JPWO2019041842A5 - - Google Patents

Description

This application claims the priority of the Chinese Patent Application No. 201710762836.9, entitled "Method and Apparatus for Image Processing", filed with State Intellectual Property Office of China on Aug. 28, 2017, all of which the contents of which are incorporated herein by reference.

TECHNICAL FIELD The present invention relates to the technical field of image processing, and more particularly to an image processing method and device, a storage medium, and a computer device.

Histogram equalization is the process of changing the tone histogram of the original image from one relatively concentrated tonal range to be evenly distributed over the entire tonal range, and then performing histogram equalization on the image to restore the image. This is an image processing method that enhances the contrast of the image and makes the image clearer.

However, since video images have quasi-flat areas, the histogram equalization process can degrade the quality of the quasi-flat areas of the image, significantly degrading the quality of the image. In video images, false flat areas result from overcompression of the video image during processing and transmission. A quasi-flat portion is an area in an image where a quasi-flat state exists, and the area of the quasi-flat state appears to have the same grayscale, but is composed of grayscales that are close to the actual grayscale, for example, in a video image. The sky-like scene in the image forms a block-like region of contiguous tones after being compressed. When histogram equalization is performed on such a block-shaped pseudo-flat image, the gradation difference in the pseudo-flat portion increases, and as a result, an irregular block shape visible to the naked eye occurs in the pseudo-flat portion. The image quality of the quasi-flat portions in the image is degraded, and the image quality is significantly reduced.

SUMMARY OF THE INVENTION In order to solve the problem that the image quality of the pseudo-flat portion deteriorates and the image quality of the image deteriorates when histogram equalization is performed in the prior art, the object of the present invention is to provide an image processing method and apparatus, a storage medium and a computer apparatus. is to provide

One aspect of the image processing method according to the present invention is
generating an original grayscale image of the original image;
performing a histogram equalization process on the original grayscale image to obtain an equalized grayscale image;
Based on the original grayscale image, a first marking area including an area in which pixel positions in the original image are adjacent and the standard deviation is smaller than a set value, and an area other than the first marking area. generating a determinant map including the second marking area;
Based on the original grayscale image, the equalized grayscale image, and the determinant distribution diagram, the grayscale value of the pixel corresponding to the second marking area is the grayscale of the corresponding pixel in the equalized grayscale image. obtaining a final grayscale image, wherein grayscale values of pixels corresponding to the first marking area are grayscale values of corresponding pixels in the original grayscale image;
Reconstructing a processed image based on the final grayscale image.

In one aspect of the image processing method according to the present invention, generating a determining factor distribution map based on the original grayscale image includes:
Each pixel in the original grayscale image is calculated by calculating the standard deviation of each pixel in the original grayscale image based on the gradation values of pixels in a specific area centered on each pixel in the original grayscale image. obtaining a standard deviation distribution map containing the standard deviation of the pixels;
Pixels having a standard deviation smaller than the set value in the standard deviation distribution map are divided into the first marking regions, and regions other than the first marking region in the standard deviation distribution map are divided into the second marking regions. Obtaining the determinant distribution map by partitioning.

In another aspect of the image processing method according to the present invention, the standard deviation of each pixel of the original grayscale image is calculated based on the gradation values of the pixels within a specific region centered on each pixel of the original grayscale image. Obtaining the standard deviation distribution map by calculating
forming a tone-square diagram by calculating the square of the tone value of each pixel of the original grayscale image;
generating a first expectation map by performing mean filtering on the original grayscale image, and generating a second expectation map by performing mean filtering on the tone-square map;
obtaining a third expectation map by calculating the square value of the gradation value of each pixel in the first expectation map;
obtaining a difference image by calculating a difference in grayscale values of pixels at corresponding positions in the second expected map and the third expected map;
Obtaining the standard deviation distribution map by obtaining the square root of the gradation value of each pixel of the difference image.

In another aspect of the image processing method according to the present invention, performing average filtering on the original grayscale image comprises:
performing mean filtering on the original grayscale image using a filtering template of size m×m, wherein the value of m ranges from 10 to 20;
Performing average filtering on the tone-squared map includes:
performing mean filtering on the tone-squared map using the m×m sized filtering template.

In another aspect of the image processing method according to the invention, m is fifteen.

In another aspect of the image processing method according to the present invention, after generating the determinant distribution map based on the original grayscale image, the image processing method includes:
By updating the determinant factor distribution map, in the updated determinant factor distribution map, the second marking area includes pixels whose pixel positions in the original image are adjacent and whose standard deviation is equal to or greater than the set value. The method further includes a certain region and a region in which the number of pixels in the region is smaller than the threshold value and the pixel positions in the original image are adjacent to each other and the standard deviation is smaller than the set value.

In another aspect of the image processing method according to the present invention, the threshold is 8% to 15% of the number of pixels in the original image.

In another aspect of the image processing method according to the present invention, the threshold is 10% of the number of pixels in the original image.

In another aspect of the image processing method according to the present invention, the set value ranges from 1-5.

In another aspect of the image processing method according to the present invention, the set value is one.

One aspect of the image processing apparatus according to the present invention is
a generation circuit for generating an original grayscale image of the original image;
a histogram equalization circuit for performing a histogram equalization process on the original grayscale image to obtain an equalized grayscale image;
Based on the original grayscale image, a first marking area including an area in which pixels in the original image have adjacent pixel positions and a standard deviation is smaller than a set value, and an area other than the first marking area. a first processing circuit for generating a determinant map including a second marking area;
Based on the original grayscale image, the equalized grayscale image, and the determinant distribution diagram, the grayscale value of the pixel corresponding to the second marking area is the grayscale of the corresponding pixel in the equalized grayscale image. a second processing circuit for obtaining a final grayscale image in which the grayscale values of the pixels corresponding to the first marking area are the grayscale values of the corresponding pixels in the original grayscale image;
A third processing circuit is included for reconstructing a processed image based on said final grayscale image.

In one aspect of the image processing device according to the present invention,
The first processing circuit calculates the standard deviation of each pixel of the original grayscale image based on the gradation values of pixels in a specific area centered on each pixel of the original grayscale image. , obtaining a standard deviation distribution map containing the standard deviation of each pixel in the original grayscale image,
Pixels having a standard deviation smaller than the set value in the standard deviation distribution map are divided into the first marking regions, and regions other than the first marking region in the standard deviation distribution map are divided into the second marking regions. By dividing, the determinant distribution map is obtained.

In another aspect of the image processing device according to the present invention, the first processing circuit
a first computational subcircuit for forming a tone-squared diagram by computing the squared tone value of each pixel of the original grayscale image;
A mean filtering sub-sub which generates a first expectation map by performing mean filtering on the original grayscale image and generates a second expectation map by performing mean filtering on the tone-square map. a circuit;
a second calculation subcircuit for obtaining a third expectation map by calculating the square value of the grayscale value of each pixel of the first expectation map;
a third calculation sub-circuit for obtaining a difference image by calculating differences in grayscale values of pixels at corresponding positions in the second expectation map and the third expectation map;
A fourth calculation sub-circuit for obtaining the standard deviation distribution map by obtaining the square root of the gradation value of each pixel of the difference image.

In another aspect of the image processing apparatus according to the present invention, the averaging filtering subcircuit uses a filtering template of size m×m to respectively average Filtering is performed and the value of m ranges from 10-20.

In another aspect of the image processing device according to the invention, m is fifteen.

In another aspect of the image processing device according to the present invention, the first processing circuit generates the determinant factor distribution map based on the original grayscale image, and then updates the determinant factor distribution map, thereby: In the updated determinant factor distribution map, the second marking area includes an area having pixels adjacent to each other in the original image and having a standard deviation equal to or greater than the set value, and A region in which the number of pixels in a region containing adjacent pixels whose standard deviation is smaller than the set value is smaller than the threshold is included.

In another aspect of the image processing device according to the present invention, the threshold is 8% to 15% of the number of pixels in the original image.

In another aspect of the image processing device according to the present invention, the threshold is 10% of the number of pixels of the original image.

In another aspect of the image processing device according to the present invention, the range of the setting value is 1-5.

In another aspect of the image processing device according to the present invention, the set value is one.

In one aspect of the storage medium according to the present invention, at least one instruction is stored on the storage medium, and the instruction is loaded and executed by a processor to cause the operation performed in any of the image processing methods described above. achieve

In one aspect of the invention, there is provided a computer apparatus comprising a memory, a processor and a computer program stored in said memory and running on said processor, said processor, when said computer program is executed, produces an image as described above. configured to accomplish the operations performed in the processing method;

The technical solutions provided by the embodiments of the present invention have the following beneficial effects.

In an embodiment of the present invention, an original grayscale image and an equalized grayscale image are first generated based on the original image, and then a determinant distribution map is generated based on the original grayscale image. In the determinant factor distribution diagram, the first marking area includes an area in which the pixel positions in the original image are adjacent to each other and the standard deviation is smaller than the set value. Since the gradation difference in the pseudo-flat portion is small, the standard deviation of the pixels in the pseudo-flat portion is also small, so the area corresponding to the first marking area in the original image is the pseudo-flat area. The pixel tone of a grayscale image in which the pseudo-flat area is not expanded, i.e. not histogram equalized, is represented by the pixel tone values of the original grayscale image, and the second marking area is processed by histogram equalization. It is expressed in terms of values, ie pixel tone values of the equalized grayscale image. In such a processing method, since the pseudo-flat portion is not expanded, the gradation difference of the pseudo-flat portion is not enlarged, the image quality deterioration of the pseudo-flat portion in the image is avoided, and the image quality of the image is ensured.

In order to describe the technical solutions of the embodiments of the present invention more clearly, the following drawings used in the description of the embodiments are briefly described. The drawings in the following description are some embodiments of the present invention, and those skilled in the art can easily obtain other drawings without creative work.

4 is a flow chart of an image processing method according to an embodiment of the present invention; FIG. 4 is a flow chart of generating a determinant distribution map in an embodiment of the present invention; FIG. 1 is a structural diagram of an image processing device according to an embodiment of the present invention; FIG. 3 is a structural diagram of a first processing circuit in an embodiment of the present invention; FIG. 1 is a structural diagram of an image processing device according to an embodiment of the present invention; FIG.

To make the objectives, technical solutions and advantages of the present invention clearer, the embodiments of the present invention are described in more detail below with reference to the accompanying drawings.

FIG. 1 is a flow chart of an image processing method according to an embodiment of the present invention. Referring to FIG. 1, the image processing method comprises:
It includes a step 101 of generating an original grayscale image of the original image.

Here, the original image may be an image obtained by various methods such as an image obtained by photographing or a video image. The original image is a format image of red (Red, R), green (Green, G), and blue (Blue, B).

In embodiments of the present invention, generating a source grayscale image of the source image includes obtaining the source image and processing the source image to obtain the source grayscale image.

Here, processing the original image includes, but is not limited to, generating the original grayscale image by converting the original image into a format such as YUV. In YUV format, Y represents luminance and U and V represent chrominance, so if there is only a Y component and no U and V components, the acquired image is a grayscale image. Step 101 can perform an image transformation on this basis to generate the original grayscale image.

At step 102, a histogram equalization process is performed on the original grayscale image to obtain an equalized grayscale image.

In embodiments of the present invention, performing a histogram equalization process on the original grayscale image can be achieved according to the following steps.

(1) Statistics the number of pixels n _i of each gray level in the original grayscale image, i = 0, 1, ..., L-1, where L is the total number of gray levels in the original grayscale image, The total number of gray levels in a grayscale image is the number of gray levels represented by a pixel in the original grayscale image, eg, L=256, where i ranges from 0-255.

(2) Compute the original image histogram, ie, the probability density P _i (r _i ) for each gray level in the original grayscale image.
P _i (r _i ) = n _i /n (1)
In equation (1), n is the total number of pixels in the original grayscale image, r _i is the ith gray level, and n _i is the number of pixels at the ith gray level. The original image histogram can be obtained by sequentially calculating the probability density of each gray level with the formula.

The horizontal axis of the original image histogram indicates the gray level, the vertical axis indicates the probability density, and the original image histogram contains the probability density of each gray level.

(3) Calculate the cumulative distribution function s _k (r _k ) of the original grayscale image.

In equation (2), r _k is the kth gray level, and the cumulative distribution function value of the kth gray level is the sum of the probability densities of the 0th to kth gray levels.

(4) Calculate the gray level gk in the equalized grayscale image of each pixel of the original grayscale image according to the cumulative distribution function s _k (r _k ) of the original grayscale image.
gk=INT[gm*s _k (r _k )] (3)
In equation (3), INT[] means rounding down the decimal point to an integer, gm=L-1.

For a pixel with gray level k in the original grayscale image, the gray level in the equalized grayscale image is gk. In this way, it is possible to calculate the equalized gray level of each pixel in the original grayscale image.

(5) modifying the gray level of the original gray scale image according to the calculated gray level of each pixel of the original gray scale image in the equalized gray scale image to obtain the equalized gray scale image;

In step 103, based on the original grayscale image, a first marking area including an area having pixels whose pixel positions in the original image are adjacent and whose standard deviation is smaller than a set value, and a marking area other than the first marking area. Generate a determinant map including a second marking area that is the area of .

Here, in the present invention, the standard deviation is obtained by filtering the square value image of the gradation value of each pixel in the original grayscale image, and then filtering the original grayscale image and subtracting the squared image. This means that the square root is obtained for the gradation value of each pixel of the difference image. Here, filtering refers to calculating the average value of gradation values within a specific range around the pixel as the gradation value of the pixel. The standard deviation can reflect the tone difference between each pixel in the original grayscale image and its surrounding pixels.

FIG. 2 is a flow chart of step 103 in an embodiment of the present invention, see FIG. 2, step 103 includes:
Each pixel in the original grayscale image is calculated by calculating the standard deviation of each pixel in the original grayscale image based on the gradation values of pixels in a specific area centered on each pixel in the original grayscale image. It includes a step 103a of obtaining a standard deviation distribution map containing the standard deviation of the pixels.

As an example, step 103a includes forming a tone-squared map by calculating the square of the tone value of each pixel of the original grayscale image; A first expectation map and a second expectation map are respectively generated by performing average filtering on each of the original grayscale images, i.e., performing average filtering on the original grayscale image to generate the first expectation map and generating a second expectation map by performing average filtering on the grayscale squared map; and calculating the squared value of the grayscale value of each pixel in the first expectation map. obtaining a third expected map by calculating a difference in tone values of pixels at corresponding positions in the second expected map and the third expected map, thereby obtaining a difference image; and obtaining the standard deviation distribution map by obtaining the square root of the tone value of each pixel of the difference image.

The standard deviation distribution map obtained by this method can reflect the gradation difference between each pixel and its surrounding pixels. It can be divided into regions.

Wherein performing average filtering on the original grayscale image and the tone-squared map, respectively, uses a filtering template of size m×m to obtain the original grayscale image and the tone-squared image. Including performing mean filtering on each figure. Mean filtering refers to selecting, for a pending pixel (x,y), a filtering template consisting of pixels adjacent to the pending pixel. Find the average grayscale value of all pixels in the filtering template, then average grayscale at pixel (x,y) and average grayscale at pixel (x,y) pending processing as the grayscale of the filtered image. assign a value. A filtering template is used to limit the number of pixels for calculating the grayscale average value. When a filtering template is used, the pixel awaiting processing is located at the center (at or near the center of the pixel) of the pixel selected by the filtering template. The range of values for m is 10-20.

By using m in this range, the standard deviation distribution map can better indicate pseudo-flat regions in the original image. As a result, the quasi-flat area is not expanded, and image quality deterioration is avoided.

As an example, m is fifteen. Using a value of 15 for m allows maximum discrimination of pseudo-flat regions in the original image.

Here, the second expected map and the third expected map have the same number of rows and columns, and the pixel at the corresponding position means the same row in the second expected map and the third expected map. Refers to pixels in columns.

In step 103b, pixels whose standard deviation in the standard deviation distribution map is smaller than a set value are divided into the first marking regions, and regions other than the first marking region in the standard deviation distribution map are divided into the second marking regions. By dividing into marking regions, the determinant distribution map is obtained.

During the generation of the determinant distribution map described above, first, the standard deviation distribution map of the original grayscale image is calculated, and the first marking area is divided according to the relationship between the standard deviation of each pixel in the standard deviation distribution map and the set value. Since the gradation difference in the quasi-flat portion is small, the standard deviation of the pixels in the quasi-flat portion is also small . marking area), and the remaining area is determined as the second marking area.

Furthermore, when dividing the first marking area and the second marking area, it can be realized by a binarization method. As an example, the process of dividing the first marking area and the second marking area is as follows. That is, pixels with standard deviations smaller than the set value in the standard deviation distribution map set the corresponding pixel values in the determinant distribution map to 0, and pixels with standard deviations greater than the set value in the standard deviation distribution map are the determinant distribution. In the figure, the corresponding pixel value is set to 1, so that pixels with a pixel value of 0 in the determinant distribution map belong to the first marking area, and pixels with a pixel value of 1 belong to the second marking area.

After dividing the area using the binarization method, in subsequent processing, it is determined whether the pixel belongs to the first marking area or the second marking area according to the pixel value of the determinant distribution map. It's easy and convenient.

In an embodiment of the present invention, the set value ranges from 1-5. The use of the range settings ensures that the determinant map better shows the quasi-flat regions in the original image. As a result, the quasi-flat area is not expanded, and image quality deterioration is avoided.

As an example, the setting value is one. By using 1 as the setting value, pseudo-flat regions in the original image can be distinguished maximally.

Optionally, after generating the determinant distribution map based on the original grayscale image, the image processing method comprises:
By updating the determinant factor distribution map, in the updated determinant factor distribution map, the second marking area includes pixels whose pixel positions in the original image are adjacent and whose standard deviation is equal to or greater than the set value. The method further includes a certain region and a region in which the number of pixels in the region is smaller than the threshold value and the pixel positions in the original image are adjacent to each other and the standard deviation is smaller than the set value.

During the updating of the determinant distribution map described above, a fill-in-the-blank algorithm is actually used, and the number of pixels in the region in which the pixel positions in the original image are adjacent and the pixels whose standard deviation is smaller than the set value is greater than the threshold value. Divide the smaller area into a second marking area, and during processing, perform the same processing on the hole and the pixels around it, so that only the perimeter of a region or the center of a region (hole) is extended and processed. It is possible to avoid sharp changes in contrast between the center and the periphery of a region in the captured image.

In an embodiment of the present invention, the threshold is 8%-15% of the number of pixels in the original image. By using a threshold for the range, the hole-filling algorithm will only run if there are sufficiently few pixels in the center of the region, so that regions with more than the threshold number of pixels will be processed by the hole-filling algorithm, It is possible to avoid the inability to optimize the image quality of pseudo-flat areas.

As an example, the threshold range is 10%. At this time, the area with the number of pixels above the threshold can be filled with a hole-filling algorithm to avoid the inability to optimize the image quality of the pseudo-flat area. Also, abrupt changes in contrast between the center and periphery of regions in the processed image can be avoided.

In step 104, based on the original grayscale image, the equalized grayscale image and the determinant distribution diagram, the grayscale value of the pixel corresponding to the second marking area is the corresponding value in the equalized grayscale image. and the grayscale values of the pixels corresponding to the first marking area are the grayscale values of the corresponding pixels in the original grayscale image to obtain a final grayscale image.

At step 104, the final grayscale image can be calculated according to the following equation.
I(i,j)=G(i,j)*H(i,j)+(1-G(i,j))*A(i,j) (4)
In equation (4), I is the final grayscale image and G is the determinant map (pixels in the first marking area are 0 and pixels in the second marking area are 1) , H is the equalized grayscale image and A is the original grayscale image. where I(i,j) is the pixel at row i, column j in the final grayscale image, G(i,j) is the pixel at row i, column j in the determinant map, and H(i, j) is the pixel at row i and column j in the equalized grayscale image, and A(i, j) is the pixel at row i and column j in the original grayscale image. As a result of calculation according to the above formula, the gradation value of the pixel corresponding to the second marking area in I is the gradation value of the corresponding pixel in H, and the gradation value of the pixel corresponding to the first marking area in I is The grayscale value is the grayscale value of the corresponding pixel in A.

At step 105, the processed image is reconstructed based on the final grayscale image.

In this step, the processed original image is obtained by restoring the final grayscale image to an R (red) G (green) B (blue) image. That is, the YUV format image is restored to an RGB image.

In an embodiment of the present invention, an original grayscale image and an equalized grayscale image are first generated based on the original image, and then a determinant distribution map is generated based on the original grayscale image. In the determinant factor distribution diagram, the first marking area includes an area in which the pixel positions in the original image are adjacent to each other and the standard deviation is smaller than the set value. Since the gradation difference in the pseudo-flat portion is small, the standard deviation of the pixels in the pseudo-flat portion is also small, so the area corresponding to the first marking area in the original image is the pseudo-flat area. The pixel tone of a grayscale image in which the pseudo-flat area is not expanded, i.e. not histogram equalized, is represented by the pixel tone values of the original grayscale image, and the second marking area is processed by histogram equalization. It is expressed in terms of values, ie pixel tone values of the equalized grayscale image. In such a processing method, since the pseudo-flat portion is not expanded, the gradation difference of the pseudo-flat portion is not enlarged, the image quality deterioration of the pseudo-flat portion in the image is avoided, and the image quality of the image is ensured.

FIG. 3 is a structural diagram of an image processing device according to an embodiment of the present invention. Referring to FIG. processing circuit 204 and a third processing circuit 205 .

Here, the generating circuit 201 generates an original grayscale image of the original image. A histogram equalization circuit 202 performs a histogram equalization process on the original grayscale image to obtain an equalized grayscale image. Based on the original grayscale image, the first processing circuit 203 creates a first marking area including an area in which pixels in the original image have adjacent pixel positions and have a standard deviation smaller than a set value, and the first marking area. generating a determinant distribution map including a second marking area that is an area other than the marking area of . The second processing circuit 204, based on the original grayscale image, the equalized grayscale image, and the determinant distribution diagram, determines the gradation value of the pixel corresponding to the second marking area as the equalized grayscale image. The grayscale value of the corresponding pixel in the scale image, wherein the grayscale value of the pixel corresponding to the first marking area is the grayscale value of the corresponding pixel in the original grayscale image. obtain. A third processing circuit 205 restores the processed image based on the final grayscale image.

Here, the generation circuit 201, the histogram equalization circuit 202, the first processing circuit 203, the second processing circuit 204, and the third processing circuit 205 are all realized by one individual circuit. A circuit may be a chip or an integrated circuit, such as a central processing unit (CPU), graphics processing unit (GPU), and the like. Any two or more of the generation circuit 201, the histogram equalization circuit 202, the first processing circuit 203, the second processing circuit 204, and the third processing circuit 205 may be realized by the same circuit. , the histogram equalization circuit 202, the first processing circuit 203, the second processing circuit 204 and the third processing circuit 205 are implemented by the same graphics processor.

In an embodiment of the present invention, an original grayscale image and an equalized grayscale image are first generated based on the original image, and then a determinant distribution map is generated based on the original grayscale image. In the determinant factor distribution diagram, the first marking area includes an area in which the pixel positions in the original image are adjacent to each other and the standard deviation is smaller than the set value. Since the gradation difference in the pseudo-flat portion is small, the standard deviation of the pixels in the pseudo-flat portion is also small, so the area corresponding to the first marking area in the original image is the pseudo-flat area. The pixel tone of a grayscale image in which the pseudo-flat area is not expanded, i.e. not histogram equalized, is represented by the pixel tone values of the original grayscale image, and the second marking area is processed by histogram equalization. It is expressed in terms of values, ie pixel tone values of the equalized grayscale image. In such a processing method, since the pseudo-flat portion is not expanded, the gradation difference of the pseudo-flat portion is not enlarged, the image quality deterioration of the pseudo-flat portion in the image is avoided, and the image quality of the image is ensured.

In the embodiment of the present invention, the first processing circuit 203 calculates each pixel of the original grayscale image based on the gradation values of pixels within a specific area centered on each pixel of the original grayscale image. By calculating the standard deviation of , a standard deviation distribution map including the standard deviation of each pixel in the original grayscale image is obtained. Pixels with a standard deviation smaller than a set value in the standard deviation distribution diagram are divided into the first marking area, and areas other than the first marking area in the standard deviation distribution diagram are divided into the second marking area. By doing so, the determinant distribution map is obtained.

During the generation of the determinant distribution map described above, the first processing circuit 203 first calculates the standard deviation distribution map of the original grayscale image, and the relationship between the standard deviation of each pixel in the standard deviation distribution map and the set value Since the gradation difference in the pseudo-flat portion is small, the standard deviation of the pixels in the pseudo-flat portion is also small . A pseudo-flat area (that is, the first marking area) is determined.

Furthermore, when dividing the first marking area and the second marking area, the first processing circuit 203 can be realized by a binarization method. A specific process is as follows. That is, pixels with standard deviations smaller than the set value in the standard deviation distribution map set the corresponding pixel values in the determinant distribution map to 0, and pixels with standard deviations greater than the set value in the standard deviation distribution map are the determinant distribution. In the figure, the corresponding pixel value is set to 1, so that pixels with a pixel value of 0 in the determinant distribution map belong to the first marking area, and pixels with a pixel value of 1 belong to the second marking area.

After dividing the area using the binarization method, in subsequent processing, it is determined whether the pixel belongs to the first marking area or the second marking area according to the pixel value of the determinant distribution map. It's easy and convenient.

FIG. 4 is a structural diagram of the first processing circuit 203 in an embodiment of the present invention, please refer to FIG. It includes a second computational sub-circuit 233 , a third computational sub-circuit 234 and a fourth computational sub-circuit 235 .

Here, the first calculation subcircuit 231 forms a grayscale square diagram by calculating the square value of the grayscale value of each pixel of the original grayscale image. The mean filtering subcircuit 232 generates a first expectation map by performing mean filtering on the original grayscale image, and generates a second expectation map by performing mean filtering on the tone-square map. Generate diagrams. The second calculation subcircuit 233 obtains the third expectation map by calculating the square value of the gradation value of each pixel of the first expectation map. The third calculation subcircuit 234 obtains a difference image by calculating the difference in grayscale value of pixels at corresponding positions in the second expectation map and the third expectation map. The fourth calculation sub-circuit 235 obtains the standard deviation distribution map by obtaining the square root of the gradation value of each pixel of the difference image. The standard deviation distribution map obtained by this method can reflect the gradation difference between each pixel and its surrounding pixels. It can be divided into regions.

In an embodiment of the present invention, the average filtering subcircuit 232 performs average filtering on the original grayscale image and the tone-square map, respectively, using a filtering template of size m×m; The range of values for m is 10-20.

By using m in this range, the standard deviation distribution map can better indicate pseudo-flat regions in the original image. As a result, the quasi-flat area is not expanded, and image quality deterioration is avoided.

As an example, m is fifteen. Using a value of 15 for m allows maximum discrimination of pseudo-flat regions in the original image.

In an embodiment of the present invention, the first processing circuit 203, after generating the determinant distribution map based on the original grayscale image, updates the determinant distribution map to obtain the updated decision In the factor distribution diagram, the second marking area includes an area in which the pixel positions in the original image are adjacent and the standard deviation is equal to or greater than the set value, and an area in which the pixel positions in the original image are adjacent and the standard deviation is equal to or greater than the set value. includes a region in which the number of pixels in a region having pixels smaller than the set value is smaller than the threshold.

During the updating of the determinant distribution map described above, a fill-in-the-blank algorithm is actually used, and the number of pixels in the region in which the pixel positions in the original image are adjacent and the pixels whose standard deviation is smaller than the set value is greater than the threshold value. Divide the smaller area into a second marking area, and during processing, perform the same processing on the hole and the pixels around it, so that only the perimeter of a region or the center of a region (hole) is extended and processed. It is possible to avoid sharp changes in contrast between the center and the periphery of a region in the captured image.

In an embodiment of the present invention, the threshold is 8%-15% of the number of pixels in the original image. By using a threshold for the range, the hole-filling algorithm will only run if there are sufficiently few pixels in the center of the region, so that regions with more than the threshold number of pixels will be processed by the hole-filling algorithm, It is possible to avoid the inability to optimize the image quality of pseudo-flat areas.

As an example, the threshold range is 10%. At this time, the area with the number of pixels above the threshold can be filled with a hole-filling algorithm to avoid the inability to optimize the image quality of the pseudo-flat area. Also, abrupt changes in contrast between the center and periphery of regions in the processed image can be avoided.

In an embodiment of the present invention, the set value ranges from 1-5. The use of the range settings ensures that the determinant map better shows the quasi-flat regions in the original image. As a result, the quasi-flat area is not expanded, and image quality deterioration is avoided.

As an example, the setting value is one. By using 1 as the setting value, pseudo-flat regions in the original image can be distinguished maximally.

In an embodiment of the present invention, the second processing circuit 204 can calculate the final grayscale image according to the following formula: I(i,j)=G(i,j)*H(i , j) + (1-G(i,j))*A(i,j).
where I is the final grayscale image, G is the determinant map (where pixels in the first marking area are 0 and pixels in the second marking area are 1), and H is the is the equalized grayscale image and A is the original grayscale image. where I(i,j) is the pixel at row i, column j in the final grayscale image, G(i,j) is the pixel at row i, column j in the determinant map, and H(i, j) is the pixel at row i and column j in the equalized grayscale image, and A(i, j) is the pixel at row i and column j in the original grayscale image. As a result of calculation according to the above formula, the gradation value of the pixel corresponding to the second marking area in I is the gradation value of the corresponding pixel in H, and the gradation value of the pixel corresponding to the first marking area in I is The grayscale value is the grayscale value of the corresponding pixel in A.

When the image processing apparatus according to the above embodiment performs image processing, the division of each functional module has been described as an example. can be achieved, namely by dividing the internal structure of the device into different functional modules to achieve all or part of the functions described above. It should be noted that the embodiments of the image processing apparatus and the image processing method according to the above embodiments belong to the same concept, and the specific implementation processes thereof refer to the method embodiments. It is not explained here.

FIG. 5 is a structural diagram of an image processing device according to an embodiment of the present invention. The image processing apparatus may be a device such as a computer.

Computer 300 includes a central processing unit (CPU) 301, a system memory 304 including random access memory (RAM) 302 and read only memory (ROM) 303, and a system bus 305 connecting system memory 304 and central processing unit 301. including. Computer 300 further includes a basic input/output system (I/O system) 306 that facilitates the transfer of information between components within the computer.

A basic input/output system 306 includes a display 308 for displaying information and an input device 309 such as a mouse and keyboard for user input of information. Here, both display 308 and input device 309 are connected to central processing unit 301 through input/output controller 310 connected to system bus 305 . The basic input/output system 306 may further include an input/output controller 310 that receives and processes input from a number of other devices such as keyboards, mice, and electronic styluses. Similarly, input/output controller 310 provides output to a display screen, printer, or other type of output device.

Without loss of generality, computer-readable media can include computer storage media and communication media. Computer storage media includes volatile, nonvolatile, removable and non-removable media implementing any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Including media. Computer storage media may include RAM, ROM, EPROM, EEPROM, flash memory or other solid state storage technology, CD-ROM, DVD or other optical storage devices, tape cartridges, magnetic tapes, disk storage devices or other magnetic storage. Including equipment. Of course, those of ordinary skill in the art will appreciate that the computer storage medium is not limited to the above.

According to embodiments of the present invention, computer 300 may operate by being connected via a network, such as the Internet, to a remote computer operating on the network. That is, computer 300 can be connected to network 312 via network interface unit 311 connected to system bus 305; It can be connected to a computer system (not shown).

The memory includes one or more programs, and the one or more programs are configured to be stored in the memory and executed by the CPU. When the CPU executes a program stored in memory, the methods shown in any of FIGS. 1-2 can be implemented.

In an exemplary embodiment, a computer readable storage medium containing instructions is also provided, such as a memory containing instructions, which are loaded and executed by central processing unit 301 of computer 300 to produce the Complete the method shown in one of 2. For example, a computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage, and the like.

It can be understood by those skilled in the art that all or part of the above embodiments may be completed by hardware, or by using a program to direct the relevant hardware. . The program can be stored in a computer-readable storage medium. The storage medium may be a read-only memory, a magnetic disk, an optical disk, or the like.

The above are only preferred embodiments of the present invention, and are not intended to limit the present invention, and any changes, equivalent substitutions, improvements, etc. made within the spirit and scope of this disclosure are within the scope of the present invention. shall be included.

201 generation circuit 202 histogram equalization circuit 203 first processing circuit 204 second processing circuit 205 third processing circuit

Claims (20)

generating an original grayscale image of the original image;
performing a histogram equalization process on the original grayscale image to obtain an equalized grayscale image;
Based on the original grayscale image, a first marking area including an area in which pixels in the original image have adjacent pixel positions and a standard deviation is smaller than a set value, and an area other than the first marking area. generating a determinant map including the second marking area;
Based on the original grayscale image, the equalized grayscale image, and the determinant distribution diagram, the grayscale value of the pixel corresponding to the second marking area is the grayscale of the corresponding pixel in the equalized grayscale image. obtaining a final grayscale image, wherein grayscale values of pixels corresponding to the first marking area are grayscale values of corresponding pixels in the original grayscale image;
A method of image processing comprising reconstructing a processed image based on said final grayscale image. Generating a determinant distribution map based on the original grayscale image includes:
Each pixel in the original grayscale image is calculated by calculating the standard deviation of each pixel in the original grayscale image based on the gradation values of pixels in a specific area centered on each pixel in the original grayscale image. obtaining a standard deviation distribution map containing the standard deviation of the pixels;
Pixels having a standard deviation smaller than the set value in the standard deviation distribution map are divided into the first marking regions, and regions other than the first marking region in the standard deviation distribution map are divided into the second marking regions. 2. The image processing method of claim 1, comprising obtaining the determinant distribution map by segmenting. Obtaining a standard deviation distribution chart by calculating the standard deviation of each pixel of the original grayscale image based on the gradation values of pixels in a specific area centered on each pixel in the original grayscale image. teeth,
forming a tone-square diagram by calculating the square of the tone value of each pixel of the original grayscale image;
generating a first expectation map by performing mean filtering on the original grayscale image, and generating a second expectation map by performing mean filtering on the tone-square map;
obtaining a third expectation map by calculating the square value of the gradation value of each pixel in the first expectation map;
obtaining a difference image by calculating a difference in grayscale values of pixels at corresponding positions in the second expected map and the third expected map;
3. The image processing method according to claim 2, comprising obtaining the standard deviation distribution map by obtaining a square root of the tone value of each pixel of the difference image. performing average filtering on the original grayscale image,
performing mean filtering on the original grayscale image using a filtering template of size m×m, wherein the value of m ranges from 10 to 20;
Performing average filtering on the tone-squared map includes:
4. The image processing method of claim 3, comprising performing mean filtering on the tone-squared map using the mxm sized filtering template. 5. The image processing method according to claim 4, wherein m is fifteen. After generating the determinant distribution map based on the original grayscale image,
By updating the determinant factor distribution map, in the updated determinant factor distribution map, the second marking area includes pixels whose pixel positions in the original image are adjacent and whose standard deviation is equal to or greater than the set value. 3. The method according to claim 2, further comprising: a region, and a region in which the pixel positions in the original image are adjacent and the number of pixels in the region is smaller than the threshold value . Image processing method. 7. The image processing method according to claim 6, wherein said threshold is 8% to 15% of the number of pixels of said original image. 8. The image processing method according to claim 7, wherein said threshold is 10% of the number of pixels of said original image. 9. The image processing method according to any one of claims 1 to 8, wherein the set value ranges from 1 to 5. a generation circuit for generating an original grayscale image of the original image;
a histogram equalization circuit for performing a histogram equalization process on the original grayscale image to obtain an equalized grayscale image;
Based on the original grayscale image, a first marking area including an area in which pixels in the original image have adjacent pixel positions and a standard deviation is smaller than a set value, and an area other than the first marking area. a first processing circuit for generating a determinant map including a second marking area;
Based on the original grayscale image, the equalized grayscale image, and the determinant distribution diagram, the grayscale value of the pixel corresponding to the second marking area is the grayscale of the corresponding pixel in the equalized grayscale image. a second processing circuit for obtaining a final grayscale image in which the grayscale values of the pixels corresponding to the first marking area are the grayscale values of the corresponding pixels in the original grayscale image;
An image processing apparatus comprising a third processing circuit for reconstructing a processed image based on said final grayscale image. The first processing circuit calculates the standard deviation of each pixel of the original grayscale image based on the gradation values of pixels in a specific area centered on each pixel of the original grayscale image. , obtaining a standard deviation distribution map containing the standard deviation of each pixel in the original grayscale image,
Pixels having a standard deviation smaller than the set value in the standard deviation distribution map are divided into the first marking regions, and regions other than the first marking region in the standard deviation distribution map are divided into the second marking regions. 11. The image processing apparatus according to claim 10, wherein said determining factor distribution map is obtained by dividing. The first processing circuit,
a first computational subcircuit for forming a tone-squared diagram by computing the squared tone value of each pixel of the original grayscale image;
a mean filtering sub-sub which generates a first expectation map by performing mean filtering on the original grayscale image and generates a second expectation map by performing mean filtering on the tone-squared map; a circuit;
a second calculation subcircuit for obtaining a third expectation map by calculating the square value of the grayscale value of each pixel of the first expectation map;
a third calculation sub-circuit for obtaining a difference image by calculating differences in grayscale values of pixels at corresponding positions in the second expectation map and the third expectation map;
12. The image processing apparatus according to claim 11, further comprising a fourth calculation sub-circuit for obtaining said standard deviation distribution map by calculating the square root of the gradation value of each pixel of said difference image. The averaging filtering subcircuit performs averaging filtering respectively on the original grayscale image and the tone-squared map using a filtering template of size m×m, where the value of m ranges from 10 to 13. The image processing apparatus according to claim 12, wherein the image processing apparatus is 20. 14. The image processing apparatus according to claim 13, wherein m is fifteen. After generating the determinant factor distribution map based on the original grayscale image, the first processing circuit updates the determinant factor distribution map so that in the updated determinant factor distribution map, the second The marking area is an area in which there are pixels whose pixel positions are adjacent in the original image and whose standard deviation is equal to or greater than the set value, and pixels whose pixel positions are adjacent in the original image and whose standard deviation is smaller than the set value 12. The image processing apparatus according to claim 11, wherein the number of pixels in a certain region includes a region smaller than a threshold. 16. The image processing apparatus according to claim 15, wherein said threshold is 8% to 15% of the number of pixels of said original image. 17. The image processing apparatus according to claim 16, wherein said threshold is 10% of the number of pixels of said original image. 18. The image processing apparatus according to any one of claims 10 to 17, wherein the set value ranges from 1 to 5. Characterized in that at least one command is stored, said command being loaded and executed by a processor to achieve the operations performed in the image processing method according to any one of claims 1 to 9. storage medium. 1. A computing device comprising a memory, a processor and a computer program stored in said memory and running on said processor,
A computer device, characterized in that the processor is arranged to perform the operations performed in the image processing method of any one of claims 1 to 9 when the computer program is executed.

Images