CN112184597A - Image restoration device and method - Google Patents
Image restoration device and method Download PDFInfo
- Publication number
- CN112184597A CN112184597A CN202011225419.9A CN202011225419A CN112184597A CN 112184597 A CN112184597 A CN 112184597A CN 202011225419 A CN202011225419 A CN 202011225419A CN 112184597 A CN112184597 A CN 112184597A
- Authority
- CN
- China
- Prior art keywords
- image
- module
- restoration
- convolution
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses an image restoration device and method. The image restoration device comprises an image acquisition module, an image preprocessing module, a target sample sampling module, a restoration processing module and a restoration image display module. The image acquisition module is used for acquiring original fuzzy image information; the image preprocessing module receives the image information acquired by the image acquisition module and preprocesses the image information; the target sample sampling module performs sliding sampling on the image obtained by the image preprocessing module to obtain a target sampling sample image; the restoration processing module inputs the target sampling sample image into a trained convolutional neural network model, and restores the image; and the restoration image display module receives the output restoration image of the restoration processing module and displays the restoration image. The invention also provides an image restoration method, and by applying the method, the original content of the image can be displayed, the image quality is improved, and the visual experience of an observer can be improved.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to an image restoration device and method.
Background
As an important medium for transferring information, image information is very important. However, due to various reasons, images are mosaiced, stained or noise interference is generated on the images by external environment, and the images displayed on the image receiving end have poor presentation effect, and in many research fields such as optics, medicine, astronomy and meteorology, the requirements on the definition and quality of the images are high, so that the blurred images must be processed. For this reason, image restoration technology has been developed and is an indispensable and high-tech technology in modern times.
Most of the existing image restoration technologies have an unobvious image processing effect, and the display effect of the final restored image is not good enough.
Therefore, a new technical solution is provided to solve the above problems.
Disclosure of Invention
In view of the above, the present invention provides an image restoration apparatus and method to solve the above technical problems.
In one aspect, to achieve the above object, the present invention provides an image restoration apparatus comprising:
an image restoration device comprises an image acquisition module, an image preprocessing module, a target sample sampling module, a restoration processing module and a restoration image display module.
In the above image restoration device, the image acquisition module is configured to acquire image information to be processed having a pixel loss region.
In the image restoration device, the image preprocessing module is electrically connected to the image acquisition module, and the image preprocessing module receives and preprocesses the image information acquired by the image acquisition module.
In the image restoration device, the target sample sampling module is electrically connected to the image preprocessing module, and the target sample sampling module performs sliding sampling on the image obtained by the image preprocessing module through a sliding window to obtain a target sampling sample image corresponding to the target image.
In the above image restoration device, the restoration processing module is electrically connected to the target sample sampling module, and the restoration processing module performs restoration processing on the image by inputting the target sample image into a trained convolutional neural network model.
In the above-described image restoration device, the restoration image display module is electrically connected to the restoration processing module, and the restoration image display module receives and displays an output restoration image of the restoration processing module.
In the image restoration device, the image acquisition module comprises a camera and a controller, the camera sends acquired image information to be processed with pixel loss areas to the controller, pixel correction is realized by adopting a correction method based on Gamma correction, and the dark color part and the light color part in an image signal are detected and increased in proportion, so that the image contrast effect is improved.
In the above image restoration apparatus, the image preprocessing module includes an interference information removing unit, a denoising unit and an image normalization unit, the interference information removing unit includes a defogging module, a stain removing module and a demosaicing module, the defogging module is configured to estimate and correct the transmittance in the foggy image model by using a dark channel prior method, perform defogging recovery on the original foggy image by using a defogging recovery formula, and perform initial assignment on the ideal image, the stain removing module adjusts the gray level according to the local gray level fluctuation degree of the pixel neighborhood to suppress background interference, so as to achieve adaptive removal of stains, the dynamic adjustment of the gray level can be achieved by a convolution operation method of weighted average of the pixel neighborhood, and the demosaicing module achieves adaptive removal of stains by separately performing convolution operation on a luminance signal Y and three chrominance signals R, G, B, demosaicing processing is carried out, luminance and chrominance signals of the current pixel position are recovered from input original data, and the luminance and chrominance signals of the demosaiced current pixel position are converted through matrix operation to obtain corresponding original color gamut image signals; the denoising unit obtains a preliminary denoising image by performing two-dimensional wiener filtering on an original image, calculates residual quantity of pixels according to the original image and the preliminary denoising image, calculates a weight matrix of each pixel by adopting a non-local mean filtering method according to the residual quantity, and performs non-local mean calculation on the image to be processed according to the weight matrix so as to realize denoising processing on the image to be processed and obtain a denoising image; the image normalization unit is used for performing normalization processing on the denoised image to obtain a normalized image.
In the image restoration device, the image normalization unit includes a detection module, a preliminary normalization module, an estimation module, a positioning module, and an extraction module, and the detection module is configured to detect a plurality of feature points in the denoised image; the preliminary normalization module normalizes the denoised image based on the plurality of feature points detected by the detection module, thereby obtaining a preliminary normalized image; the estimation module estimates original features in the denoised image; the positioning module positions a characteristic region of a preliminary normalized image based on normalized characteristic points of the preliminary normalized image and original characteristics in the denoised image estimated by the estimation module; the extraction module extracts features from a feature region in the preliminary normalized image.
In the image restoration device, the restoration processing module includes a training determining module, a first result judging module, a first adjusting module, a first convolution module, a second result judging module, an inverse normalization module, and a second adjusting module; the training module trains the characteristic parameters corresponding to the first convolution module and the second convolution module respectively through a data set containing original blurred image samples to obtain a first characteristic parameter set corresponding to the first convolution module and a second characteristic parameter set corresponding to the second convolution module, wherein the data set containing the original blurred image samples comprises a training set and a testing set, the training set is used for training a model, and the testing set is used for testing the model obtained by training and is used for approximating the generalization capability of the model; the first result judging module calculates loss by using a cross entropy function to obtain accuracy, judges whether the accuracy reaches a specific value or not, and controls the training module to stop training when the accuracy reaches the specific value; the determination module is configured to determine whether a resolution of the target sample image is consistent with an input dimension of the first convolution module; the first adjusting module is used for carrying out scaling processing on the target sampling sample image to obtain a first image when the resolution of the target sampling sample image is inconsistent with the input dimension of the first convolution module; the first convolution module is a positive convolution module, and the first convolution module performs convolution processing on the first image based on the first characteristic parameter set obtained by training of the training module to obtain a first characteristic of the first image; the second convolution module is a deconvolution module, and the second convolution module performs deconvolution processing on the first feature representation of the first image obtained by the first convolution module based on the second feature parameter set obtained by the training of the training module to obtain a second image; the second result judging module is used for judging whether the restoration error between the original images corresponding to the second image reaches a preset value or not, and if the restoration error reaches the preset value, the convolution process is stopped; the reverse normalization module is used for performing reverse normalization processing on the obtained second image to obtain a first restored image; the second adjusting module is used for adjusting the resolution of the first restoration image so that the resolution of the second image is the same as the resolution of the original blurred image.
In the above-described image restoration device, the restored image display module includes a display control unit and a display unit; the display control unit comprises an image depth of field analysis module and a bending degree control module, wherein the image depth of field analysis module is used for analyzing the depth of field of the display image, the image depth of field analysis module comprises an average value of the depth of field of a plurality of display images in a fixed time period, and the bending degree control module controls the bending degree of the display unit in a grading mode according to the average value; the display unit includes a light emitting device including a nitride semiconductor light emitting element, a red phosphor and a green phosphor, and a color filter having a blue pixel, a red pigment and a green pigment, and an image display device configured by combining the light emitting device and the color filter can realize higher white luminance and higher color purity, thereby being capable of reducing color difference of a displayed image.
In another aspect, the present invention provides an image restoration method, including:
acquiring to-be-processed image information with a pixel loss area;
carrying out interference information unit, denoising and normalization processing on the image to obtain a normalized image;
performing sliding sampling on the normalized image to obtain a target sampling sample image;
training the characteristic parameters corresponding to the first convolution module and the second convolution module respectively;
calculating loss to obtain accuracy, and judging the accuracy to control the training process;
adjusting the resolution of the target sampling sample image to be consistent with the input dimension of the first convolution module to obtain a first image;
performing convolution processing and deconvolution processing on the first image to obtain a second image;
judging whether the recovery error reaches a preset value or not, if not, returning to the step of training the characteristic parameters corresponding to the first convolution module and the second convolution module, repeating the steps, and if so, stopping the convolution process;
performing inverse normalization processing and resolution adjustment on the second image;
and displaying the restored image.
In the above image restoration method, the normalization process includes detecting a plurality of feature points in the denoised image; normalizing the denoised image based on the plurality of feature points detected by the detection module, thereby obtaining a preliminary normalized image; an estimation module estimates original features in the denoised image; based on the normalized feature points of the preliminary normalized image and the original features in the denoised image estimated by the estimation module, positioning the feature region of the preliminary normalized image; extracting features from a feature region in the preliminary normalized image.
In the above image restoration method, the displaying the restored image includes displaying control and displaying; the display control includes image depth analysis for analyzing the depth of field of the display image, including analyzing an average of the depths of field of the plurality of display images over a fixed period of time, and curvature control for controlling the curvature of the display unit in stages according to the average.
In summary, the image restoration apparatus and method provided by the present invention are adopted, and the beneficial effects of the present invention are: the original image which is printed with mosaic and stained or generates noise interference due to external environment is collected, preprocessed and restored by using the image restoration device, so that the original content of the image can be displayed, the condition that the visual experience of a user is not influenced by the existence of interference information on the original image is ensured, the image quality is improved, meanwhile, the display stability of a white point is excellent, the image displayed by the image restoration display module of the bending degree of the display module can be controlled according to the depth of field of the displayed image, the effect is better, and the visual experience of an observer can be improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention.
Fig. 1 is a schematic composition diagram of an image restoration apparatus according to the present invention.
Fig. 2 is a flowchart of an image restoration method provided by the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
As shown in fig. 1, an image restoration apparatus provided by the present invention includes an image acquisition module, an image preprocessing module, a target sample sampling module, a restoration processing module, and a restored image display module.
The connection relationship between the above modules of the present invention will be further described in detail with reference to the accompanying drawings.
In the above image restoration apparatus provided by the present invention, the image acquisition module is configured to acquire image information to be processed having a pixel loss area, the image acquisition module includes a camera and a controller, the camera sends the acquired image information to be processed having the pixel loss area to the controller, and implements pixel correction by using a correction method based on Gamma correction, and implements non-linear tone editing on an image by editing a Gamma curve of the image, thereby detecting a dark color portion and a light color portion in an image signal and increasing a ratio of the dark color portion and the light color portion, so as to improve an image contrast effect, and the Gamma correction is implemented by matlab.
In the image restoration device provided by the present invention, the image preprocessing module is electrically connected to the image acquisition module, and the image preprocessing module receives and preprocesses the image information acquired by the image acquisition module.
In the image restoration device provided by the present invention, the image preprocessing module includes an interference information removing unit, a denoising unit, and an image normalization unit;
the interference information removing unit comprises a defogging module, a stain removing module and a demosaicing module, wherein the defogging module is used for estimating and correcting the transmissivity in the foggy image model by using a dark channel prior method, performing defogging recovery on the original foggy image by using a defogging recovery formula and performing initial assignment on an ideal image, and the defogging recovery formula isWherein s is a pixel point to be processed, t is a transmittance, and t is0And the lower threshold of the transmissivity t is defined as A, global background light, i and O, wherein i is an original foggy image and O is a defogged image.
The stain removing module is used for inhibiting background interference by adjusting the gray level according to the local gray level fluctuation degree of the pixel neighborhood so as to realize self-adaptive removal of stains, and the dynamic adjustment of the gray level can be realized by a convolution operation method of weighted average of the pixel neighborhood;
the demosaicing module recovers the brightness and the chrominance signals of the current pixel position from the input original data by respectively conducting demosaicing processing on the brightness signal Y and the three chrominance signals R, G, B, and converts the brightness and the chrominance signals of the demosaiced current pixel position into corresponding original color gamut image signals through matrix operation;
the denoising unit obtains a preliminary denoising image by performing two-dimensional wiener filtering on an original image, calculates residual quantity of pixels according to the original image and the preliminary denoising image, calculates a weight matrix of each pixel by adopting a non-local mean filtering method according to the residual quantity, and performs non-local mean calculation on the image to be processed according to the weight matrix so as to realize denoising processing on the image to be processed and obtain a denoising image;
the image normalization unit is used for performing normalization processing on the denoised image to obtain a normalized image.
In the image restoration device provided by the present invention, the image normalization unit includes a detection module, a preliminary normalization module, an estimation module, a positioning module, and an extraction module, and the detection module is configured to detect a plurality of feature points in the denoised image; the preliminary normalization module normalizes the denoised image based on the plurality of feature points detected by the detection module, thereby obtaining a preliminary normalized image; the estimation module estimates original features in the denoised image; the positioning module positions a characteristic region of a preliminary normalized image based on normalized characteristic points of the preliminary normalized image and original characteristics in the denoised image estimated by the estimation module; the extraction module extracts features from a feature region in the preliminary normalized image.
In the image restoration device provided by the present invention, the target sample sampling module is electrically connected to the image preprocessing module, and the target sample sampling module performs sliding sampling on the image obtained by the image preprocessing module through a sliding window to obtain a target sample image corresponding to the target image.
In the above-described image restoration device according to the present invention, the restoration processing module is electrically connected to the target sample sampling module, and the restoration processing module performs restoration processing on the image by inputting the target sample image into a trained convolutional neural network model.
In the image restoration device provided by the present invention, the restoration processing module includes a training determining module, a first result judging module, a first adjusting module, a first convolution module, a second result judging module, an inverse normalization module, and a second adjusting module; the training module trains the characteristic parameters corresponding to the first convolution module and the second convolution module respectively through a data set containing original blurred image samples to obtain a first characteristic parameter set corresponding to the first convolution module and a second characteristic parameter set corresponding to the second convolution module, wherein the data set containing the original blurred image samples comprises a training set and a testing set, the training set is used for training a model, and the testing set is used for testing the model obtained by training and is used for approximating the generalization capability of the model; the first result judging module calculates loss by using a cross entropy function to obtain accuracy, judges whether the accuracy reaches a specific value or not, and controls the training module to stop training when the accuracy reaches the specific value; the determination module is configured to determine whether a resolution of the target sample image is consistent with an input dimension of the first convolution module; the first adjusting module is used for carrying out scaling processing on the target sampling sample image to obtain a first image when the resolution of the target sampling sample image is inconsistent with the input dimension of the first convolution module; the first convolution module is a positive convolution module, and the first convolution module performs convolution processing on the first image based on the first characteristic parameter set obtained by training of the training module to obtain a first characteristic of the first image; the second convolution module is a deconvolution module, and the second convolution module performs deconvolution processing on the first feature representation of the first image obtained by the first convolution module based on the second feature parameter set obtained by the training of the training module to obtain a second image; the second result judging module is used for judging whether the restoration error between the original images corresponding to the second image reaches a preset value or not, and if the restoration error reaches the preset value, the convolution process is stopped; the reverse normalization module is used for performing reverse normalization processing on the obtained second image to obtain a first restored image; the second adjusting module is used for adjusting the resolution of the first restoration image so that the resolution of the second image is the same as the resolution of the original blurred image.
In the image restoration device according to the present invention, the restoration image display module is electrically connected to the restoration processing module, and the restoration image display module receives and displays an output restoration image of the restoration processing module.
In the image restoration device according to the present invention, the restored image display module includes a display control unit and a display unit; the display control unit comprises an image depth of field analysis module and a bending degree control module, wherein the image depth of field analysis module is used for analyzing the depth of field of the display image, the image depth of field analysis module comprises an average value of the depth of field of a plurality of display images in a fixed time period, and the bending degree control module controls the bending degree of the display unit in a grading mode according to the average value; the display unit includes a light emitting device including a nitride semiconductor light emitting element, a red phosphor and a green phosphor, and a color filter having a blue pixel, a red pigment and a green pigment, and an image display device configured by combining the light emitting device and the color filter can realize higher white luminance and higher color purity, thereby being capable of reducing color difference of a displayed image.
As shown in fig. 2, the image restoration method provided by the present invention includes the following steps:
in step S1, to-be-processed image information having a pixel loss region is acquired;
in step S2, performing interference information unit, denoising, and normalization processing on the image to obtain a normalized image, where the normalization processing includes detecting a plurality of feature points in the denoised image; normalizing the denoised image based on the plurality of feature points detected by the detection module, thereby obtaining a preliminary normalized image; an estimation module estimates original features in the denoised image; based on the normalized feature points of the preliminary normalized image and the original features in the denoised image estimated by the estimation module, positioning the feature region of the preliminary normalized image; extracting features from a feature region in the preliminary normalized image;
in step S3, sliding sampling is performed on the normalized image to obtain a target sampling sample image;
in step S4, training feature parameters corresponding to the first convolution module and the second convolution module, respectively;
in step S5, calculating the loss to obtain an accuracy, and determining the accuracy to control the training process;
in step S6, the resolution of the target sample image is adjusted to be consistent with the input dimension of the first convolution module, so as to obtain a first image;
in step S7, performing convolution processing and deconvolution processing on the first image to obtain a second image;
in step S8, it is determined whether the recovery error reaches a predetermined value, and if not, the procedure returns to the step of training the characteristic parameters corresponding to the first convolution module and the second convolution module, and the above steps are repeated, and if the recovery error reaches the predetermined value, the convolution process is stopped;
in step S9, the second image is subjected to the inverse normalization processing and the resolution adjustment;
in step S10, displaying a restoration image including display control and display; the display control includes image depth analysis for analyzing the depth of field of the display image, including analyzing an average of the depths of field of the plurality of display images over a fixed period of time, and curvature control for controlling the curvature of the display unit in stages according to the average.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes may be made to the embodiment of the present invention by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. An image restoration apparatus, comprising: the device comprises an image acquisition module, an image preprocessing module, a target sample sampling module, a restoration processing module and a restoration image display module;
the image acquisition module is used for acquiring to-be-processed image information with a pixel loss area;
the image preprocessing module is electrically connected with the image acquisition module, receives image information acquired by the image acquisition module and preprocesses the image information;
the target sample sampling module is electrically connected with the image preprocessing module and is used for performing sliding sampling on the image obtained by the image preprocessing module through a sliding window so as to obtain a target sampling sample image corresponding to the target image;
the restoration processing module is electrically connected with the target sample sampling module, and the restoration processing module carries out restoration processing on the image by inputting the target sample image into a trained convolutional neural network model;
the restoration image display module is electrically connected with the restoration processing module, and receives the output restoration image of the restoration processing module for display.
2. The image restoration device according to claim 1, wherein the image acquisition module comprises a camera and a controller, the camera sends acquired image information to be processed with pixel loss areas to the controller, pixel correction is realized by a correction method based on Gamma correction, and an image contrast effect is improved by detecting a dark color part and a light color part in an image signal and increasing the ratio of the dark color part and the light color part.
3. The image restoration device according to claim 1, wherein the image preprocessing module comprises a de-interference information unit, a de-noising unit and an image normalization unit, the de-interference information unit comprises a defogging module, a stain removal module and a de-mosaic module, the defogging module is used for estimating and correcting the transmittance in the foggy image model by using a dark channel prior method, defogging and restoring the original foggy image by using a defogging and restoring formula and initially assigning an ideal image, the stain removal module is used for suppressing background interference by adjusting gray levels according to local gray level fluctuation degrees of pixel neighborhoods to realize self-adaptive stain removal, the dynamic adjustment of the gray levels can be realized by a convolution operation method of weighted average of the pixel neighborhoods, and the de-mosaic module is used for respectively carrying out weighted average on a luminance signal Y and three chrominance signals R, B and C, G. B, demosaicing processing is carried out, luminance and chrominance signals of the current pixel position are recovered from input original data, and the luminance and chrominance signals of the demosaiced current pixel position are converted through matrix operation to obtain corresponding original color gamut image signals; the denoising unit obtains a preliminary denoising image by performing two-dimensional wiener filtering on an original image, calculates residual quantity of pixels according to the original image and the preliminary denoising image, calculates a weight matrix of each pixel by adopting a non-local mean filtering method according to the residual quantity, and performs non-local mean calculation on the image to be processed according to the weight matrix so as to realize denoising processing on the image to be processed and obtain a denoising image; the image normalization unit is used for performing normalization processing on the denoised image to obtain a normalized image.
4. The image restoration device according to claim 3, wherein the image normalization unit comprises a detection module, a preliminary normalization module, an estimation module, a positioning module and an extraction module, wherein the detection module is used for detecting a plurality of feature points in the denoised image; the preliminary normalization module normalizes the denoised image based on the plurality of feature points detected by the detection module, thereby obtaining a preliminary normalized image; the estimation module estimates original features in the denoised image; the positioning module positions a characteristic region of a preliminary normalized image based on normalized characteristic points of the preliminary normalized image and original characteristics in the denoised image estimated by the estimation module; the extraction module extracts features from a feature region in the preliminary normalized image.
5. The image restoration device according to claim 1, wherein the restoration processing module includes a determination training module, a first result judgment module, a first adjustment module, a first convolution module, a second result judgment module, an inverse normalization module, and a second adjustment module; the training module trains the characteristic parameters corresponding to the first convolution module and the second convolution module respectively through a data set containing original blurred image samples to obtain a first characteristic parameter set corresponding to the first convolution module and a second characteristic parameter set corresponding to the second convolution module, wherein the data set containing the original blurred image samples comprises a training set and a testing set, the training set is used for training a model, and the testing set is used for testing the model obtained by training and is used for approximating the generalization capability of the model; the first result judging module calculates loss by using a cross entropy function to obtain accuracy, judges whether the accuracy reaches a specific value or not, and controls the training module to stop training when the accuracy reaches the specific value; the determination module is configured to determine whether a resolution of the target sample image is consistent with an input dimension of the first convolution module; the first adjusting module is used for carrying out scaling processing on the target sampling sample image to obtain a first image when the resolution of the target sampling sample image is inconsistent with the input dimension of the first convolution module; the first convolution module is a positive convolution module, and the first convolution module performs convolution processing on the first image based on the first characteristic parameter set obtained by training of the training module to obtain a first characteristic of the first image; the second convolution module is a deconvolution module, and the second convolution module performs deconvolution processing on the first feature representation of the first image obtained by the first convolution module based on the second feature parameter set obtained by the training of the training module to obtain a second image; the second result judging module is used for judging whether the restoration error between the original images corresponding to the second image reaches a preset value or not, and if the restoration error reaches the preset value, the convolution process is stopped; the reverse normalization module is used for performing reverse normalization processing on the obtained second image to obtain a first restored image; the second adjusting module is used for adjusting the resolution of the first restoration image so that the resolution of the second image is the same as the resolution of the original blurred image.
6. The image restoration device according to claim 1, wherein the restoration image display module includes a display control unit and a display unit; the display control unit comprises an image depth of field analysis module and a bending degree control module, wherein the image depth of field analysis module is used for analyzing the depth of field of the display image, the image depth of field analysis module comprises an average value of the depth of field of a plurality of display images in a fixed time period, and the bending degree control module controls the bending degree of the display unit in a grading mode according to the average value; the display unit includes a light emitting device including a nitride semiconductor light emitting element, a red phosphor and a green phosphor, and a color filter having a blue pixel, a red pigment and a green pigment, and an image display device configured by combining the light emitting device and the color filter can realize higher white luminance and higher color purity, thereby being capable of reducing color difference of a displayed image.
7. An image restoration method, comprising the steps of:
acquiring to-be-processed image information with a pixel loss area;
carrying out interference information unit, denoising and normalization processing on the image to obtain a normalized image;
performing sliding sampling on the normalized image to obtain a target sampling sample image;
training the characteristic parameters corresponding to the first convolution module and the second convolution module respectively;
calculating loss to obtain accuracy, and judging the accuracy to control the training process;
adjusting the resolution of the target sampling sample image to be consistent with the input dimension of the first convolution module to obtain a first image;
performing convolution processing and deconvolution processing on the first image to obtain a second image;
judging whether the recovery error reaches a preset value or not, if not, returning to the step of training the characteristic parameters corresponding to the first convolution module and the second convolution module, repeating the steps, and if so, stopping the convolution process;
performing inverse normalization processing and resolution adjustment on the second image;
and displaying the restored image.
8. The image restoration method according to claim 7, wherein the normalization process includes detecting a plurality of feature points in a denoised image; normalizing the denoised image based on the plurality of feature points detected by the detection module, thereby obtaining a preliminary normalized image; an estimation module estimates original features in the denoised image; based on the normalized feature points of the preliminary normalized image and the original features in the denoised image estimated by the estimation module, positioning the feature region of the preliminary normalized image; extracting features from a feature region in the preliminary normalized image.
9. The image restoration method according to claim 7, wherein the displaying the restored image includes display control and display; the display control includes image depth analysis for analyzing the depth of field of the display image, including analyzing an average of the depths of field of the plurality of display images over a fixed period of time, and curvature control for controlling the curvature of the display unit in stages according to the average.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011225419.9A CN112184597A (en) | 2020-11-05 | 2020-11-05 | Image restoration device and method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011225419.9A CN112184597A (en) | 2020-11-05 | 2020-11-05 | Image restoration device and method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112184597A true CN112184597A (en) | 2021-01-05 |
Family
ID=73916520
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011225419.9A Pending CN112184597A (en) | 2020-11-05 | 2020-11-05 | Image restoration device and method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112184597A (en) |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104093015A (en) * | 2014-06-23 | 2014-10-08 | 京东方科技集团股份有限公司 | Display system and control method thereof |
CN105335947A (en) * | 2014-05-26 | 2016-02-17 | 富士通株式会社 | Image de-noising method and image de-noising apparatus |
CN105844604A (en) * | 2016-05-19 | 2016-08-10 | 湖南源信光电科技有限公司 | Fast defogging algorithm based on local histogram enhancement |
CN105989331A (en) * | 2015-02-11 | 2016-10-05 | 佳能株式会社 | Facial feature extraction apparatus, facial feature extraction method, image processing equipment and image processing method |
CN106530227A (en) * | 2016-10-27 | 2017-03-22 | 北京小米移动软件有限公司 | Image restoration method and device |
CN106557750A (en) * | 2016-11-22 | 2017-04-05 | 重庆邮电大学 | It is a kind of based on the colour of skin and the method for detecting human face of depth y-bend characteristics tree |
CN107112395A (en) * | 2014-10-14 | 2017-08-29 | 大日本印刷株式会社 | Device used for image display module and image display device |
CN108364262A (en) * | 2018-01-11 | 2018-08-03 | 深圳大学 | A kind of restored method of blurred picture, device, equipment and storage medium |
CN110223258A (en) * | 2019-06-12 | 2019-09-10 | 西南科技大学 | A kind of multi-mode fast video image defogging method and device |
CN111429360A (en) * | 2020-01-14 | 2020-07-17 | 湖南城市学院 | Dual-display device for graphic image processing and control method thereof |
CN111539893A (en) * | 2020-04-28 | 2020-08-14 | 南京图格医疗科技有限公司 | Bayer image joint demosaicing denoising method based on guided filtering |
-
2020
- 2020-11-05 CN CN202011225419.9A patent/CN112184597A/en active Pending
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105335947A (en) * | 2014-05-26 | 2016-02-17 | 富士通株式会社 | Image de-noising method and image de-noising apparatus |
CN104093015A (en) * | 2014-06-23 | 2014-10-08 | 京东方科技集团股份有限公司 | Display system and control method thereof |
CN107112395A (en) * | 2014-10-14 | 2017-08-29 | 大日本印刷株式会社 | Device used for image display module and image display device |
CN105989331A (en) * | 2015-02-11 | 2016-10-05 | 佳能株式会社 | Facial feature extraction apparatus, facial feature extraction method, image processing equipment and image processing method |
CN105844604A (en) * | 2016-05-19 | 2016-08-10 | 湖南源信光电科技有限公司 | Fast defogging algorithm based on local histogram enhancement |
CN106530227A (en) * | 2016-10-27 | 2017-03-22 | 北京小米移动软件有限公司 | Image restoration method and device |
CN106557750A (en) * | 2016-11-22 | 2017-04-05 | 重庆邮电大学 | It is a kind of based on the colour of skin and the method for detecting human face of depth y-bend characteristics tree |
CN108364262A (en) * | 2018-01-11 | 2018-08-03 | 深圳大学 | A kind of restored method of blurred picture, device, equipment and storage medium |
CN110223258A (en) * | 2019-06-12 | 2019-09-10 | 西南科技大学 | A kind of multi-mode fast video image defogging method and device |
CN111429360A (en) * | 2020-01-14 | 2020-07-17 | 湖南城市学院 | Dual-display device for graphic image processing and control method thereof |
CN111539893A (en) * | 2020-04-28 | 2020-08-14 | 南京图格医疗科技有限公司 | Bayer image joint demosaicing denoising method based on guided filtering |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110730336B (en) | Demosaicing method and device | |
JP5213670B2 (en) | Imaging apparatus and blur correction method | |
CN108389215B (en) | Edge detection method and device, computer storage medium and terminal | |
CN107451969A (en) | Image processing method, device, mobile terminal, and computer-readable storage medium | |
CN111861896A (en) | A UUV-Oriented Color Compensation and Restoration Method for Underwater Images | |
TWI779948B (en) | Lens dirt detection method for camera module | |
CN109087254A (en) | Unmanned plane image haze sky and white area adaptive processing method | |
CN107464225A (en) | Image processing method, device, computer readable storage medium and mobile terminal | |
CN107317967A (en) | Image processing method, device, mobile terminal, and computer-readable storage medium | |
Peng et al. | Single image restoration using scene ambient light differential | |
US9258460B2 (en) | Image processing method and apparatus for correcting an image | |
CN103313068B (en) | White balance corrected image processing method and device based on gray edge constraint gray world | |
CN111836103B (en) | Anti-occlusion processing system based on data analysis | |
CN110782400B (en) | Self-adaptive illumination uniformity realization method and device | |
CN112435184A (en) | Haze sky image identification method based on Retinex and quaternion | |
CN105809677B (en) | Image edge detection method and system based on bilateral filter | |
CN107277369B (en) | Image processing method, device, computer readable storage medium and computer equipment | |
Jiwani et al. | Single image fog removal using depth estimation based on blur estimation | |
CN117408906B (en) | Low-light image enhancement method and system | |
CN107392870B (en) | Image processing method, image processing device, mobile terminal and computer readable storage medium | |
CN107424134A (en) | Image processing method, device, computer readable storage medium and computer equipment | |
CN112184597A (en) | Image restoration device and method | |
CN107317968A (en) | Image defogging method, device, computer storage medium and mobile terminal | |
JP2018160024A (en) | Image processing device, image processing method and program | |
CN114120167A (en) | A method and device for video image exposure measurement and automatic correction |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210105 |
|
RJ01 | Rejection of invention patent application after publication |