CN118261929A

CN118261929A - Checkerboard corner detection method and device, electronic equipment and storage medium

Info

Publication number: CN118261929A
Application number: CN202211633397.9A
Authority: CN
Inventors: 温捷文; 何高志; 杨育智
Original assignee: Shenzhen Lan You Technology Co Ltd
Current assignee: Shenzhen Lan You Technology Co Ltd
Filing date: 2022-12-19
Publication date: 2024-06-28

Abstract

The invention provides a checkerboard corner detection method, a device, electronic equipment and a storage medium. Performing image preprocessing by acquiring an acquired original image to acquire a first image; acquiring a characteristic response diagram corresponding to the image according to the central line of the angular points of the checkerboard, and carrying out preliminary angular point screening according to the characteristic response diagram; determining initial centers of the checkerboards, performing corner prediction of the checkerboards, and determining final checkerboards based on multidirectional growth. Compared with the prior art, the problem of edge gradient noise sensitivity is avoided by solving the characteristic response diagram according to the checkerboard center line principle; through the flexible and stable angular point prediction mode, detection failure caused by the fact that a certain angular point cannot be detected is avoided, the number and the size of the chequers do not need to be specified, and a plurality of chequers can be detected simultaneously.

Description

Checkerboard corner detection method and device, electronic equipment and storage medium

Technical Field

The invention relates to the technical field of camera calibration, in particular to a checkerboard corner detection method and device, electronic equipment and a storage medium.

Background

The camera calibration technology is an indispensable technology for machine vision, robot, intelligent driving and the like, and the checkerboard corner detection is the most critical step of the camera calibration technology. The quality of the detected angular points directly determines the calibration precision of the camera. At present, in the aspect of detecting the angular points of the checkerboard, available implementation schemes and patent inventions are not few.

Most of schemes described in the prior art aim at specific application scenes, the algorithm complexity is high, no open source scheme can be used for reference, the reproduction difficulty is high, and the practicability is not guaranteed;

OpenCV and matlab toolboxes have a very well available implementation, but algorithms require the specification of the number and size of checkerboards in advance, which is difficult to do in many automation applications; if the checkerboard is slightly shielded or the inclination angle is large, the detection fails; the situation that one picture contains a plurality of checkerboards cannot be processed; often, fisheye camera edges cannot be effectively identified, resulting in detection failure.

The description of the background art is only for the purpose of facilitating an understanding of the relevant art and is not to be taken as an admission of prior art.

Disclosure of Invention

According to an example embodiment of the disclosure, a method, a device, an electronic device and a storage medium for detecting checkerboard corner are provided.

In a first aspect of the present disclosure, a method for detecting checkerboard corner is provided, the method comprising:

acquiring an acquired original image, and performing image preprocessing to acquire a first image;

Acquiring a characteristic response diagram corresponding to the first image according to the central line of the checkerboard angular points, and performing preliminary angular point screening according to the characteristic response diagram;

determining initial centers of the checkerboards, performing corner prediction of the checkerboards, and determining final checkerboards based on multidirectional growth.

In a further embodiment, obtaining a feature response graph corresponding to the first image according to the checkerboard corner center line includes:

Sequentially rotating the first image according to a preset angle, and filtering according to a square frame after each rotation to find an integral;

The block filtered images are rotated in turn back to the aligned image centered on the first image, and the filtered images are combined to obtain a feature response map.

In a further embodiment, performing preliminary corner screening according to the feature response graph includes: traversing the characteristic values according to the characteristic response diagram, taking any characteristic value neighborhood, solving the maximum value in the neighborhood, and filtering the rest value in the neighborhood;

And performing preliminary corner screening according to the characteristic response diagram, wherein the method further comprises the following steps: sub-pixel extraction is carried out on the first image, and scaling is carried out according to a preset scaling proportion; and fitting saddle points according to the polynomials, wherein the saddle points are used as initial corner points.

In a further embodiment, determining a checkerboard initial center, performing checkerboard corner prediction, including an initialization step and a prediction step;

The initializing step comprises the following steps: initializing a checkerboard to 0, and taking the initial corner point as the initial center of the checkerboard; searching nearest eight neighborhood corner points according to the two main gradient directions of the initial corner point;

The predicting step includes: based on the eight neighborhood corner points in the initializing step, the position of the next adjacent corner point is predicted according to the angle difference and the modular length difference increment of the known adjacent corner points.

In a further embodiment, predicting the position of the next neighboring corner from the angular difference and the modulo length difference increment of the known neighboring corner based on the eight neighboring corner in the initializing step comprises:

Let the known three corner points be p1, p2, p3 respectively, then the two vectors v1, v2 formed by them are

v1＝p2-p1

v2＝p3-p2

The angle difference angle is:

angle＝2·b₂-b₁

b₁＝atan2(v1[1],v1[0])

b₂＝atan2(v2[1],v2[0])

the mode length difference mol is as follows:

mol＝2·d₂-d₁

d₁＝||v1||

d₂＝||v2||

the position pred of the next neighboring corner point is:

pred＝p3+ε·mol·ang

ang＝[cos(angle),sin(angle)]

wherein epsilon is an empirical constant;

The step is iterated continuously, and prediction is traversed from the four directions of up, down, left and right in whole rows or columns at a time, so that a complete checkerboard is obtained.

In a further embodiment, determining the final checkerboard based on the multi-directional growth includes:

Growing 4 chessboards from the 4 directions respectively, and finding out the chessboard with the minimum energy; if the energy of the checkerboard after growth is smaller than that before growth, the checkerboard is successfully grown, and the newly grown checkerboard is used for replacing the checkerboard before growth; otherwise, the iterative growth is carried out, and the search is continued.

In a further embodiment, acquiring an acquired original image, performing image preprocessing to acquire a first image, comprises:

collecting a plurality of images with the same visual angle, and performing superposition averaging to obtain an average image;

And carrying out Gaussian filtering on the mean value image, and then carrying out scale normalization processing on the filtered image to obtain a first image.

In a second aspect of the disclosure, there is provided a checkerboard corner detection apparatus, the apparatus comprising a preprocessing module, a feature response module, and a prediction module,

The preprocessing module is used for acquiring an acquired original image and executing image preprocessing to acquire a first image;

The characteristic response module is used for acquiring a characteristic response diagram corresponding to the first image according to the central line of the checkerboard angular points and carrying out preliminary angular point screening according to the characteristic response diagram;

And the prediction module is used for determining the initial center of the checkerboard, performing the corner prediction of the checkerboard and determining the final checkerboard based on multidirectional growth.

In a third aspect of the present disclosure, there is provided an electronic apparatus including: one or more processors, memory for storing one or more computer programs; characterized in that the computer program is configured to be executed by the one or more processors, the program comprising steps for performing the checkerboard corner detection method according to the first aspect of the present disclosure.

In a third aspect of the present disclosure, there is provided a storage medium storing a computer program; the program is loaded and executed by a processor to implement the checkerboard corner detection method steps according to the first aspect of the present disclosure.

It should be understood that the description in this summary is not intended to limit the critical or essential features of the disclosed embodiments of the invention, nor is it intended to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following description.

Optional features and other effects of embodiments of the invention are described in part below, and in part will be apparent from reading the disclosure herein.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a prior art tessellation corner detection method;

FIG. 2 is a flow chart of a checkerboard corner detection method disclosed by an embodiment of the invention;

Fig. 3 is a diagram showing a comparison between a feature of extracting a corner point disclosed in an embodiment of the present invention and a feature of extracting a corner point by a conventional method;

fig. 4 is a block diagram of an overall execution flow based on checkerboard corner detection disclosed in an embodiment of the present invention.

Detailed Description

The present application will be described in further detail with reference to the accompanying drawings, for the purpose of making the objects, technical solutions and advantages of the present application more apparent. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

It should be noted that the terms "first," "second," and the like in the description and the claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

The block diagrams depicted in the figures are merely functional entities and do not necessarily correspond to physically separate entities. That is, the functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.

The flow diagrams depicted in the figures are exemplary only, and do not necessarily include all of the elements and operations/steps, nor must they be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the order of actual execution may be changed according to actual situations.

It should be noted that: references herein to "a plurality" means two or more.

The implementation details of the technical scheme of the embodiment of the application are described in detail below:

In the traditional scheme, the detection of checkerboard corner is realized by the existence of more mature checkerboard corner detection in the OpenCV and matlab toolboxes. The method comprises the following specific steps of: performing binarization operation on the checkerboard picture by using a local average self-adaptive thresholding method; step 2: expanding and separating each black block quadrangle from the binarized image, shrinking the black block quadrangle, and breaking the connection; step 3: detecting quadrangles, and calculating convex hulls of each outline. Removing some interference quadrilaterals by utilizing constraints such as length-width ratio, perimeter, area and the like; step 4: each quadrangle is taken as a unit, two adjacent quadrangles are quadrangles at the boundary, and four adjacent quadrangles are internal quadrangles. The sequence numbers of each quadrangle can be ordered according to the adjacent relation, then two opposite points of the two quadrangles are diagonally opposite, and the middle point of the connecting line is taken as the corner point. The algorithm is simple and practical, and is the most widely applied algorithm.

Next, in some conventional schemes, a flowchart of a checkerboard corner detection method disclosed in patent application number CN201310451928.7 is shown in fig. 1. Referring to fig. 1, the drawing stool step of the detection comprises the following steps: step 100: detecting checkerboard images by using a Harris corner detection algorithm to obtain candidate corners; step 120: the coordinates of the candidate corner points are accurate to the sub-pixel level; step 130: respectively obtaining square symmetrical templates by taking each candidate corner point as a center; step 140: and processing the candidate corner points by using the square symmetrical templates, removing the pseudo corner points, and obtaining corner points. The scheme has short calculation time and high detection precision.

In the traditional scheme, the general scheme aims at a specific application scene, the algorithm complexity is high, an open source scheme is not available for reference, the reproduction difficulty is high, and the practicability is not guaranteed. OpenCV and matlab toolboxes have a very well available implementation, but (1) the algorithm requires that the number and size of checkerboards be specified in advance, which is difficult to do in many automation applications; if the checkerboard is slightly shielded or the inclination angle is large, the detection fails; the situation that one picture contains a plurality of checkerboards cannot be processed; often, fisheye camera edges cannot be effectively identified, resulting in detection failure.

Fig. 2 shows a flowchart of a checkerboard corner detection method according to an embodiment of the present invention.

The method comprises the following steps:

in step S210, an acquired original image is acquired, and image preprocessing is performed to acquire a first image.

In a further embodiment, acquiring an acquired original image, performing image preprocessing to acquire a first image, comprises: collecting a plurality of images with the same visual angle, and performing superposition averaging to obtain an average image; and carrying out Gaussian filtering on the mean value image, and then carrying out scale normalization processing on the filtered image to obtain a first image.

Specifically, the image preprocessing step of the present embodiment includes 2 steps.

Step 210-1: images of the same view angle are acquired for a plurality of times, and the images are subjected to superposition and averaging according to the following formula so as to reduce possible random thermal noise. This step has good noise reduction for cheaper cameras. Generally take n=4;

Where img represents the image after superposition averaging, img _i represents the acquired ith image.

Step 210-2: firstly, gaussian filtering is carried out on the image, and then scale normalization processing is carried out on the filtered image. Here, the gaussian kernel size kernel=5 and sigma=1 are empirically set.

Step S220, obtaining a characteristic response diagram corresponding to the first image according to the central line of the checkerboard angular points, and performing preliminary angular point screening according to the characteristic response diagram.

In a further embodiment, obtaining a feature response graph corresponding to the first image according to the checkerboard corner center line includes: sequentially rotating the first image according to a preset angle, and filtering according to a square frame after each rotation to find an integral; the block filtered images are rotated in turn back to the aligned image centered on the first image, and the filtered images are combined to obtain a feature response map.

In the embodiment of the application, the characteristic response graph is obtained according to the central line of the checkerboard angular points. Fig. 3 shows a comparison diagram of the features of the extracted corner points disclosed in the embodiment of the present application and the features of the extracted corner points in the conventional method. Therein, as shown in fig. 3 (b), the conventional algorithm basically extracts corner points with image edge gradient features, but this method is very sensitive to image noise. The present embodiment proposes to extract corner points according to the principle of symmetry center lines of the corner points. Fig. 3 (a) shows the two centerlines. The sum of all pixel values along the center line through the bright area is higher than the sum of the pixel values of the other center line through the dark area. In fact, if the two paths for summation belong to the center line of one tessellation corner, the difference between the two summations will be the largest.

In theory, the smaller the contact area through which the two paths pass, the smaller the value will be. Let the function f _c (x, y) calculate the difference between the maximum integral and the minimum integral of the pixel values for all possible centerlines around a given image point f (x, y), which can convert the image into a feature response map.

In step S220, the obtaining, according to the central line of the corner points of the checkerboard, a feature response map corresponding to the first image may specifically include the following steps:

Step S220-1: the image is rotated. Rotating the image img in step 1-2 by 0 and about the center of the image

Step S220-2: block filtering to integrate. Performing convolution operation on the image by using square block filtering with the kernel sizes of m multiplied by n and n multiplied by m respectively;

Step S220-3: to be used for And (2) rotating the angle again, and repeating the step 2-2. Until the rotation interval is [0, pi ]. This step makes it possible to determine as many corner points as possible; theoretically, the smaller the rotation angle is, the more corner points are detected, but the same corner point can be repeatedly detected;

Step S220-4: sequentially rotating the image after the block filtering back to be aligned by taking the original image as the center;

Step S220-5: the filtered images are combined to obtain a characteristic response map. This step 2 can be generalized by the following formula:

f_rotate[x,y,α]＝f[xcos(α)-ysin(α),xsin(α)+ycos(α)]

f_r[x,y,α]＝f_rotate(f_blur(f_rotate(f(x,y),-α)),α)

f_c[x,y]～[maxf_r[x,y,α]-minf_r[x,y,α]]²

Wherein the method comprises the steps of F _blur [ x, y ] represents the block filtering of the pixel (x, y) neighborhood, f _rotate [ x, y, α ] represents the rotation angle α, f _r [ x, y, α ] of the pixel (x, y) with the image center as the rotation center, and f _rotate [ x, y, α ] is the anti-rotation result of the pixel (x, y) and f _c [ x, y ] is the feature response map.

In a further embodiment, performing preliminary corner screening according to the feature response graph includes: and traversing the characteristic values according to the characteristic response diagram, taking any characteristic value neighborhood, solving the maximum value in the neighborhood, and filtering the residual value in the neighborhood.

In the embodiment of the application, the characteristic response diagram is taken, the characteristic value is traversed, a certain characteristic value neighborhood is taken, the maximum value in the neighborhood is obtained, and the rest value in the neighborhood is filtered. This step can filter out most of the clutter. Empirically, the sliding window size of the neighborhood is taken to be 5.

Specifically, in this embodiment, the subpixel extraction may directly use the opencv existing interface; scaling the image, repeating the steps 210-2, step S220, preliminary corner screening, and sub-pixel extraction to detect more and more complete corners. Empirically, scaling=2.0 when the image Height <640 and Width < 480; when Height > =640 and Width > =480, scale=0.5;

The polynomial fits saddle points, and the saddle points are used as corner points. The step can accurately position the angular points and simultaneously filter out the impurity points. The formula is as follows:

Where x _i＝x+i,y_j = y + j, k is that the sliding window is too small, a ₁,a₂,…,a₆ represents the polynomial coefficients that need to be fitted.

Step S230, determining initial centers of the checkerboards, performing corner prediction of the checkerboards, and determining final checkerboards based on multi-directional growth.

the initializing step comprises the following steps: initializing a checkerboard to 0, and taking the initial corner point as the initial center of the checkerboard; and searching nearest eight neighborhood corner points according to the two main gradient directions of the initial corner point.

In the embodiment of the application, a 3×3 checkerboard is initialized to 0, a certain corner left in the step 6 is selected as the initial center of the checkerboard, and the nearest eight neighborhood corner is searched according to two main gradient directions of the corner.

Further, on the basis of the eight neighborhoods in the initializing step, the position of the next adjacent corner point is predicted according to the angle difference and the modular length difference increment of the known 3 adjacent corner points. Wherein 3 of the known 3 neighboring corner points, i.e. the eight neighborhood corner points found in the initializing step, are more precisely 3 neighboring corner points per row and column of 3*3 corner points.

Let three corner points be p1, p2, p3 respectively, then two vectors v1, v2 formed by the corner points are

v1＝p2-p1

v2＝p3-p2

The angle difference angle is:

angle＝2·b₂-b₁

b₁＝atan2(v1[1],v1[0])

b₂＝atan2(v2[1],v2[0])

the mode length difference mol is as follows:

mol＝2·d₂-d₁

d₁＝||v1||

d₂＝||v2||

the position pred of the next neighboring corner point is:

pred＝p3+ε·mol·ang

ang＝[cos(angle),sin(angle)]

Wherein epsilon is an empirical constant; empirically, if the camera is a fisheye camera, epsilon=0.6 can be made. The step is iterated continuously, and prediction is traversed from the four directions of up, down, left and right in a whole row or a whole column each time, so that a complete checkerboard can be obtained.

In a further embodiment, determining the final checkerboard based on the multi-directional growth includes: growing 4 chessboards from the 4 directions respectively, and finding out the chessboard with the minimum energy; if the energy of the checkerboard after growth is smaller than that before growth, the checkerboard is successfully grown, and the newly grown checkerboard is used for replacing the checkerboard before growth; otherwise, the iterative growth is carried out, and the search is continued.

Specifically, in the present embodiment, a multidirectional search is performed. And respectively growing 4 chessboards from the 4 directions, and finding out the chessboard with the minimum energy. If the energy of the checkerboard after growth is smaller than that before growth, the checkerboard is successfully grown, and the newly grown checkerboard is used for replacing the checkerboard before growth; otherwise, the iterative growth is carried out, and the search is continued.

The checkerboard energy calculation formula is:

E(P,Y)＝E_corners(Y)+E_struct(P,Y)

E_corners(Y)＝-|{y|y≠O}|

Wherein, p= { P1, P2, P3 …, pn } is the corner set, Y is the label corresponding to the corner set, and O is the label of the possible miscellaneous point of the corresponding label. E (P, Y) is the total energy value and is divided into two parts: e _corners (Y) is the negative of the total number of corner points in the current board, and E _struct (P, Y) predicts the matching degree of the fourth corner point with three neighboring corner points. In an embodiment of the present application, a general execution flow diagram based on tessellation corner detection is shown in fig. 4. Step 1: preprocessing an image; step 2: solving a characteristic response graph according to the central line of the checkerboard angular points; step 3: preliminary corner screening; step 4: sub-pixel extraction; step 5: scaling the image, and repeating the steps 2 to 4; step 6: fitting saddle points by polynomials, wherein the saddle points are used as corner points; step 7: initializing a checkerboard; step 8: predicting checkerboard corner points; and 9, multi-direction searching. According to the embodiment, the characteristic response diagram is obtained by obtaining the checkerboard center line principle, so that the robustness is good; multiple corner screening techniques are used; multiple checkerboards can be detected, and failure in detecting picture corner points caused by shadows, shielding and the like is avoided.

Compared with the prior art, the checkerboard corner detection method based on the embodiment has the advantages that:

1. High accuracy: the checkerboard angular point detection algorithm provided by the invention has high accuracy and less false detection rate; the applicability to high-distortion fisheye cameras is good.

2. The robustness is good: (1) The central line principle is adopted for extraction, so that the problem of edge gradient noise sensitivity is avoided; (2) The flexible and stable corner prediction mode can not cause detection failure because a certain corner cannot be detected;

3. the applicability is strong: multiple checkerboards can be detected simultaneously without specifying the number and size of the checkerboards.

In a second aspect of the disclosure of this embodiment, there is provided a checkerboard corner detection apparatus, the apparatus including a preprocessing module, a feature response module, and a prediction module;

The characteristic response module is also used for sequentially rotating the first image according to a preset angle and integrating according to square filtering after each rotation; the block filtered images are rotated in turn back to the aligned image centered on the first image, and the filtered images are combined to obtain a feature response map.

The characteristic response module is also used for traversing the characteristic values according to the characteristic response graph, taking any characteristic value neighborhood, solving the maximum value in the neighborhood, and filtering the residual value in the neighborhood;

The prediction module comprises an initialization sub-module and a prediction sub-module;

An initialization sub-module for initializing the checkerboard to 0 and taking the initial corner point as the initial center of the checkerboard; searching nearest eight neighborhood corner points according to the two main gradient directions of the initial corner point;

And the prediction submodule is used for predicting the position of the next adjacent corner point according to the angle difference and the module length difference increment of the known adjacent corner point based on the eight adjacent corner points in the initializing step.

In a further embodiment, the prediction submodule is further configured to perform:

v1＝p2-p1

v2＝p3-p2

The angle difference angle is:

angle＝2·b₂-b₁

b₁＝atan2(v1[1],v1[0])

b₂＝atan2(v2[1],v2[0])

the mode length difference mol is as follows:

mol＝2·d₂-d₁

d₁＝||v1||

d₂＝||v2||

the position pred of the next neighboring corner point is:

pred＝p3+ε·mol·ang

ang＝[cos(angle),sin(angle)]

wherein epsilon is an empirical constant;

In a further embodiment, the prediction module is further configured to grow 4 chequers from the 4 directions, respectively, and find the chequer with the smallest energy; if the energy of the checkerboard after growth is smaller than that before growth, the checkerboard is successfully grown, and the newly grown checkerboard is used for replacing the checkerboard before growth; otherwise, the iterative growth is carried out, and the search is continued.

Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. In addition, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices, or elements, or may be an electrical, mechanical, or other form of connection.

The elements described as separate components may or may not be physically separate, and as such, those skilled in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, where the elements and steps of the examples are generally described functionally in the foregoing description of the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

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

The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.

Claims

1. A method for detecting corner points of a checkerboard, the method comprising:

2. The method for detecting a corner of a checkerboard according to claim 1, wherein obtaining a feature response map corresponding to the first image according to a corner center line of the checkerboard comprises:

And sequentially carrying out inverse rotation on the first image after the block filtering to rotate back to an image aligned by taking the first image as a center, and combining the filtered images to obtain a characteristic response diagram.

3. The checkerboard corner detection method according to claim 2, wherein the preliminary corner screening according to the characteristic response graph includes: traversing the characteristic values according to the characteristic response diagram, taking any characteristic value neighborhood, solving the maximum value in the neighborhood, and filtering the rest value in the neighborhood;

4. A method of detecting corner points of a checkerboard according to claim 3, wherein determining an initial center of the checkerboard, performing corner point prediction, includes an initializing step and a predicting step;

5. The method of detecting checkerboard corner points according to claim 4, wherein predicting the position of the next neighboring corner point based on the angular difference and the modulo length difference increment of the known neighboring corner point based on the eight neighboring corner points in the initializing step, comprises:

v1＝p2-p1

v2＝p3-p2

The angle difference angle is:

angle＝2·b₂-b₁

b₁＝atan2(v1[1],v1[0])

b₂＝atan2(v2[1],v2[0])

the mode length difference mol is as follows:

mol＝2·d₂-d₁

d₁＝||v1||

d₂＝||v2||

the position pred of the next neighboring corner point is:

pred＝p3+ε·mol·ang

ang＝[cos(angle),sin(angle)]

wherein epsilon is an empirical constant;

6. The method of detecting corner points of a checkerboard of claim 5, wherein determining a final checkerboard based on multi-directional growth includes:

7. The method of detecting checkerboard corner according to claim 1, wherein acquiring an acquired original image, performing image preprocessing to acquire a first image, includes:

8. The utility model provides a checkerboard angular point detection device which characterized in that, the device includes preprocessing module, characteristic response module, and prediction module, wherein:

9. An electronic device, the electronic device comprising: one or more processors, memory for storing one or more computer programs; characterized in that the computer program is configured to be executed by the one or more processors, the program comprising steps for performing the checkerboard corner detection method of any of claims 1-7.

10. A storage medium storing a computer program; the program being loaded and executed by a processor to implement the tessellation corner detection method steps of any of claims 1 to 7.