding optimized characteristic image, then deconvolution processing is carried out to obtain a deconvolution image, and the size of the deconvolution image is adjusted to obtain a target optimized image. By the method, each channel characteristic image is whitened according to the channel image information acquired from the first image information, and then the whitening characteristic image of each channel is optimized pixel by pixel through the cross-channel image information, so that each pixel in the image can be enhanced in a targeted manner, the effect of image enhancement is greatly enhanced, the intelligent enhancement of the image quality is realized, and the image quality is improved.

The embodiment of the present invention further provides an image enhancement processing apparatus, which can be configured in a user terminal or a management server, and the image enhancement processing apparatus is configured to execute any embodiment of the aforementioned image enhancement processing method. Specifically, referring to fig. 8, fig. 8 is a schematic block diagram of an image enhancement processing apparatus according to an embodiment of the present invention.

As shown in fig. 8, the image enhancement processing apparatus 100 includes a multi-channel feature image acquisition unit 110, a channel image information acquisition unit 120, a whitening-feature-image acquisition unit 130, a cross-channel extraction unit 140, an optimized feature image acquisition unit 150, a deconvolution image acquisition unit 160, and a target optimized image acquisition unit 170.

The multi-channel feature image obtaining unit 110 is configured to, if a first image input by a user is received, process the first image according to a preset channelization processing rule to obtain a corresponding multi-channel feature image.

In one embodiment, the pixel division processing unit 110 includes sub-units: the second image acquisition unit is used for carrying out size adjustment on the first image according to the size information to obtain a second image matched with the size information; and the convolution processing unit is used for carrying out convolution processing on the second image according to the multi-channel convolution kernel to obtain a corresponding multi-channel characteristic image.

A channel image information obtaining unit 120, configured to extract channel image information corresponding to the channel feature image of each channel from the multi-channel feature image according to a preset image information extraction rule.

In a specific embodiment, the channel image information obtaining unit 120 includes sub-units: the mean vector acquiring unit is used for calculating a mean vector of each channel characteristic image according to the mean calculation formula; and the covariance matrix acquisition unit is used for calculating a covariance matrix of each channel characteristic image according to the matrix calculation formula and the mean vector.

The whitening characteristic image obtaining unit 130 is configured to perform whitening processing on the channel characteristic images corresponding to the channel image information according to the channel image information of each channel, so as to obtain a whitening characteristic image corresponding to each channel.

In a specific embodiment, the whitening-feature-image obtaining unit 130 includes sub-units: the difference information calculation unit is used for calculating the pixel difference between each pixel value in each channel characteristic image and the mean vector of the channel characteristic image to obtain difference information corresponding to each channel characteristic image; and the whitening characteristic image acquisition unit is used for performing inverse transformation on each covariance matrix and multiplying each pixel difference value in the difference value information of the corresponding channel characteristic image respectively to obtain a whitening characteristic image corresponding to each channel characteristic image.

A cross-channel extracting unit 140, configured to perform cross-channel extraction on the whitening feature image to obtain corresponding cross-channel image information.

In one embodiment, the cross-channel extraction unit 140 includes sub-units: a cross-channel mean value calculation unit, configured to perform cross-channel mean calculation on multiple pixel values corresponding to each pixel in the whitening feature image to obtain a cross-channel mean value corresponding to each pixel; and the cross-channel standard deviation calculating unit is used for calculating the cross-channel standard deviation of each pixel according to a preset standard deviation calculation formula and the cross-channel mean value.

And an optimized feature image obtaining unit 150, configured to perform pixel-by-pixel optimization processing on the whitened feature image of each channel according to a preset optimization model and the cross-channel image information, to obtain an optimized feature image corresponding to each channel.

In a specific embodiment, the optimized feature image obtaining unit 150 includes sub-units: a standard pixel value obtaining unit, configured to perform normalization calculation on each pixel value of each whitening feature image according to the normalization calculation formula and the cross-channel image information, to obtain a standard pixel value of each pixel in each whitening feature image; the convolution dimensionality reduction calculation unit is used for performing convolution dimensionality reduction calculation on the cross-channel image information according to the convolution operator to obtain cross-channel image information subjected to dimensionality reduction; and the superposition calculation unit is used for carrying out superposition calculation on the standard pixel value of each pixel in each whitening characteristic image and the cross-channel image information after dimensionality reduction to obtain an optimized characteristic image corresponding to each whitening characteristic image.

And the deconvolution image acquisition unit 160 is configured to perform deconvolution processing on the optimized feature image according to a preset deconvolution processing rule to obtain a corresponding deconvolution image.

And a target optimized image obtaining unit 170, configured to perform size adjustment on the deconvolution image according to the image size of the first image, and use the obtained enhanced image as a target optimized image corresponding to the first image.

In a specific embodiment, the target-optimized image obtaining unit 170 includes sub-units: the size ratio information acquisition unit is used for acquiring corresponding size ratio information according to the image size of the first image and the image size of the deconvolution image; and the size adjusting unit is used for performing up-sampling processing on pixels contained in the deconvolution image according to the size proportion information so as to perform size adjustment on the deconvolution image to obtain the target optimization image.

The image enhancement processing device provided by the embodiment of the invention applies the image enhancement processing method, processes a first image input by a user according to a channelization processing rule to obtain a multi-channel characteristic image, extracts channel image information of each channel from the multi-channel characteristic image, respectively whitens the channel characteristic image of the corresponding channel according to the channel image information to obtain a whitened characteristic image, extracts cross-channel to obtain cross-channel image information, respectively optimizes the whitened characteristic image of each channel pixel by pixel to obtain a corresponding optimized characteristic image, performs deconvolution processing to obtain a deconvolution image, and performs size adjustment on the deconvolution image to obtain a target optimized image. By the method, each channel characteristic image is whitened according to the channel image information acquired from the first image information, and then the whitening characteristic image of each channel is optimized pixel by pixel through the cross-channel image information, so that each pixel in the image can be enhanced in a targeted manner, the effect of image enhancement is greatly enhanced, the intelligent enhancement of the image quality is realized, and the image quality is improved.

The image enhancement processing apparatus described above may be implemented in the form of a computer program that can be run on a computer device as shown in fig. 9.

Referring to fig. 9, fig. 9 is a schematic block diagram of a computer device according to an embodiment of the present invention. The computer device may be a user terminal or a management server for executing an image enhancement processing method for intelligently enhancing image quality.

Referring to fig. 9, the computer device 500 includes a processor 502, memory, and a network interface 505 connected by a system bus 501, where the memory may include a storage medium 503 and an internal memory 504.

The storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032, when executed, may cause the processor 502 to perform an image enhancement processing method, wherein the storage medium 503 may be a volatile storage medium or a non-volatile storage medium.

The processor 502 is used to provide computing and control capabilities that support the operation of the overall computer device 500.

The internal memory 504 provides an environment for running the computer program 5032 in the storage medium 503, and when the computer program 5032 is executed by the processor 502, the processor 502 can be enabled to execute the image enhancement processing method.

The network interface 505 is used for network communication, such as providing transmission of data information. Those skilled in the art will appreciate that the configuration shown in fig. 9 is a block diagram of only a portion of the configuration associated with aspects of the present invention and is not intended to limit the computing device 500 to which aspects of the present invention may be applied, and that a particular computing device 500 may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

The processor 502 is configured to run a computer program 5032 stored in the memory to implement the corresponding functions in the image enhancement processing method.

Those skilled in the art will appreciate that the embodiment of a computer device illustrated in fig. 9 does not constitute a limitation on the specific construction of the computer device, and that in other embodiments a computer device may include more or fewer components than those illustrated, or some components may be combined, or a different arrangement of components. For example, in some embodiments, the computer device may only include a memory and a processor, and in such embodiments, the structures and functions of the memory and the processor are consistent with those of the embodiment shown in fig. 9, and are not described herein again.

It should be understood that, in the embodiment of the present invention, the Processor 502 may be a Central Processing Unit (CPU), and the Processor 502 may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, and the like. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

In another embodiment of the invention, a computer-readable storage medium is provided. The computer readable storage medium may be a volatile or non-volatile computer readable storage medium. The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps included in the image enhancement processing method described above.

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described apparatuses, devices and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described in a functional general in the foregoing description for the purpose of illustrating clearly the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the embodiments provided by the present invention, it should be understood that the disclosed apparatus, device and method can be implemented in other ways. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only a logical division, and there may be other divisions when the actual implementation is performed, or units having the same function may be grouped into one unit, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may also be an electric, mechanical or other form of connection.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present invention.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention essentially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product stored in a computer-readable storage medium, which includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned computer-readable storage media comprise: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disk.

While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.