CN110400278B - A fully automatic correction method, device and equipment for image color and geometric distortion - Google Patents

Abstract

The invention discloses a fully automatic correction method for image color and geometric distortion, which acquires image information including a target object and an extended two-dimensional code; identifies and calculates the positioning mark of the extended two-dimensional code to obtain camera external parameters; The external parameter performs three-dimensional rotation and translation transformation of the target image; model fitting is performed on the edge of the color card of the extended QR code and the edge of the gray card, and the internal parameters of the camera are calculated; the camera internal parameters are used to perform lens geometry on the target image. Distortion correction; identify the color blocks and gray blocks of the extended two-dimensional code to perform color correction and gray correction on the target image; obtain the real image of the target. The invention reduces the misjudgment probability of image recognition and improves the accuracy of the image recognition result. The invention also discloses an automatic correction device, system and storage medium for image color and geometric distortion, which have corresponding technical effects.

Description

A fully automatic correction method, device and equipment for image color and geometric distortion

technical field

The present invention relates to the technical field of image processing, in particular to a fully automatic correction method, device, system and computer-readable storage medium for image color and geometric distortion.

Background technique

The current artificial intelligence technology is developing rapidly, and the computer vision technology is widely used. When using an image acquisition device to capture images, due to the difference in the lighting used, the different angles of the camera when shooting, the existence of geometric distortion of the camera, etc., the images captured by the image capture device have strong color, grayscale and geometric distortion. If these distortions and distortions cannot be calibrated, the probability of misjudgment of image recognition will increase, which will result in relatively large errors in image recognition results.

To sum up, how to effectively solve the problem that the color, grayscale and geometric distortion of the image cause a relatively large error in the image recognition result is an urgent problem for those skilled in the art.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a fully automatic correction method for image color and geometric distortion, which greatly reduces the misjudgment probability of image recognition and greatly improves the accuracy of image recognition results; another purpose of the present invention is A fully automatic correction device, system and computer-readable storage medium for image color and geometric distortion are provided.

In order to solve the above-mentioned technical problems, the present invention provides the following technical solutions:

A fully automatic correction method for image color and geometric distortion, including:

Obtain the image information including the target object and the extended two-dimensional code; wherein, the extended two-dimensional code includes a color card composed of a plurality of different color regions, a grayscale card composed of a plurality of different grayscale regions, and an image containing a positioning mark. QR code;

Identify and calculate the positioning mark of the extended two-dimensional code in the image information to obtain camera external parameters; perform a three-dimensional image of the target object corresponding to the target object in the image information according to the camera external parameters Rotation and translation transformation to get the image after rotation and translation correction;

By performing pattern fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information, the camera internal parameters are calculated by using the lens distortion model; Correct the lens geometric distortion of the image after translation correction to obtain the image after lens geometric distortion correction;

Identify the color blocks of the color card and the grayscale blocks of the grayscale card in the image information to obtain a recognition result; perform color correction and grayscale on the image after the lens geometric distortion correction according to the recognition result Correction to obtain the real image of the target.

In a specific embodiment of the present invention, the positioning mark includes three position identifiers that are respectively located at three corners of the extended two-dimensional code in black and white. Identify and calculate the positioning marks to obtain the camera external parameters, including:

Calculate the coordinates of the center points of the three position identifiers in the positioning marks respectively;

Using the geometric relationship of the three position identifiers and the position of each of the position identifiers in the image information, an inverse perspective transformation matrix of the image information is calculated to obtain the camera extrinsic parameters.

In a specific embodiment of the present invention, by performing pattern fitting on the edge of the color card and the edge of the gray card of the extended two-dimensional code in the image information, the camera is calculated by using the lens distortion model. Internal reference, including:

By performing pattern fitting on the edge of the color card and the edge of the gray card of the extended two-dimensional code in the image information, the camera internal parameters are calculated in combination with the lens distortion model and the steepest descent algorithm.

In a specific embodiment of the present invention, identifying the color blocks of the color card and the gray blocks of the gray card in the image information includes:

A polynomial regression method is used to identify the color blocks of the color card and the gray blocks of the gray card in the image information.

In a specific embodiment of the present invention, identifying the color blocks of the color card and the gray blocks of the gray card in the image information includes:

The BP neural network method is used to identify the color blocks of the color card and the gray blocks of the gray card in the image information.

In a specific embodiment of the present invention, identifying the color blocks of the color card and the gray blocks of the gray card in the image information includes:

A support vector machine algorithm is used to identify the color blocks of the color card and the gray blocks of the gray scale in the image information.

A fully automatic correction device for image color and geometric distortion, comprising:

The image information acquisition module is used to acquire the image information including the target object and the extended two-dimensional code; wherein, the extended two-dimensional code includes a color card composed of a plurality of different color regions, and a gray scale composed of a plurality of different grayscale regions. Degree card, a QR code containing a positioning mark;

a tilt distortion correction module, used for identifying and calculating the positioning mark of the extended two-dimensional code in the image information, to obtain camera external parameters; according to the camera external parameters, the target object in the image information The corresponding target image is subjected to three-dimensional rotation and translation transformation to obtain an image after rotation and translation correction;

a lens geometric distortion correction module, configured to perform mode fitting on the edge of the color card of the extended two-dimensional code in the image information and the edge of the gray card, and use the lens distortion model to calculate the camera internal parameters; using The camera internal parameter corrects the lens geometric distortion of the image after rotation and translation correction to obtain an image after lens geometric distortion correction;

A real image obtaining module is used to identify the color blocks of the color card and the gray blocks of the gray scale card in the image information, and obtain the identification result; after correcting the geometric distortion of the lens according to the identification result The image is subjected to color correction and grayscale correction to obtain a real image of the target, so that a patient corresponding to the target can be diagnosed according to the real image.

An automatic correction system for image color and geometric distortion, comprising an extended two-dimensional code, an image acquisition device, an image preprocessing terminal and a server, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, A grayscale card composed of different grayscale blocks, including the QR code of the positioning mark; of which:

The image acquisition device is used to acquire the image information of the extended two-dimensional code and the target object; and send the image information to the image preprocessing terminal;

The image preprocessing terminal is configured to identify and calculate the positioning mark of the extended two-dimensional code in the image information, and obtain camera external parameters; Performing three-dimensional rotation and translation transformation on the image of the target object corresponding to the target to obtain an image after rotation and translation correction; and sending the rotation and translation corrected image to the server;

The server is configured to perform mode fitting on the edge of the color card of the extended two-dimensional code in the image information and the edge of the gray card, and use the lens distortion model to calculate the camera internal parameters; use the The camera internal parameter corrects the lens geometric distortion of the image after rotation and translation correction to obtain an image after lens geometric distortion correction; Recognition to obtain a recognition result; color correction and grayscale correction are performed on the image after geometric distortion correction of the lens according to the recognition result to obtain the real image of the target, and the patient corresponding to the target is determined according to the real image. Diagnose.

In a specific embodiment of the present invention, the positioning mark includes three position identifiers that are located in black and white at three corners of the extended two-dimensional code, respectively,

The image preprocessing terminal is specifically configured to calculate the center point coordinates of the three position identifiers in the positioning mark respectively; The position in the image information is calculated, and the inverse perspective transformation matrix of the image information is calculated to obtain the camera extrinsic parameters.

A computer-readable storage medium having a computer program stored on the computer-readable storage medium, when the computer program is executed by a processor, realizes the steps of the fully automatic correction method for image color and geometric distortion as described above.

The invention provides an automatic correction method for image color and geometric distortion: acquiring image information including a target object and an extended two-dimensional code; wherein, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, which are composed of multiple A grayscale card composed of different grayscale blocks, including the two-dimensional code of the positioning mark; identify and calculate the positioning mark of the extended two-dimensional code in the image information to obtain the camera external parameters; according to the camera external parameters, the target object in the image information is identified and calculated The corresponding target image is transformed by three-dimensional rotation and translation, and the image after rotation and translation correction is obtained; by pattern fitting the edge of the color card and the edge of the gray card in the image information, the lens distortion model is used to calculate Camera internal parameters; use the camera internal parameters to correct the lens geometric distortion of the image after rotation and translation correction, and obtain the image after lens geometric distortion correction; identify the color blocks of the color card and the gray blocks of the gray card in the image information, and get the recognition Results: According to the recognition results, color correction and grayscale correction are performed on the image after lens geometric distortion correction, and the real image of the target object is obtained.

It can be seen from the above technical solutions that by presetting a color card composed of a plurality of different color blocks, a gray scale card composed of a plurality of different gray blocks, and an extended two-dimensional code containing a two-dimensional code of a positioning mark, the extended two-dimensional code containing the extended two-dimensional code can be obtained. The two-dimensional code and the image information of the target object, use the extended two-dimensional code to obtain the camera external parameters, complete the rotation and translation correction of the target object image; use the lens distortion model to calculate the camera internal parameters, complete the lens geometric distortion correction of the target object image; By identifying the color blocks of the color card and the gray blocks of the gray scale in the image information, the color correction and gray scale correction of the target image are completed. Finally, a real image is obtained, so that image recognition can be performed based on the real image, which greatly reduces the misjudgment probability of image recognition and greatly improves the accuracy of the image recognition result.

Correspondingly, the embodiments of the present invention also provide a fully automatic correction device, system and computer-readable storage medium for image color and geometric distortion corresponding to the above-mentioned automatic correction method for image color and geometric distortion, which have the above technical effects. This will not be repeated here.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings without creative efforts.

Fig. 1 is a kind of implementation flow chart of the automatic correction method of image color and geometric distortion in the embodiment of the present invention;

2 is a schematic structural diagram of an extended two-dimensional code in an embodiment of the present invention;

3 is a structural diagram of an image containing a target and an extended two-dimensional code obtained in an embodiment of the present invention;

4 is another implementation flow chart of the fully automatic correction method for image color and geometric distortion in the embodiment of the present invention;

5 is another implementation flow chart of the fully automatic correction method for image color and geometric distortion in the embodiment of the present invention;

6 is a structural diagram of a BP neural network in an embodiment of the present invention;

7 is another implementation flow chart of the fully automatic correction method for image color and geometric distortion in the embodiment of the present invention;

8 is a structural block diagram of a fully automatic correction device for image color and geometric distortion in an embodiment of the present invention;

9 is a structural block diagram of a fully automatic correction system for image color and geometric distortion in an embodiment of the present invention;

FIG. 10 is a schematic diagram of a fully automatic correction system for image color and geometric distortion in an embodiment of the present invention.

The figures are marked as follows:

11-image, 21-camera, 22-camera, 23-user mobile terminal, 31-client, 4-server.

Detailed ways

In order to make those skilled in the art better understand the solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Example 1:

Referring to FIG. 1, FIG. 1 is an implementation flowchart of an automatic correction method for image color and geometric distortion in an embodiment of the present invention, and the method may include the following steps:

S101: Acquire image information including a target object and an extended two-dimensional code.

Wherein, the extended two-dimensional code includes a color card composed of a plurality of different color regions, a grayscale card composed of a plurality of different gray-scale regions, and a two-dimensional code including a positioning mark.

Referring to FIG. 2, FIG. 2 is a schematic structural diagram of an extended two-dimensional code in an embodiment of the present invention. An extended two-dimensional code including a color card composed of a plurality of different color regions, a grayscale card composed of a plurality of different gray-scale regions, and a two-dimensional code containing a positioning mark can be preset. The two-dimensional code is also called QR (Quick Response) Code, which has the characteristics of omnidirectional (360°) reading. The two-dimensional code itself has a positioning mark composed of three black and white large squares, which are located on the upper left of the two-dimensional code. Corner, upper right corner, lower left corner, the purpose is to determine the size and position of the QR code. The calibration mark composed of three black and white small squares nested and the timing mark composed of two black and white straight lines are easy to determine the position and rotation angle of the QR code. Add a color card and a grayscale card below the QR code. The color card is composed of multiple rectangles of different colors, while the grayscale card adopts a self-made multi-level grayscale card, which is composed of rectangles of different grayscales. The design is arranged in the order of the first order followed by the nth order and then the second order, so that the grayscale cards arranged in this way can be better clustered in the later stage to calculate and separate each grayscale block. Different colors and different grayscales are divided using black borders, making it easy for computer vision to extract colors, grayscales, and divide grids of different colors and grayscales. Color and grayscale are used for color and grayscale correction of the system, and lattice is used to correct the geometric distortion of the lens. The grayscale card is different from the general grayscale gradient, and the grayscale is divided into multiple discrete color levels, graduating from white to black.

When a photo or video of a target needs to be collected, an extended two-dimensional code provided by the embodiment of the present invention can be placed next to the target, and an appropriate distance is maintained between the target and the extended two-dimensional code, and the target object can be obtained by placing an extended two-dimensional code provided by the embodiment of the present invention. and the image information of the extended QR code. Referring to FIG. 3 , FIG. 3 is a structural diagram of an image including a target object and an extended two-dimensional code obtained in an embodiment of the present invention. In FIG. 3, image information of a human face is acquired, an extended two-dimensional code provided by an embodiment of the present invention is placed beside the human face, and image information including the human face and the extended two-dimensional code is acquired.

The color card and the grayscale card can be Macbeth color chips, but are not limited to Macbeth color chips.

S102: Identify and calculate the positioning mark of the extended two-dimensional code in the image information to obtain camera external parameters; perform three-dimensional rotation and translation transformation on the target image corresponding to the target object in the image information according to the camera external parameters to obtain rotation and translation corrections post image.

After acquiring the image information including the target object and the extended two-dimensional code, the positioning marks of the extended two-dimensional code in the image information can be identified and calculated to obtain the Euler angle and translation vector, and then the camera extrinsic parameters can be obtained. According to the external parameters of the camera, three-dimensional rotation and translation transformation is performed on the target image corresponding to the target in the image information, and the image after rotation and translation correction is obtained.

S103: By performing pattern fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information, the camera internal parameters are calculated by using the lens distortion model; Correction to obtain the image after lens geometric distortion correction.

Pattern fitting can be performed on the edge of the color card of the extended QR code and the edge of the gray card in the image information. For example, the RANSAC method can be used to identify the edge of the color card and the edge of the gray card, and the lens distortion model can be used to calculate the camera internal parameters. The camera's internal parameters are used to correct the lens geometric distortion of the image after rotation and translation correction, and the image after lens geometric distortion correction is obtained. For pictures with lens distortion, only the calibrated distortion coefficient and internal parameter matrix can be used to restore the distorted picture to an undistorted picture. The process of calculating the camera internal parameters using the lens distortion model is as follows:

First, consider the lens distortion-free model, which is expressed by the following formula:

Among them, (X, Y, Z) represents a point in the world coordinate system, (u, v) represents the pixel coordinates of the image plane corresponding to this point in the world coordinates, and f _x and f _y are the calibrated camera lateral directions, respectively. and the longitudinal focal length, u ₀ and v ₀ are the calibrated camera image center points, R and T respectively represent the rotation matrix and translation matrix in the camera external parameters, z _c is the object point (X, Y, Z) in the camera coordinate system The coordinate along the Z axis below.

Let P be:

The lens distortion-free model becomes:

P is a matrix with 3 rows and 4 columns. Can be set to:

At this point, the point (X, Y, Z) of the world coordinate system is connected to the point (u, v) in the pixel plane coordinate system through P. The P matrix can be obtained by solving the six known points in world coordinates and their corresponding pixel coordinates. In general calibration work, there are often more than 6 known points on the target, so that the number of equations greatly exceeds the number of unknowns, so that the least squares method can be used to solve the problem to reduce the influence of errors.

The target generally refers to a lens calibration plate, or a calibration reference, and in this embodiment of the present invention refers to an extended two-dimensional code.

则镜头无畸变模型如下：Once the P matrix is obtained, the internal and external parameter matrices can be reversed. Suppose a point given n sets of pixel coordinates to a point in the world coordinate system

Then the lens distortion-free model is as follows:

Wherein, φ(P, M _i ) is the image point projected by the camera at the object-side point M _i =(X _i , Y _i , Z _i ).

The above considers the lens without distortion model. Now consider the lens imaging model after adding distortion. The lens imaging model after adding distortion is as follows:

Among them, R and T respectively represent the rotation matrix and translation matrix in the camera external parameters, and (x, y, z) are the coordinates transformed from the three-dimensional world coordinate point to the camera coordinate system. Radial distortion and tangential distortion are expressed as follows:

δ _x1 =x'(1+k ₁ r ² +k ₂ r ⁴ +...);

δ _y1 =y'(1+k ₁ r ² +k ₂ r ⁴ +...);

δ _x2 =[2p ₁ xy+p ₂ (r ² +2x ² )](1+p ₃ r ² +…);

δ _y2 =[p ₁ (r ² +2y ² )+2p ₂ xy](1+p ₃ r ² +…);

是加入了畸变量后在成像坐标系下的坐标；δ_x和δ_y是x和y方向上的畸变量，通常δ_x和δ_y包含两部分，一部分是径向畸变，另一部分是切向畸变；r²＝x′²+y′²，k1和k2就是要标定的径向畸变系数，更高阶的系数一般会忽略。p1和p2就是要标定的切向畸变系数，更高阶的系数也会忽略掉。in,

is the coordinate in the imaging coordinate system after adding the distortion amount; δ _x and δ _y are the distortion amounts in the x and y directions, usually δ _x and δ _y contain two parts, one part is radial distortion, and the other part is tangential Distortion; r ² =x′ ² +y′ ² , k1 and k2 are radial distortion coefficients to be calibrated, and higher-order coefficients are generally ignored. p1 and p2 are the tangential distortion coefficients to be calibrated, and higher-order coefficients are also ignored.

So the total distortion is expressed as:

δ _x =δ _x1 +δ _x2 ;

δ _y =δ _y1 +δ _y2 ;

可以表示为：image point coordinates

It can be expressed as:

Assume:

Suppose a point given n sets of pixel coordinates to a point in the world coordinate system:

The lens imaging model after adding distortion is as follows:

Among them, φ(R,T,A,k ₁ ,k ₂ ,p ₁ ,p ₂ ,M _i ) is the image point projected by the camera at the object point Mi =(X _i ,Y _i ,Z _{i )} .

The above is a nonlinear optimization process. The nonlinear optimization method can be used to find a solution that meets the requirements, so as to extract the internal parameter coefficients of the above formula and obtain the camera internal parameters.

S104: Identify the color blocks of the color card and the gray scale blocks of the gray scale in the image information to obtain the identification result; perform color correction and gray scale correction on the image after lens geometric distortion correction according to the identification result to obtain the real image of the target object .

Identify the color block of the color card and the gray scale of the gray card in the image information, and obtain the identification result; according to the identification result, perform color correction and gray correction on the image after lens geometric distortion correction to obtain the real image of the target object. For example, the polynomial regression method, the BP neural network method or the support vector machine algorithm can be used to identify the color card and gray card of the extended two-dimensional code in the image information.

It can be seen from the above technical solutions that by presetting a color card composed of a plurality of different color blocks, a gray scale card composed of a plurality of different gray blocks, and an extended two-dimensional code containing a two-dimensional code of a positioning mark, the extended two-dimensional code containing the extended two-dimensional code can be obtained. The two-dimensional code and the image information of the target object, use the extended two-dimensional code to obtain the camera external parameters, complete the rotation and translation correction of the target object image; use the lens distortion model to calculate the camera internal parameters, complete the lens geometric distortion correction of the target object image; By identifying the color blocks of the color card and the gray blocks of the gray scale in the image information, the color correction and gray scale correction of the target image are completed. Finally, a real image is obtained, so that image recognition can be performed based on the real image, which greatly reduces the misjudgment probability of image recognition and greatly improves the accuracy of the image recognition result.

It should be noted that, based on the foregoing first embodiment, the embodiment of the present invention also provides a corresponding improvement solution. In subsequent embodiments, the same steps or corresponding steps in the above-mentioned first embodiment can be referred to each other, and corresponding beneficial effects can also be referred to each other, which will not be repeated in the following improved embodiments.

Embodiment 2:

Referring to FIG. 4, FIG. 4 is another implementation flowchart of the fully automatic correction method for image color and geometric distortion in the embodiment of the present invention, and the method may include the following steps:

S401: Acquire image information including a target object and an extended two-dimensional code.

Wherein, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, a grayscale card composed of a plurality of different gray blocks, and a two-dimensional code including a positioning mark.

S402: Calculate the coordinates of the center points of the three position identifiers in the positioning mark respectively.

First load the captured image information, first perform grayscale conversion on the loaded image information, and then perform configuration verification on the image information. Next, scan the image information line by line. The pixel points are scanned in one line in increments, and the filtering is completed, the edge gradient is obtained, the gradient threshold is adaptive, the edge is determined, and it is converted into a light and dark width flow. After the edge is determined, the current edge is subtracted from the last saved edge to obtain a width, and the current edge information is saved. Then calculate the saved light and dark width flow, describe the current width flow as a custom line segment structure, including information such as endpoints and lengths at both ends, and store the horizontal line segment structure variables that meet the conditions into a horizontal line segment collection of a container middle. Scan the entire image column by column, the steps are the same as the line by line scan, the scanning path is N-shaped, find the edge and finally obtain the vertical light and dark height flow, and store the vertical line segments that conform to the QR code into the vertical line segment set of the container. middle. Therefore, in the process of parsing the two-dimensional code, firstly, the coordinates of the center points of the three position identifiers in the positioning mark should be calculated separately. The horizontal and vertical line segments obtained before can be screened with a ratio of 1:1:3:1:1, and the horizontal and vertical intersections can be obtained after clustering, and the three positions in the positioning mark can be obtained. The coordinates of the center point of the identifier.

The location identifiers may be three black and white location identifiers located at three corners of the extended two-dimensional code. For example, the three location identifiers may be located at the upper left corner, the lower right corner and the lower left corner of the extended two-dimensional code, respectively.

It should be noted that, the embodiments of the present invention do not limit the shapes of the three location identifiers, for example, they may be set as squares or circles.

S403: Using the geometric relationship of the three position identifiers and the position of each position identifier in the image information, calculate an inverse perspective transformation matrix of the image information to obtain the camera extrinsic parameters.

After calculating the center point coordinates of the three position identifiers in the positioning mark, the inverse perspective transformation matrix of the image information can be calculated by using the geometric relationship of the three position identifiers and the position of each position identifier in the image information, thereby obtaining Camera external parameters.

S404: Perform three-dimensional rotation and translation transformation on the target image corresponding to the target in the image information according to the external parameters of the camera, to obtain an image after rotation and translation correction.

S405: By performing pattern fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information, and combining the lens distortion model and the steepest descent algorithm, calculate the camera internal parameters; use the camera internal parameters to correct the rotation and translation. The lens geometric distortion of the image is corrected to obtain an image after lens geometric distortion correction.

After the lens distortion model is obtained, the camera intrinsic parameters can be extracted from the model using the steepest descent algorithm. The gradient descent method is simple to implement. When the objective function is a convex function, the solution of the gradient descent method is a global solution. The optimization idea of the gradient descent method is to use the negative gradient direction of the current position as the search direction, because this direction is the fastest descent direction of the current position, so it is also called the "steepest descent method". The closer the steepest descent method is to the target value, the smaller the step size and the slower the progress. In general, the solution is not guaranteed to be the global optimal solution, and the speed of the gradient descent method is not necessarily the fastest. Therefore, in the embodiment of the present invention, the Levenberg-Marquardt method, that is, the steepest descent algorithm, is used to obtain the parameters. The Levenberg-Marquardt algorithm can provide a numerical solution of nonlinear minimization (local minimization). . This algorithm can modify the parameters during execution to combine the advantages of the Gauss-Newton algorithm and the gradient descent method, and improve the shortcomings of the two.

S406: Use the polynomial regression method to identify the color blocks of the color card and the gray blocks of the gray scale in the image information, and obtain the identification result; perform color correction and gray scale correction on the image after the lens geometric distortion correction according to the identification result, and obtain the target real images of things.

The polynomial regression method can be used to identify the color blocks of the color card and the gray blocks of the gray scale in the image information. Suppose there are N color blocks on the color card, and the color tristimulus values of the i-th color block are _Roi , _Goi , and _Boi in the standard space, and the No. The color tristimulus values of i color blocks are R _i , G _i , B _i , where i=1, 2, 3....N, then:

_Roi =a ₁₁ v _1i +a ₁₂ v _2i +...+a _1i v _ji ;

G _oi =a ₂₁ v _1i +a ₂₂ v _2i +...+a _2i v _ji ;

_Boi =a ₃₁ v _1i +a ₃₂ v _2i +...+a _3i v _ji ;

Among them, v _ji (j=1,...,J) consists of polynomials of R _i , G _i , B _i , and there are various polynomial forms, for example, V=[R,G,B,1], V=[R, G, B, RG, RB, GB, 1], V=[R, G, B, RGB, 1], etc. The form of V can be combined into different forms as required.

The matrix form of the above formula is:

X=A ^T *V;

Among them, X is the R, G, B tristimulus value matrix of the corrected image, the dimension is 3×M; V is the matrix composed of the polynomial terms corresponding to the R, G, B values of all pixels of the original image, the dimension The number is J×M; M is the total number of pixels of the original image. A is a conversion coefficient matrix with dimension J×3, A can be obtained by least squares optimization, and A is the required model parameter. By bringing A into the above matrix formula, the R, G, and B values of each pixel of the corrected image can be calculated to realize online color correction. By designing a reasonable polynomial term form, the polynomial regression algorithm is used to identify the color card and gray card of the extended two-dimensional code in the image information, so as to obtain a good identification result.

Embodiment three:

Referring to FIG. 5, FIG. 5 is another implementation flowchart of the fully automatic correction method for image color and geometric distortion in the embodiment of the present invention, and the method may include the following steps:

S501: Acquire image information including the target object and the extended two-dimensional code.

Wherein, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, a grayscale card composed of a plurality of different gray blocks, and a two-dimensional code including a positioning mark.

S502: Calculate the coordinates of the center points of the three position identifiers in the positioning mark respectively.

S503: Calculate the inverse perspective transformation matrix of the image information by using the geometric relationship of the three position identifiers and the position of each position identifier in the image information to obtain the camera extrinsic parameters.

S504: Perform three-dimensional rotation and translation transformation on the target image corresponding to the target in the image information according to the external parameters of the camera, to obtain an image after rotation and translation correction.

S505: By performing pattern fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information, and combining the lens distortion model and the steepest descent algorithm, calculate the camera internal parameters; use the camera internal parameters to correct the rotation and translation. The lens geometric distortion of the image is corrected to obtain an image after lens geometric distortion correction.

S506: Use the BP neural network method to identify the color blocks of the color card and the gray scale blocks of the gray scale in the image information, and obtain a recognition result; perform color correction and grayscale on the image after the lens geometric distortion correction according to the recognition result. degree correction to get the real image of the target.

A feedforward (BP) neural network method can be used to identify the color blocks of the color card and the gray blocks of the gray scale in the image information. Specifically, the processing end may use the normalized squared difference matching method to perform template matching on the standard color blocks in the picture, and template matching is a technique of finding the part in one image that best matches another template image. It is a method of matching the actual image and the input image by sliding the template image block on the input image. Use T( ) to represent the template image, I( ) to represent the image to be matched, R( ) to represent the matching result, in the formula (x, y) represent the pixels of the image to be matched, and (x', y') represent The pixel points of the template image, the matching method is as follows:

Cut out the matched color card in the picture and save it. The color card has multiple color blocks, and the color values in the color blocks are clustered. The color value of the cluster center represents the color value Q1 of each color block after collection, and the color value in the color card can be used as the BP neural network. The reference value Q2 is output. For the trained BP neural network, Q1 is the actual color input value, and Q2 is the expected color output value.

The structure of the designed BP neural network is shown in Figure 6. The output layer and the input layer are 3 neurons respectively, and the BP neural network has a hidden layer. The number of neurons in the hidden layer is 7. The activation function of the BP neural network is a nonlinear function, and the data is normalized and input to the neural network. The BP neural network is trained using the data of multiple color blocks, and the color correction model coefficients are obtained. The color correction model obtained by inputting the color tristimulus value data R, G, and B of the entire image is color corrected and saved.

Embodiment 4:

Referring to FIG. 7, FIG. 7 is another implementation flowchart of the fully automatic correction method for image color and geometric distortion in the embodiment of the present invention, and the method may include the following steps:

S701: Acquire image information including the target object and the extended two-dimensional code.

Wherein, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, a grayscale card composed of a plurality of different gray blocks, and a two-dimensional code including a positioning mark.

S702: Calculate the coordinates of the center points of the three position identifiers in the positioning mark respectively.

S703: Using the geometric relationship of the three position identifiers and the position of each position identifier in the image information, calculate the inverse perspective transformation matrix of the image information to obtain the camera extrinsic parameters.

S704: Perform three-dimensional rotation and translation transformation on the target image corresponding to the target in the image information according to the external parameters of the camera, to obtain an image after rotation and translation correction.

S705: By performing pattern fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information, and combining the lens distortion model and the steepest descent algorithm, calculate the camera internal parameters; use the camera internal parameters to correct the rotation and translation The lens geometric distortion of the image is corrected to obtain an image after lens geometric distortion correction.

S706: Use the support vector machine algorithm to identify the color blocks of the color card and the gray blocks of the gray scale in the image information, and obtain the identification result; perform color correction and gray scale correction on the image after the lens geometric distortion correction according to the identification result, and obtain Real image of the target.

The support vector machine algorithm can be used to identify the color blocks of the color card and the gray blocks of the gray card in the image information, or the support vector machine algorithm can be used to use the color blocks on the standard color card as the training sample set, and the nonlinear kernel can be used. The function transforms the training data into a high-dimensional feature space. First collect the color patch of the color card reference image shot in the standard environment and the color patch RGB value of the color card image shot in the non-standard environment, and then convert them from the RGB color space to the Lab color space. The Lab color model is determined by brightness (L) is composed of three elements a and b related to color. L represents Luminosity, a represents the range from magenta to green, and b represents the range from yellow to blue. The training data in the training set of the obtained Lab color space is divided into three components, L, a, and b, thereby obtaining three training subsets of L, a, and b, and then using the color of the color card image taken in a non-standard environment. The block is used as the source, the color block of the reference image of the color card taken in the standard environment is used as the target, and the three subsets of L, a, b are trained separately, and three support vector regression models and L, a, b can be obtained. The support vector regression functions f _{L_SVR} , f _{a_SVR} , f _{b_SVR} . Finally, the RGB color space of the image to be corrected is converted to the Lab color space, and the regression function obtained by training is used for regression. Suppose the color value of the i-th pixel of the image to be corrected is L _i , a _i , b _i , where i=1, 2, 3,..., N, N is the total number of image pixels, respectively through the support vector Regression functions f _{L_SVR} , f _{a_SVR} , f _{b_SVR} , calculate the corrected color value L_SVR _i , a_SVR _i , b_SVR _i of the pixel, and its form is:

L_SVR _i =f _{L_SVR} (L _i );

a_SVR _i =f _{a_SVR} (a _i );

b_SVR _i =f _{b_SVR} (b _i );

Finally, the image corrected by the support vector machine algorithm is obtained. The specific form and solution process of the support vector regression model and the support vector regression functions f _{L_SVR} , f _{a_SVR} , f _{b_SVR} are as follows:

First of all, a loss function needs to be defined. In the field of medical images, the ε-insensitive loss function is mainly used. Compared with the least squared difference loss function, the Laplace loss function, and the Huber loss function, the ε-insensitive loss function can ignore the real loss function. The error is within a certain range of the value, and it can help to obtain less support vectors, which has good robustness.

The ε-insensitive loss function can be expressed as:

L _ε (y)=max(0,|f(x)-y|-ε);

where y represents the value calculated by the ε-insensitive loss function.

The kernel function adopts the Gaussian radiation kernel function. For any vector x _i , x _j ∈X∈R ⁿ , δ is the bandwidth, which controls the local scope of the Gaussian kernel function:

Solve the regression parameters to get the Lagrange multipliers α,α ^* :

The constraints satisfied by the above formula are:

And satisfy the Karush-Kuhn-Tucker (KKT) condition:

因此支持向量是指那些拉格朗日乘子大于零的点。Solving the above regression parameter equation under the above constraints, the Lagrange multiplier can be obtained

So support vectors are those points where the Lagrange multiplier is greater than zero.

The regression equation of the following form is obtained as:

in:

为偏置，x_r和x_s分别为核函数的某个x向量。K(x _i ,x _j ) is the kernel function,

is the bias, x _r and x _s are a certain x vector of the kernel function, respectively.

Therefore, the specific forms of the support vector regression functions f _{L_SVR} , f _{a_SVR} , and f _{b_SVR} are the same as the above regression equations.

By cutting out the matched grayscale blocks in the picture and saving them. The grayscale card has multiple grayscales. The multiple grayscales are clustered and their values are calculated. According to the accurate values of the standard multiple grayscales, the coefficient of variation of the grayscale of the picture after shooting is calculated. The variation coefficients of different gray levels can be used to correct different gray levels of the picture, and finally an image without color distortion and gray level distortion can be obtained.

Corresponding to the above method embodiments, the embodiments of the present invention also provide a fully automatic correction device for image color and geometric distortion. The fully automatic correction device for image color and geometric distortion described below is the same as the image color and geometric distortion described above. The fully automatic correction methods for distortion can be referenced against each other.

Referring to FIG. 8, FIG. 8 is a structural block diagram of a fully automatic correction device for image color and geometric distortion in an embodiment of the present invention, and the device may include:

The image information acquisition module 81 is used to acquire the image information including the target object and the extended two-dimensional code; wherein, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, and a gray scale composed of a plurality of different gray blocks Card, QR code containing locator mark;

The rotation and translation correction module 82 is used to identify and calculate the positioning mark of the extended two-dimensional code in the image information, and obtain the camera external parameters; according to the camera external parameters, perform three-dimensional rotation and translation on the target image corresponding to the target in the image information. Transform to get the image after rotation and translation correction;

The lens geometric distortion correction module 83 is used to perform mode fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information, and use the lens distortion model to calculate the camera internal parameters; use the camera internal parameters to correct rotation and translation The lens geometric distortion of the post image is corrected to obtain an image after lens geometric distortion correction;

The real image obtaining module 84 is used to identify the color blocks of the color card and the gray scale blocks of the gray scale in the image information to obtain the identification result; to perform color correction and gray scale correction on the image after the lens geometric distortion correction according to the identification result, A real image of the target is obtained so that the patient corresponding to the target can be diagnosed according to the real image.

It can be seen from the above technical solutions that by presetting a color card composed of a plurality of different color blocks, a gray scale card composed of a plurality of different gray blocks, and an extended two-dimensional code containing a two-dimensional code of a positioning mark, the extended two-dimensional code containing the extended two-dimensional code can be obtained. The two-dimensional code and the image information of the target object, use the extended two-dimensional code to obtain the camera external parameters, complete the rotation and translation correction of the target object image; use the lens distortion model to calculate the camera internal parameters, complete the lens geometric distortion correction of the target object image; By identifying the color blocks of the color card and the gray blocks of the gray scale in the image information, the color correction and gray scale correction of the target image are completed. Finally, a real image is obtained, so that image recognition can be performed based on the real image, which greatly reduces the misjudgment probability of image recognition and greatly improves the accuracy of the image recognition result.

In a specific embodiment of the present invention, the positioning mark includes three position identifiers that are respectively located at three corners of the extended two-dimensional code in black and white, and the rotation and translation correction module 82 includes a camera extrinsic parameter acquisition sub-module, which obtains the camera extrinsic parameters. Submodules include:

The coordinate calculation unit is used to calculate the coordinates of the center point of the three position identifiers in the positioning mark respectively

The camera extrinsic parameter obtaining unit is used for calculating the inverse perspective transformation matrix of the image information by using the geometric relationship of the three position identifiers and the position of each position identifier in the image information to obtain the camera extrinsic parameters.

In a specific embodiment of the present invention, the lens geometric distortion correction module 83 includes a camera internal parameter calculation sub-module,

The camera internal parameter calculation sub-module is a module that calculates the camera internal parameters by combining the lens distortion model and the steepest descent algorithm by performing pattern fitting on the edge of the color card of the extended two-dimensional code and the edge of the gray card in the image information.

In a specific embodiment of the present invention, the real image obtaining module 94 includes a color card and a gray card identification sub-module,

The color card and gray card identification sub-module is specifically a module that uses the polynomial regression method to identify the color blocks of the color card and the gray blocks of the gray card in the image information.

In a specific embodiment of the present invention, the real image obtaining module 84 includes a color card and a gray card identification sub-module,

The color card and gray card identification sub-module is specifically a module that uses the BP neural network method to identify the color blocks of the color card and the gray blocks of the gray card in the image information.

In a specific embodiment of the present invention, the real image obtaining module 84 includes a color card and a gray card identification sub-module,

The color card and gray card identification sub-module is specifically a module that uses the support vector machine algorithm to identify the color blocks of the color card and the gray blocks of the gray card in the image information.

Corresponding to the above method embodiments, the embodiments of the present invention also provide a fully automatic correction system for image color and geometric distortion. The fully automatic correction system for image color and geometric distortion described below is the same as the image color and geometric distortion described above. The fully automatic correction methods for distortion can be referenced against each other.

Referring to FIG. 9, FIG. 9 is a structural block diagram of a fully automatic correction system for image color and geometric distortion in an embodiment of the present invention. The system may include an extended two-dimensional code 1, an image acquisition device 2, an image preprocessing terminal 3, and a server 4. The extended two-dimensional code 1 includes a color card composed of a plurality of different color blocks, a grayscale card composed of a plurality of different gray blocks, and a two-dimensional code containing a positioning mark; wherein:

The image acquisition device 2 is used for acquiring the image information of the extended two-dimensional code 1 and the target object; and sending the image information to the image preprocessing terminal 3;

The image preprocessing terminal 3 is used to identify and calculate the positioning mark of the extended two-dimensional code 1 in the image information, and obtain the camera external parameters; according to the camera external parameters, perform three-dimensional rotation and translation on the target image corresponding to the target object in the image information Transform to obtain the image after rotation and translation correction; and send the image after rotation and translation correction to the server 4;

The server 4 is used to perform pattern fitting on the edge of the color card of the extended two-dimensional code 1 and the edge of the gray card in the image information, and use the lens distortion model to calculate the camera internal parameters; Correct the lens geometric distortion to obtain the image after lens geometric distortion correction; identify the color block of the color card and the gray block of the gray card in the image information, and obtain the recognition result; according to the recognition result, color the image after the lens geometric distortion correction. Correction and grayscale correction are performed to obtain the real image of the target, and the patient corresponding to the target is diagnosed according to the real image.

It can be seen from the above technical solutions that by presetting a color card composed of a plurality of different color blocks, a gray scale card composed of a plurality of different gray blocks, and an extended two-dimensional code containing a two-dimensional code of a positioning mark, the extended two-dimensional code containing the extended two-dimensional code can be obtained. The two-dimensional code and the image information of the target object, use the extended two-dimensional code to obtain the camera external parameters, complete the rotation and translation correction of the target object image; use the lens distortion model to calculate the camera internal parameters, complete the lens geometric distortion correction of the target object image; By identifying the color blocks of the color card and the gray blocks of the gray scale in the image information, the color correction and gray scale correction of the target image are completed. Finally, a real image is obtained, so that image recognition can be performed based on the real image, which greatly reduces the misjudgment probability of image recognition and greatly improves the accuracy of the image recognition result.

In a specific embodiment of the present invention, the positioning mark includes three position identifiers located in black and white at the three corners of the extended two-dimensional code 1, respectively,

The image preprocessing terminal 3 is specifically used to calculate the center point coordinates of the three position identifiers in the positioning mark respectively; using the geometric relationship of the three position identifiers and the position of each position identifier in the image information, calculate the inverse of the image information. Perspective transformation matrix to get camera extrinsic parameters.

In a specific embodiment of the present invention, the server 4 is specifically configured to perform pattern fitting on the edge of the color card of the extended two-dimensional code 1 and the edge of the gray card in the image information, combining the lens distortion model and the steepest The downhill algorithm calculates the camera internal parameters.

In a specific embodiment of the present invention, the server 4 is specifically configured to use the polynomial regression method to identify the color blocks of the color card and the gray blocks of the gray scale in the image information.

In a specific embodiment of the present invention, the server 4 is specifically configured to identify the color blocks of the color card and the gray blocks of the gray scale in the image information by using the BP neural network method.

In a specific implementation manner of the present invention, the server 4 is specifically configured to use a support vector machine algorithm to identify the color blocks of the color card and the gray blocks of the gray scale card in the image information.

In a specific example application, referring to FIG. 10 , FIG. 10 is a schematic diagram of a fully automatic correction system for image color and geometric distortion in an embodiment of the present invention. If currently in the process of diagnosing a patient by observing the color of the patient's face, first, the image 11 including the patient's face image and the extended two-dimensional code can be captured by the camera 21 , the video camera 22 or the user's mobile terminal 23 . If the image is collected by the camera 21 or the camera 22, the collected image 11 can be sent to the client 31, and the camera 21, the camera 22 and the client 31 can be connected through wired communication or wireless communication. The client 31 is used to perform three-dimensional rotation and translation transformation of the facial image; if the image is captured by the user's mobile terminal 23, the user's mobile terminal 23 can directly use the user's mobile terminal 23 to perform three-dimensional rotation and translation transformation of the facial image. After performing three-dimensional rotation and translation transformation on the facial image, the image after rotation and translation correction is sent to the server 4 through the client 31 or the user's mobile terminal 23, and the client 31 and the server 4 can be connected through wired communication. It is also possible to communicate through wireless communication, and use the server 4 to further perform lens geometric distortion correction and color and gray correction on the facial image, and finally obtain the patient's real face image. The diagnostic result is returned to the user's mobile terminal 23 or client 31, which greatly improves the recognition rate of the diagnostic system and greatly reduces the misjudgment rate of the diagnostic system.

The fully automatic correction system for image color and geometric distortion shown in Figure 10 can also be an anti-theft or sign-in system for face recognition. The collection process and processing process of face images can refer to the above process. The image information undergoes three-dimensional rotation and translation transformation, lens geometric distortion correction, and color and grayscale correction, which can greatly improve the authenticity of face images, and improve the recognition effect and efficiency. When the target object mentioned in the implementation of the present invention is not limited to a human face, it can also be other image recognition objects, which is not limited in this embodiment of the present invention.

Corresponding to the above method embodiments, the present invention also provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the following steps can be implemented:

Obtain the image information including the target object and the extended two-dimensional code; wherein, the extended two-dimensional code includes a color card composed of a plurality of different color blocks, a gray scale card composed of a plurality of different gray blocks, and a two-dimensional image containing a positioning mark. Identify and calculate the positioning marks of the extended two-dimensional code in the image information to obtain the camera external parameters; perform three-dimensional rotation and translation transformation on the target image corresponding to the target object in the image information according to the camera external parameters to obtain the rotation and translation corrections After image; by pattern fitting the edge of the color card of the extended QR code and the edge of the gray card in the image information, the lens distortion model is used to calculate the camera internal parameters; the camera internal parameters are used to correct the lens geometric distortion of the image after rotation and translation Perform correction to obtain the image after lens geometric distortion correction; identify the color block of the color card and the gray scale block of the gray card in the image information, and obtain the recognition result; according to the recognition result, perform color correction and grayscale on the image after lens geometric distortion correction. degree correction to get the real image of the target.

The computer-readable storage medium may include: a USB flash drive, a removable hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, etc., which can store program codes. medium.

For the introduction of the computer-readable storage medium provided by the present invention, please refer to the foregoing method embodiments, which will not be repeated in the present invention.

The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments may be referred to each other. As for the apparatus, system and computer-readable storage medium disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and for related parts, please refer to the description of the method section.

The principles and implementations of the present invention are described herein by using specific examples, and the descriptions of the above embodiments are only used to help understand the technical solutions and core ideas of the present invention. It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, several improvements and modifications can also be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims (10)

1. A full-automatic correction method for image color and geometric distortion is characterized by comprising the following steps:

acquiring image information containing a target object and an extended two-dimensional code; the expanded two-dimensional code comprises a color card consisting of a plurality of different color blocks, a gray card consisting of a plurality of different gray blocks, and a two-dimensional code containing a positioning mark;

identifying and calculating the positioning mark of the extended two-dimensional code in the image information to obtain camera external parameters; performing three-dimensional rotation and translation transformation on the target object image corresponding to the target object in the image information according to the camera external parameters to obtain an image subjected to rotation and translation correction;

performing mode fitting on the edge of the color card and the edge of the gray card of the extended two-dimensional code in the image information, and calculating camera internal parameters by using a lens distortion model; correcting the lens geometric distortion of the image after the rotation and translation correction by using the camera internal parameters to obtain a corrected image of the lens geometric distortion;

identifying color blocks of the color card and gray blocks of the gray card in the image information to obtain an identification result; and performing color correction and gray correction on the image after the geometric distortion correction of the lens according to the identification result to obtain a real image of the target object.

2. The method according to claim 1, wherein the positioning mark includes three position identifiers respectively located at three corners of the extended two-dimensional code, and the positioning mark of the extended two-dimensional code in the image information is identified and calculated to obtain a camera external parameter, including:

respectively calculating the coordinates of the central points of three position identifiers in the positioning mark;

and calculating an inverse perspective transformation matrix of the image information by using the geometric relationship of the three position identifiers and the positions of the position identifiers in the image information to obtain the external parameters of the camera.

3. The method according to claim 1 or 2, wherein calculating camera parameters using a lens distortion model by performing pattern fitting on the edges of the color card and the edges of the grayscale card of the extended two-dimensional code in the image information comprises:

and performing mode fitting on the edge of the color card and the edge of the gray card of the extended two-dimensional code in the image information, and calculating camera internal parameters by combining a lens distortion model and a steepest descent algorithm.

4. The method according to claim 3, wherein identifying color blocks of the color card and gray blocks of the gray card in the image information comprises:

and identifying the color blocks of the color card and the gray blocks of the gray card in the image information by utilizing a polynomial regression method.

5. The method according to claim 3, wherein identifying color blocks of the color card and gray blocks of the gray card in the image information comprises:

and identifying the color blocks of the color card and the gray blocks of the gray card in the image information by using a BP neural network method.

6. The method according to claim 3, wherein identifying color blocks of the color card and gray blocks of the gray card in the image information comprises:

and identifying the color blocks of the color card and the gray blocks of the gray card in the image information by using a support vector machine algorithm.

7. A device for fully automatic correction of color and geometric distortions of an image acquisition system, comprising: the image information acquisition module is used for acquiring image information containing a target object and the extended two-dimensional code; the expanded two-dimensional code comprises a color card consisting of a plurality of different color blocks, a gray card consisting of a plurality of different gray blocks, and a two-dimensional code containing a positioning mark;

the rotation and translation correction module is used for identifying and calculating the positioning mark of the extended two-dimensional code in the image information to obtain camera external parameters; performing three-dimensional rotation and translation transformation on the target object image corresponding to the target object in the image information according to the camera external parameters to obtain an image subjected to rotation and translation correction;

the lens geometric distortion correction module is used for performing mode fitting on the edge of the color card and the edge of the gray card of the extended two-dimensional code in the image information and calculating camera internal parameters by using a lens distortion model; correcting the lens geometric distortion of the image after the rotation and translation correction by using the camera internal parameters to obtain a corrected image of the lens geometric distortion;

the real image obtaining module is used for identifying color blocks of the color card and gray blocks of the gray card in the image information to obtain an identification result; and performing color correction and gray scale correction on the image after the lens geometric distortion correction according to the identification result to obtain a real image of the target object so as to diagnose a patient corresponding to the target object according to the real image.

8. A full-automatic correction system for image color and geometric distortion is characterized by comprising an extended two-dimensional code, image acquisition equipment, an image preprocessing terminal and a server, wherein the extended two-dimensional code comprises a color card consisting of a plurality of different color blocks, a gray card consisting of a plurality of different gray blocks and a two-dimensional code containing a positioning mark; wherein:

the image acquisition equipment is used for acquiring the image information of the extended two-dimensional code and the target object; sending the image information to the image preprocessing terminal;

the image preprocessing terminal is used for identifying and calculating the positioning mark of the extended two-dimensional code in the image information to obtain camera external parameters; performing three-dimensional rotation and translation transformation on the target object image corresponding to the target object in the image information according to the camera external parameters to obtain an image subjected to rotation and translation correction; sending the image subjected to the rotation and translation correction to the server;

the server is used for performing mode fitting on the edge of the color card and the edge of the gray card of the extended two-dimensional code in the image information and calculating camera internal parameters by using a lens distortion model; correcting the lens geometric distortion of the image after the rotation and translation correction by using the camera internal parameters to obtain a corrected image of the lens geometric distortion; identifying color blocks of the color card and gray blocks of the gray card in the image information to obtain an identification result; and performing color correction and gray scale correction on the image after the lens geometric distortion correction according to the identification result to obtain a real image of the target object, and diagnosing a patient corresponding to the target object according to the real image.

9. The system of claim 8, wherein the position marker includes three position identifiers respectively located at three corners of the extended two-dimensional code, which are black and white alternated,

the image preprocessing terminal is specifically configured to calculate coordinates of center points of the three position identifiers in the positioning mark respectively; and calculating an inverse perspective transformation matrix of the image information by using the geometric relationship of the three position identifiers and the positions of the position identifiers in the image information to obtain the external parameters of the camera.

10. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, which computer program, when being executed by a processor, carries out the steps of the method for fully automatic correction of color and geometric distortion of an image according to any one of claims 1 to 6.

Images