CN115760996A

CN115760996A - Corner screening method and device, computer equipment and computer-readable storage medium

Info

Publication number: CN115760996A
Application number: CN202211340016.8A
Authority: CN
Inventors: 请求不公布姓名
Original assignee: Shenzhen Xhorse Electronics Co Ltd
Current assignee: Shenzhen Xhorse Electronics Co Ltd
Priority date: 2022-10-29
Filing date: 2022-10-29
Publication date: 2023-03-07

Abstract

The application relates to a corner screening method, a corner screening device, computer equipment and a storage medium. The method comprises the following steps: acquiring a corner point set to be screened of an image; the image comprises target mark points with target corner points; for each corner point to be screened in the corner point set to be screened, determining a target sampling range based on pixel values of pixel points in a reference sampling range of the corner point to be screened; acquiring pixel points positioned in a target sampling range to obtain a target sampling point set; and under the condition that the target sampling point set conforms to the mark point characteristics of the target mark points, determining the corner points to be screened as target corner points. The method can remove the false angular points and improve the accuracy of the obtained angular points.

Description

Corner screening method and device, computer equipment and computer-readable storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for corner screening, a computer device, and a computer-readable storage medium.

Background

The 3D scanner comprises a binocular vision structure, a light source structure and the like, and the correlation between the scanning precision and the camera calibration is strong in the using process, so that a large number of pictures need to be acquired in the calibration process, and the internal and external parameters of the camera are calculated. The calibration plate usually uses mark points with obvious characteristics, and the common mark points are divided into circular mark points and coding mark points. The circular mark point is distorted and changed into an ellipse during the photographing process, the features may change, so that the calibration plate is made with index points having angular point features. However, even if the corner has obvious features, it still extracts many false corners by using common algorithms under the influence of environment. The traditional corner screening mode has the problem that the obtained corners are inaccurate.

Disclosure of Invention

In view of the above, it is necessary to provide a corner screening method, an apparatus, a computer device and a storage medium capable of improving the accuracy of corner screening.

A method of corner screening, the method comprising:

acquiring a set of corner points to be screened of an image; the image comprises target mark points with target corner points;

for each corner point to be screened in the corner point set to be screened, determining a target sampling range based on pixel values of pixel points in a reference sampling range of the corner point to be screened;

acquiring pixel points positioned in the target sampling range to obtain a target sampling point set;

and under the condition that the target sampling point set conforms to the mark point characteristics of the target mark points, determining the corner point to be screened as a target corner point.

A device for screening angular points is disclosed, the device comprises:

the device comprises a module for acquiring a to-be-screened corner set of an image; the image comprises target mark points with target corner points;

the target sampling range determining module is used for determining a target sampling range for each corner point to be screened in the corner point set to be screened based on pixel values of pixel points in a reference sampling range of the corner point to be screened;

the target sampling point set acquisition module is used for acquiring pixel points positioned in the target sampling range to obtain a target sampling point set;

and the target corner point determining module is used for determining the corner point to be screened as the target corner point under the condition that the target sampling point set conforms to the mark point characteristics of the target mark point.

A computer device includes a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the corner screening method in the embodiment when executing the computer program.

A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the steps of the corner screening method in the embodiments of the present application.

In this embodiment, a to-be-screened corner point set in an image is obtained, and a target mark point in the image may be distorted due to the influence of factors such as a shooting angle, a shooting distance and the like, so that a target sampling range needs to be determined based on pixel values of pixel points in a reference sampling range of the to-be-screened corner points, a smaller target sampling range can be screened from a general larger range, and a more appropriate range of points which can better meet the characteristics of the mark point can be screened; and under the condition that the target sampling point set conforms to the mark point characteristics of the target mark points, determining the corner points to be screened as target corner points, eliminating false corner points through the mark point characteristics, and accurately obtaining real corner points.

Drawings

FIG. 1 is a diagram of a target landmark point with a target corner point in one embodiment;

FIG. 2 is a schematic diagram of a target landmark point with a target corner point in another embodiment;

FIG. 3 is a schematic illustration of a calibration image in one embodiment;

FIG. 4 is a schematic flow chart of a corner point screening method in one embodiment;

FIG. 5 is a schematic illustration of a reference sampling radius range in one embodiment;

FIG. 6 is a diagram illustrating directional pixel points in one embodiment;

FIG. 7 is a schematic illustration of a target sampling range in one embodiment;

FIG. 8 is a schematic view of a corner point to be screened in one embodiment;

figure 9 is a schematic diagram of a target corner point in one embodiment;

FIG. 10 is a block diagram of a corner point screening method in an embodiment;

FIG. 11 is a diagram illustrating an internal structure of a computer device in one embodiment.

Detailed Description

It should be understood that the specific embodiments described herein are merely illustrative of and not restrictive on the broad application.

The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, and not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without any inventive step are within the scope of protection of the present application.

It should be noted that all directional indicators (such as upper, lower, left, right, front, and rear … …) in the embodiments of the present application are only used to explain the relative position relationship between the components, the motion situation, etc. in a specific posture (as shown in the drawings), and if the specific posture is changed, the directional indicator is changed accordingly, and the connection may be a direct connection or an indirect connection.

In addition, descriptions in this application as to "first", "second", etc. are for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicit to the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between various embodiments may be combined with each other, but must be realized by a person skilled in the art, and when the technical solutions are contradictory or cannot be realized, such a combination should not be considered to exist, and is not within the protection scope of the present application.

The terms "first," "second," and the like as used herein may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one datum from another. For example, a first set of sample points may be referred to as a second set of sample points, and similarly, a second set of sample points may be referred to as a first set of sample points, without departing from the scope of the present application. Both the first set of sample points and the second set of sample points are sets of sample points, but they are not the same set of sample points.

It is to be understood that "connection" in the following embodiments is to be understood as "electrical connection", "communication connection", and the like if the connected circuits, modules, units, and the like have communication of electrical signals or data with each other.

In one embodiment, as shown in fig. 1, a schematic diagram of a target landmark point having a target corner point in one embodiment is shown. Fig. 2 is a schematic diagram of a target mark point with a target corner point in another embodiment. The target corner points in fig. 1 and 2 are the center points of circles. It is understood that the mark point may also be divided into 5 equal parts, 6 equal parts, 8 equal parts, etc. And the round shape can be changed into square, pentagonal, hexagonal, flower-shaped and the like. Fig. 3 is a schematic diagram of a calibration image in one embodiment. The calibration image of fig. 3 includes common calibration points and target marker points having corner points. There are 3 target markers in fig. 3. Many false corners are still extracted from the calibration image by the conventional algorithm.

Fig. 4 is a schematic flow chart of a corner screening method in an embodiment, which is applied to a computer device, where the computer device may be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices. A method of corner screening, comprising:

step 402, acquiring a set of corner points to be screened of an image; the image contains landmark points with target corner points.

Wherein the image may be a calibration image. The set of the corner points to be screened is extracted by a corner point extraction algorithm such as a Harris corner point extraction algorithm.

The corner point refers to an extreme point, i.e., a point with a more prominent pixel value attribute. The properties of the corner points may be self-defining. The corner point may particularly be located at an intersection of pixels of different pixel value ranges. The set of corner points to be screened at least includes the corner points in the mark points, and may also include other pixel points, i.e., pseudo corner points.

Specifically, the computer equipment adopts a corner extraction algorithm to obtain a set of corner points to be screened of a calibration image, wherein the image comprises target mark points with target corner points. The target corner points are used for determining target mark points in the image so as to carry out image calibration.

And step 404, determining a target sampling range for each corner to be screened in the corner to be screened set based on the pixel values of the pixel points in the reference sampling range of the corner to be screened.

The reference sampling range is a preset general sampling range which takes the corner point to be screened as a reference point. For example, the range from 8 pixels to 20 pixels in radius may be taken, or a square range may be taken with the corner to be screened as the center. The target sampling range may be a smaller range of the reference sampling ranges.

Specifically, for each corner to be screened in the corner to be screened set, the computer device may determine a mutated pixel value range based on pixel values of pixel points within a reference sampling range of the corner to be screened; taking the unmutated range as the target sampling range.

And 406, acquiring pixel points positioned in the target sampling range to obtain a target sampling point set.

Specifically, the computer device obtains pixel points located in a target sampling range to obtain a target sampling point set.

And 408, determining the corner to be screened as the target corner under the condition that the target sampling point set conforms to the mark point characteristics of the target mark point.

The mark point feature may specifically be a pixel value distribution feature of the target mark point. The pixel value distribution characteristics such as pixel value distribution are in two numerical ranges, the number of pixel points in the two numerical ranges is basically the same, or the similarity of the pixel values divided in the two numerical ranges reaches preset similarity, and the like.

Specifically, under the condition that the distribution of the target sampling point set conforms to the pixel value distribution characteristics of the target mark points, the corner point to be screened is determined as the target corner point. And when the target sampling point set does not accord with the mark point characteristics of the target mark points, the to-be-screened corner point is a pseudo corner point, and the pseudo corner point is removed.

In one embodiment, determining a target sampling range based on pixel values of pixel points in a reference sampling range of a corner point to be screened includes:

sampling along one direction in a reference sampling range of a corner point to be screened to obtain a pixel value of a directional pixel point;

determining a pixel difference value between adjacent pixel points based on the pixel values of the directional pixel points;

determining a minimum difference value from the pixel difference values, and acquiring the relative position of a pixel point corresponding to the minimum difference value; the relative position is relative to the corner point to be screened;

and determining a target sampling range based on the relative position of the pixel point corresponding to the minimum difference value.

The directional pixel points refer to pixel points in the same direction. If the angular point to be screened is taken as the original point, points within a preset distance range are taken upwards or leftwards and rightwards, and directional pixel points are obtained. The pixel difference value may refer to an absolute value. If the pixel difference between the adjacent pixel points is that the direction pixel point 1 is adjacent to the direction pixel point 2, and the direction pixel point 2 is adjacent to the direction pixel point 3, then the pixel difference between the adjacent pixel points needs to calculate the difference between the pixel value of the direction pixel point 1 and the pixel value of the direction pixel point 2, and the difference between the pixel value of the direction pixel point 2 and the pixel value of the direction pixel point 3. It can be understood that the directional pixel points may also be taken at intervals, for example, if a pixel point with a radius of 8 is taken, and the next pixel point with a radius of 10 is taken, then the pixel point with a radius of 8 is adjacent to the pixel point with a radius of 10. The distinction is configured specifically based on the neighboring algorithm. The pixels can be configured to be adjacent to each other at a distance of 1 pixel, or can be configured to be adjacent to each other at a distance of 2 pixels.

The relative position refers to the position of the pixel point relative to the corner point to be screened. Such as the distance from the corner point to be screened, etc.

Specifically, the computer device samples the pixel points along a direction in a reference sampling range of the corner point to be screened to obtain pixel values of the directional pixel points. The computer equipment determines the pixel difference value between adjacent pixel points in the direction pixel points based on the pixel values of the pixel points in each direction. The computer device determines a minimum difference value from the pixel difference values. The minimum difference value corresponds to two pixel points, the relative position of a pixel point can be chosen arbitrarily. The computer device may select a range at the same relative position as the target sampling range. If the relative position is 8 pixels, the target sampling range may be a square range with a side length of 16 and the center of the corner to be screened.

Optionally, when the number of the minimum difference values is greater than one, the relative position closest to the corner point to be screened is selected from the relative positions of the pixel points corresponding to the minimum difference values.

In this embodiment, sampling is performed along a direction in a reference sampling range of a corner point to be screened to obtain pixel values of directional pixel points, a minimum difference value is determined from pixel difference values, and a relative position of a pixel point corresponding to the minimum difference value is obtained, thereby determining a target sampling range, that is, data similarities of a certain direction at different distances are compared, a difference between pixel values in the same region is small, and a difference outside the range of the corner point is subjected to sudden change such as black to gray, or white to black, so that the acquisition of pixel points outside the target corner point can be prevented, and the accuracy of corner point screening and the accuracy of the obtained corner point are improved.

In one embodiment, sampling along a direction in a reference sampling range of a corner point to be screened to obtain pixel values of directional pixel points, includes:

in the reference sampling radius range of the angular point to be screened, taking the angular point to be screened as a reference point, and obtaining pixel values of pixel points in the directions on different sampling radii along any direction;

determining a target sampling range based on the relative position of the pixel point corresponding to the minimum difference value, wherein the target sampling range comprises the following steps:

selecting the relative radius of a target pixel point from the relative radii of the pixel points corresponding to the minimum difference;

and taking a circle formed by taking the angular point to be screened as the center of a circle and taking the corresponding relative radius of the target pixel point as the radius as the target sampling range.

The reference sampling range comprises a reference sampling radius range, and the reference sampling radius range can be a circular surface or a torus. For example, the reference sampling radius is 8-20 pixels. Any direction can be the direction taking the corner point to be screened as the starting point.

Specifically, within a reference sampling radius range of a corner point to be screened, the corner point to be screened is taken as a reference point in the direction, and the computer equipment acquires pixel values of pixel points located on different sampling radii along any direction. Based on the pixel values of the directional pixel points, the computer device determines a pixel difference value between adjacent pixel points; and determining a minimum difference value from the pixel difference values, and acquiring the relative position of a pixel point corresponding to the minimum difference value. And the computer equipment randomly selects the relative radius of a target pixel point from at least two relative radii of the pixel points corresponding to the minimum difference value. And the computer equipment takes the corner point to be screened as the center of a circle and the relative radius corresponding to the target pixel point as the radius to form a circumference as a target sampling range.

Fig. 5 is a schematic diagram of a reference sampling radius range in one embodiment. The gray areas in fig. 5 are areas within the reference sample radius. In FIG. 5, the radius range of 8 pixels to 20 pixels is taken, so the range is a torus. Fig. 6 is a schematic diagram of directional pixel points in an embodiment. The directional pixel points are the points on the gray line of fig. 6. Fig. 7 is a schematic diagram of a target sampling range in one embodiment. Assuming that the relative radius is 8, the circle formed by 8 pixels is the target sampling range. The gray circle in fig. 7 is the target sampling range.

In the embodiment, sampling is performed along one direction within a reference sampling radius range of a corner point to be screened to obtain a pixel value of a directional pixel point, a minimum difference value is determined from the pixel difference values, and a relative position of the pixel point corresponding to the minimum difference value is obtained, so that a target sampling range is determined; and a small range is screened out in a large range, so that the number of selected pixel points is small, and the angular point screening efficiency is improved.

In one embodiment, within a reference sampling radius range of a corner point to be screened, taking the corner point to be screened as a reference point, along any direction, obtaining pixel values of pixel points located in directions of different sampling radii, including:

sampling each sampling radius within the reference sampling radius range by taking the corner point to be screened as the center of a circle and rotating the sampling radius for one circle to obtain a reference pixel point set containing pixel points on different sampling radii;

and acquiring pixel values of pixels along any direction and above different radiuses from the reference pixel point set.

Specifically, with reference to the sampling radius range of 8 to 20, each sampling radius may be 8, 10, 12, 14, 16, 18, and 20; or may be 8, 9, 10, 11, 12, 13 …. And taking the corner point to be screened as the center of a circle, and rotating each sampling radius for one circle to perform sampling to obtain a reference pixel point set, wherein the reference pixel point set comprises pixel points on different sampling radii. As shown in fig. 5, the points on the circle form a reference pixel point set data. The computer equipment acquires pixel values of pixel points which are located on different radiuses along any direction from the reference pixel point set based on the positions of the reference pixel points.

In this embodiment, for each sampling radius, the corner point to be screened is taken as the center of a circle, and the sampling is performed by rotating each sampling radius by one circle, so as to obtain reference pixel point sets on different sampling radii, and pixel values of pixel points located on different radii along any direction are obtained, so that a target range within a mark point range can be obtained, and the accuracy of corner point screening is improved.

In one embodiment, determining the corner point to be screened as the target corner point when the target sampling point set conforms to the feature of the target marker point includes:

acquiring pixel values of all target sampling points in a target sampling point set;

determining the pixel average value of a target sampling point set based on the pixel value of each target sampling point;

determining a pixel minimum value and a pixel maximum value in pixel values of each target sampling point;

carrying out subtraction on the pixel average value, the pixel maximum value and the pixel minimum value in pairs to obtain a difference value set;

and when the difference value set does not contain the difference value smaller than the preset pixel value, determining the corner point to be screened as the target corner point.

The preset pixel value can be set according to needs. The preset pixel value may be a smaller value, such as between 10 and 50. Landmark features include large differences between pixel values.

Specifically, the computer device calculates the sum of the pixel values of each target sampling point, and divides the sum by the total number of the target sampling points to obtain the pixel average value of the target sampling point set. The computer device determines a minimum pixel value and a maximum pixel value from the pixel values of the target sampling points. And subtracting the pixel average value from the pixel maximum value, subtracting the pixel minimum value from the pixel average value, and subtracting the pixel minimum value from the pixel maximum value to obtain three difference values, namely a difference value set. And when the difference value set does not contain the difference value smaller than the preset pixel value, the difference between the pixel values is larger, and the corner point to be screened is the target corner point. And when the difference value set comprises a difference value smaller than a preset pixel value, the difference between the pixel values is small, and the corner point to be screened is a pseudo corner point.

In this embodiment, the pixel average value, the pixel maximum value and the pixel minimum value are calculated, and a difference set is obtained by subtracting two pixels, and when the difference set does not contain a value smaller than a preset pixel value, it is described that the difference between the pixel values is large, so that it is possible to determine the corner to be screened as the target corner.

In one embodiment, when the difference set does not include a difference smaller than a preset pixel value, determining the corner to be screened as the target corner includes:

when the difference set does not contain the difference smaller than the preset pixel value, respectively counting the first number of pixel points in the first area, the second number of pixel points in the second area, the third number of pixel points in the third area and the fourth number of pixel points in the fourth area in the target sampling point set;

determining a first ratio of the first number and the third number;

determining a second ratio of the second number to the fourth number;

and determining the corner to be screened as the target corner under the condition that the first ratio meets the first characteristic condition and the second ratio meets the second characteristic condition.

when the difference value set does not contain the difference value smaller than the preset pixel value, dividing the target sampling point set into a first sampling point set and a second sampling point set based on the pixel value of the target sampling point set;

determining a first similarity of a first set of sampling points;

determining a second similarity of the second set of sample points;

and under the condition that the first similarity meets the first similarity condition and the second similarity meets the second similarity condition, determining the corner to be screened as the target corner.

respectively counting a first number of pixel points in a first area, a second number of pixel points in a second area, a third number of pixel points in a third area and a fourth number of pixel points in a fourth area in the target sampling point set;

determining a first ratio of the first number and the third number;

determining a second ratio of the second number to the fourth number;

The mark point features, namely the ratio of the target sampling points in different areas with the same color, are within a preset ratio range. If the first characteristic condition is 0.8 to 1.2, the second characteristic condition may be the same as or different from the first characteristic condition. Different from the first, second, third and fourth regions. The color ranges of the pixel points represented by the first area and the third area are the same, and the color ranges of the pixel points represented by the second area and the fourth area are also the same. The areas can be divided according to the pixel values of the pixels, if the sector area 1 where the white pixels are located is a first area, the sector area 2 where the black pixels are located is a second area, the sector area 3 where the white pixels are located is a third area, and the sector area 4 where the black pixels are located is a fourth area. When the target mark points are distributed uniformly, the corner point may be used as a coordinate origin, the first quadrant may be used as the first area, the second quadrant may be used as the second area, the third quadrant may be used as the third area, and the fourth quadrant may be used as the fourth area.

Specifically, the computer device respectively counts a first number of pixel points located in the first region, a second number of pixel points located in the second region, a third number of pixel points located in the third region, and a fourth number of pixel points located in the fourth region. The computer device determines a first ratio of the first number to the third number and determines a second ratio of the second number to the fourth number. And determining the corner to be screened as the target corner under the condition that the first ratio meets the first ratio characteristic condition and the second ratio meets the second ratio characteristic condition. And when any ratio of the first ratio and the second ratio does not meet the corresponding characteristic conditions, determining the corner to be screened as a pseudo corner and removing the pseudo corner.

In this embodiment, since the regions in the mark points are distributed uniformly, the number of target sampling points in different regions with the same color is not greatly different, and then whether the characteristic conditions are met or not is judged by counting the number of pixel points in different regions, so that the target corner points can be screened out from numerous angular points based on the characteristics of the target mark points, the efficiency is high, and the accuracy of corner screening is improved.

In one embodiment, counting a first number of pixels located in the first region, a second number of pixels located in the second region, a third number of pixels located in the third region, and a fourth number of pixels located in the fourth region in the target sampling point set, respectively, includes:

determining a median of pixel values corresponding to the target sampling point set based on the pixel values of the target sampling points;

dividing a target sampling point set based on the median of the pixel values to obtain a first sampling point set and a second sampling point set;

counting a first number of pixel points in the first sampling point set in the first area and a third number of pixel points in the third area;

and counting a second number of pixel points in the second sampling point set in the second area and a fourth number of pixel points in the fourth area.

Specifically, the target sampling point set is divided based on the median of the pixel values, pixel points larger than the median are divided into a first sampling point set, pixel points smaller than the median are divided into a second sampling point set, and pixel points equal to the median can be divided into the first sampling point set or the second sampling point set. If the first sampling point set is a white pixel point, the second sampling point set is a black pixel point. The computer device counts a first number of pixel points located in the first area in the first sampling point set and a third number of pixel points located in the third area based on the position of each sampling point in the first sampling point set. And the computer equipment counts a second number of the pixel points in the second area and a fourth number of the pixel points in the fourth area in the second sampling point set based on the position of each sampling point in the second sampling point set.

In this embodiment, the target sampling point set is divided based on the median of the pixel values, that is, the sampling points are classified, and it is known which color, such as black or white, the sampling points corresponding to each region are, so that when it is determined whether the ratio satisfies the characteristic condition, different processing can be performed on different color regions, and the accuracy of corner processing and the universality of corner screening are improved.

dividing a target sampling point set into a first sampling point set and a second sampling point set based on pixel values of the target sampling point set;

determining a first similarity of a first set of sampling points;

determining a second similarity of the second set of sample points;

Wherein the similarity is used for representing the similarity degree between the sampling points. The similarity can specifically adopt cross entropy calculation. A smaller value of the cross entropy indicates a higher degree of similarity. The first similarity refers to a similarity of the first sampling point set. The second similarity refers to a similarity of the second set of sample points. The first similarity and the second similarity indicate different meanings and are not the same similarity.

Specifically, dividing a target sampling point set into a first sampling point set and a second sampling point set based on pixel values of the target sampling point set includes: acquiring pixel values of all target sampling points in a target sampling point set; determining a median of pixel values corresponding to the target sampling point set based on the pixel values of the target sampling points; and dividing the target sampling point set based on the median of the pixel values to obtain a first sampling point set and a second sampling point set. The first similarity condition and the second similarity condition may be the same or different. For example, the similarity condition may be that the cross entropy is smaller than the cross entropy threshold, and the cross entropy thresholds are not limited to 0.1, 0.15, 0.2, and so on. And when the first similarity meets the first similarity condition and the second similarity meets the second similarity condition, determining the corner to be screened as the target corner. And when one of the first similarity and the second similarity does not meet the corresponding similarity condition, rejecting the corner to be screened.

In this embodiment, by determining the similarity between the first sampling point set and the second sampling point set and determining whether the similarity satisfies the corresponding similarity condition, and determining the corner to be screened as the target corner under the condition that the first similarity satisfies the first similarity condition and the second similarity satisfies the second similarity condition, the real corner can be accurately screened, and the accuracy of corner screening is improved.

In one embodiment, determining a similarity of the first set of sample points comprises: determining symmetrical point pairs which are symmetrical about the corner points to be screened in the first sampling point set; a first similarity of the first set of sample points is determined based on each symmetric point pair.

Wherein, if the coordinates of the corner to be screened are (0,0), the coordinates of the point A are (x, y), and the coordinates of the point B are (-x, -y), then A and B are symmetric point pairs.

Specifically, the computer device determines symmetrical point pairs in the first sampling point set, which are symmetrical with respect to the corner point to be screened, and the symmetrical point pairs take the first area and the third area as an example. The first set of sampling points includes a (x 1, y 1), C (x 3, y 3), E (x 5, y 5), and the second set of sampling points includes B (x 2, y 2), D (x 4, y 4), F (x 6, y 6). Then cross entropy is calculated based on the pairs of symmetry points, i.e. it may be

Cross entropy = f (x 1, y 1) · log ₂ (f(x1,y1)/f(x2,y2))+

f（x3,y3）·log ₂ (f(x3,y3)/f(x4,y4))+

f（x5,y5）·log ₂ (f(x5,y5)/f(x6,y6))+

f（x2,y2）·log ₂ (f(x2,y2)/f(x1,y1))+

f（x4,y4）·log ₂ (f(x4,y4)/f(x3,y3))+

f（x6,y6）·log ₂ (f(x6,y6)/f(x5,y5))

Wherein the f-function represents the pixel value of a certain pixel.

In this embodiment, since the similarity of the symmetric point pairs is optimal, the similarity of the first sampling point set is determined based on the symmetric point pairs by determining the symmetric point pairs in the first sampling point set that are symmetric with respect to the corner to be screened, so that the corner can be prevented from being mistakenly removed, and the accuracy of corner screening is improved.

In one embodiment, determining a second similarity for a second set of sample points comprises: determining symmetrical point pairs which are symmetrical about the corner points to be screened in the second sampling point set; a second similarity of the second set of sample points is determined based on the pairs of symmetrical points.

In one embodiment, a corner screening method includes:

step (a 1), acquiring a corner point set to be screened of an image. The image contains target landmark points with target corner points.

And (a 2) for each corner point to be screened in the corner point set to be screened, sampling each sampling radius in the reference sampling radius range by taking the corner point to be screened as a circle center and rotating the sampling radius for one circle, and obtaining a reference pixel point set containing pixel points on different sampling radii.

And (a 3) acquiring pixel values of pixel points along any direction and above different radiuses from the reference pixel point set.

And (a 4) determining a pixel difference value between adjacent pixel points based on the pixel values of the directional pixel points.

And (a 5) determining a minimum difference value from the pixel difference values, and acquiring the relative position of a pixel point corresponding to the minimum difference value. The relative position is the position relative to the corner point to be screened.

And (a 6) selecting the relative radius of a target pixel point from the relative radii of the pixel points corresponding to the minimum difference value.

And (a 7) taking a circle formed by taking the to-be-screened corner point as the circle center and the corresponding relative radius of the target pixel point as the radius as the target sampling range.

And (a 8) acquiring pixel points located in a target sampling range to obtain a target sampling point set.

And (a 9) acquiring the pixel values of all the target sampling points in the target sampling point set.

And (a 10) determining the pixel average value of the target sampling point set based on the pixel value of each target sampling point.

And (a 11) determining a minimum pixel value and a maximum pixel value in the pixel values of the target sampling points.

And (a 12) carrying out pairwise subtraction on the pixel average value, the pixel maximum value and the pixel minimum value to obtain a difference value set.

And (a 13) when the difference value set does not contain the difference value smaller than the preset pixel value, determining the median of the pixel values corresponding to the target sampling point set based on the pixel values of the target sampling points.

And (a 14) dividing the target sampling point set based on the median of the pixel values to obtain a first sampling point set and a second sampling point set.

And (a 15) counting the first number of pixel points in the first area and the third number of pixel points in the third area in the first sampling point set.

And (a 16) counting a second number of pixel points located in the second area and a fourth number of pixel points located in the fourth area in the second sampling point set.

Step (a 17) of determining a first ratio of the first number and the third number.

Step (a 18) of determining a second ratio of the second number to the fourth number.

And (a 19) determining symmetrical point pairs which are symmetrical about the corner point to be screened in the first sampling point set under the condition that the first ratio meets the first characteristic condition and the second ratio meets the second characteristic condition.

Step (a 20) of determining a first similarity of the first set of sample points based on each of the symmetrical point pairs.

And (a 21) determining symmetrical point pairs in the second sampling point set, wherein the symmetrical point pairs are symmetrical about the corner to be screened.

A step (a 22) of determining a second similarity of the second set of sample points based on each symmetric point pair.

And (a 23) determining the corner to be screened as the target corner under the condition that the first similarity meets the first similarity condition and the second similarity meets the second similarity condition.

In the embodiment, a smaller target sampling range can be screened from a general larger range through range screening, ratio screening and similarity screening, so that a more appropriate range of points which can better accord with the characteristics of the mark points can be screened; and under the condition that the target sampling point set conforms to the mark point characteristics of the target mark points, determining the corner points to be screened as target corner points, eliminating false corner points through the mark point characteristics, and accurately obtaining real corner points.

In one embodiment, the design rule based on the target landmark calibration plate with the corner feature (divided into 3 target landmarks and circular landmarks) is as follows: 1. the topological relation of the three angular points and the target mark points is that a right triangle 2, the target mark points serving as short right-angle sides are adjacent 3, the two target mark points of the long right-angle sides are not adjacent and have a circular mark point at the interval between the two target mark points, the centers of the two target mark points and the centers of the circular mark points are on the same straight line 4, the connecting line of the centers of the two adjacent coding points positioned on the hypotenuse does not pass through any circular mark point 5, the target mark points are circular, and are black-white alternating fan-shaped patterns 6, and other circular mark points are sequentially arranged in a linear form.

In the calibration process, a large number of calibration images are shot, and the corner features can be extracted from the calibration images. The method and the main process for removing the pseudo corner points by the local self-adaptive threshold value in the embodiment of the application mainly comprise the steps of obtaining initial corner points, sampling corner point regions, carrying out interpolation calculation on the values of the corner point regions, selecting the optimal sampling threshold value of local coding point coordinates and judging the gray scale characteristics of the fan-shaped regions of the coding points, and finally accurately obtaining 3 real target corner point coordinates.

Firstly, extracting initial corner points of an image by using a common operator, and obtaining a set of pseudo corner points and a set of real corner points. Fig. 8 is a schematic diagram of a corner point to be screened in one embodiment. To make the figure clearer, white arrows are used to point to the corner points to be screened. And extracting the corner points by the computer equipment through a common operator to obtain the corner points to be screened in the graph 8. Before removing the false corner points, the local part of the coding point has an obvious black-white alternating pattern part, so that the coding point can be considered as a judgment basis of the real corner points. In order to accurately remove the false corner points, the neighborhood of each corner point in the corner point set is sampled, and the shape of the sampling is determined as a circular domain. The maximum value and the minimum value of the sampling radius R can be determined according to pictures of the object photographed by the camera when the distance is far and near. In this embodiment, the minimum value is set to be 8 pixels, the maximum value is 20 pixels (when the camera takes a long shot, the radius of 20 pixels is just about the edge of the encoding point), the reference sampling range is from 8 pixels to 20 pixels, the data is selected from the encoding point with the radius of 0 degree, the data is rotated by one circle counterclockwise to be sampled, and the coordinate data set coord of the reference pixel point set is obtained, as shown in fig. 7. In order to accurately obtain the pixel values of the circular sampling region, the pixel values are calculated by using bilinear interpolation to obtain a pixel value set data of a reference pixel point set.

In order to further determine a proper R, coordinates and pixels with different radiuses are randomly taken out along a certain angle, adjacent coordinates of the angle are subtracted, difference values between every two angles are sorted, the smallest difference value is selected, two relative radiuses R are found through the smallest difference value, then one relative radius is randomly selected from the two relative radiuses, the corner point to be screened is taken as the center of a circle, and the circumference formed by the relative radius which is taken as the radius is taken as a target sampling range. And sampling in a target sampling range to obtain a coordinate data set Rcoord of a target sampling point set and a pixel value set Rdata of the target sampling point set, and carrying out the next operation.

And (3) obtaining proper Rdata, then carrying out preliminary screening, obtaining the minimum value, the maximum value and the average value in the whole data set from the Rdata, carrying out subtraction on the three values in pairs, if any difference value in the subtraction process is smaller than 10 pixels (preset pixel value), indicating that the pixels near the coordinate are extremely averaged, the pixel difference is not large, judging that the pixel is a pseudo-angular point, and directly judging the next point in the set, thus the judging speed of the method can be accelerated. And if the number of the remaining points is more than 3, continuing to enter the next judgment, namely performing characteristic judgment.

The target mark point is characterized by symmetrical black and white sector areas, so that whether the Rdata has the characteristic can be judged. The Rdata data sets are sorted from small to large, repeated values in the Rdata are deleted to obtain the Rdata _ delete data sets, and then the data are classified by using the median of the Rdata _ delete data sets. And classifying to blackdata when the median is less than the median, and classifying to whitetata data set otherwise. Since Rdata takes one turn in the counterclockwise rotation order, it can be judged whether data are adjacent by coordinates. The method comprises the steps of respectively counting the number of adjacent points in blackdata and whitetata by setting thresholds which are 1~2 coordinate lengths apart, respectively solving cross entropy of the blackdata and the whitetata by utilizing pixel values and formulas, judging whether the threshold of the cross entropy is larger than 0.1, judging the blackdata to be an angular point if the threshold of the cross entropy is larger than 0.1, judging the blackdata to be a pseudo angular point if the threshold of the cross entropy is smaller than 0.1, and finishing judgment of the angular point until the judgment of the angular point is finished, so that three target angular points can be obtained. The result after removing the dummy corner points is shown in fig. 9. Figure 9 is a schematic diagram of a target corner point in one embodiment. As can be seen from fig. 9, the real corner points can be screened out by the corner point screening method.

It should be understood that, although the respective steps in the flowchart of fig. 4 described above are sequentially displayed as indicated by arrows and the respective steps in the steps (a 1) to (a 23) are sequentially displayed as indicated by reference numerals, the steps are not necessarily sequentially performed in the order indicated by the arrows or numerals. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps in fig. 4 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.

In an embodiment, as shown in fig. 10, a structural block diagram of a corner point screening apparatus in an embodiment is shown. Fig. 10 provides an apparatus for corner point screening, which may be a part of a computer device using a software module or a hardware module, or a combination of the two modules, and specifically includes: a to-be-screened point set acquisition module 1002, a target sampling range determination module 1004, a target sampling point set acquisition module 1006 and a target point determination module 1008, wherein:

a to-be-screened corner set obtaining module 1002, configured to obtain a to-be-screened corner set of an image; the image comprises target mark points with target corner points;

a target sampling range determining module 1004, configured to determine, for each corner to be screened in the set of corner points to be screened, a target sampling range based on pixel values of pixel points in a reference sampling range of the corner to be screened;

a target sampling point set obtaining module 1006, configured to obtain a pixel point located in the target sampling range, and obtain a target sampling point set;

a target corner determining module 1008, configured to determine the corner to be screened as a target corner if the target sampling point set meets the landmark feature of the target landmark.

In one embodiment, the target sampling range determination module 1004 is configured to: sampling along one direction in a reference sampling range of a corner point to be screened to obtain a pixel value of a directional pixel point;

In this embodiment, sampling is performed along a direction in a reference sampling range of a corner point to be screened to obtain pixel values of directional pixel points, a minimum difference value is determined from pixel difference values, and a relative position of the pixel point corresponding to the minimum difference value is obtained, thereby determining a target sampling range, that is, data similarities of a certain direction at different distances are compared, a difference between the pixel values in the same region is small, and the difference outside the range of the corner point is suddenly increased, so that the acquisition of the pixel points outside the target corner point can be prevented, and the accuracy of corner point screening and the accuracy of the obtained corner point are improved.

In one embodiment, the target sampling range determination module 1004 is configured to: in the reference sampling radius range of the angular point to be screened, taking the angular point to be screened as a reference point, and obtaining pixel values of pixel points in the directions on different sampling radii along any direction;

relative radius of pixel point corresponding to minimum difference selecting the relative radius of a target pixel point;

In the embodiment, sampling is performed along one direction within a reference sampling radius range of a corner point to be screened to obtain a pixel value of a directional pixel point, a minimum difference value is determined from the pixel difference values, and a relative position of the pixel point corresponding to the minimum difference value is obtained, so that a target sampling range is determined; and a small range is screened out in a large range, so that the number of selected pixels is small, and the angular point screening efficiency is improved.

In one embodiment, the target sampling range determination module 1004 is configured to:

In one embodiment, the target corner determination module 1008 is configured to:

determining a first ratio of the first number and the third number;

determining a second ratio of the second number to the fourth number;

determining a first similarity of a first set of sampling points;

determining a second similarity of the second set of sample points;

determining a first ratio of the first number to the third number;

determining a second ratio of the second number to the fourth number;

counting pixel points of the first sampling point set in the first area and a third number located in a third region;

In one embodiment, the target corner determination module 1008 is configured to: dividing a target sampling point set into a first sampling point set and a second sampling point set based on pixel values of the target sampling point set;

determining a first similarity of a first sampling point set;

determining a second similarity of the second set of sample points;

and under the condition that the first similarity meets the first similarity condition and the second similarity meets the second similarity condition, determining the corner to be screened as a target corner.

In one embodiment, the target corner determination module 1008 is further configured to: determining symmetrical point pairs which are symmetrical about the corner points to be screened in the first sampling point set; a first similarity of the first set of sample points is determined based on each symmetric point pair.

In this embodiment, since the similarity of the symmetric point pairs is optimal, the similarity of the first sampling point set is determined based on the symmetric point pairs by determining the symmetric point pairs symmetric with respect to the corner to be screened in the first sampling point set, so that the corner can be prevented from being mistakenly rejected, and the accuracy of corner screening is improved.

For the specific definition of the corner point screening means, reference may be made to the definition of the corner point screening method above, and details are not described here. All or part of the modules in the corner screening apparatus can be implemented by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a terminal device, and its internal structure diagram may be as shown in fig. 11. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor for implementing a corner screening method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

Those skilled in the art will appreciate that the architecture shown in fig. 11 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory in which a computer program is stored and a processor, which when executing the computer program performs the steps of the above-described method embodiments.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

In one embodiment, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The computer instructions are read by a processor of the computer device from the computer-readable storage medium, and the computer instructions are executed by the processor to cause the computer device to perform the steps in the above-mentioned method embodiments.

It will be understood by those skilled in the art that all or part of the processes of the methods of the above embodiments may be implemented by hardware related to instructions of a computer program, which may be stored in a non-volatile computer readable storage medium, and when executed, may include the processes of the above embodiments of the methods. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.

The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.

Claims

1. A method of corner screening, the method comprising:

acquiring a set of corner points to be screened of an image; the image comprises target landmark points with target corner points;

2. The method of claim 1, wherein determining a target sampling range based on pixel values of pixel points within a reference sampling range of the corner to be filtered comprises:

sampling along one direction in a reference sampling range of the corner point to be screened to obtain a pixel value of a directional pixel point;

determining pixel difference values between adjacent pixel points based on the pixel values of the direction pixel points;

3. The method according to claim 2, wherein the sampling along a direction within a reference sampling range of the corner point to be filtered to obtain pixel values of directional pixel points comprises:

within the reference sampling radius range of the angular point to be screened, taking the angular point to be screened as a reference point, and obtaining pixel values of pixel points positioned in different sampling radius directions along any direction;

the determining a target sampling range based on the relative position of the pixel point corresponding to the minimum difference value includes:

and taking a circumference formed by taking the angular point to be screened as the center of a circle and taking the relative radius corresponding to the target pixel point as the radius as a target sampling range.

4. The method according to claim 3, wherein the obtaining pixel values of directional pixel points located on different sampling radii along any direction with the corner to be screened as a reference point within a reference sampling radius of the corner to be screened comprises:

sampling each sampling radius within a reference sampling radius range by taking the corner to be screened as a circle center and rotating the sampling radius for one circle to obtain a reference pixel point set containing pixel points on different sampling radii;

5. The method according to claim 1, wherein the determining the corner point to be screened as the target corner point in the case that the target sampling point set conforms to the landmark feature of the target landmark point comprises:

acquiring pixel values of all target sampling points in the target sampling point set;

determining a pixel average value of the target sampling point set based on the pixel value of each target sampling point;

determining the minimum value and the maximum value of pixels in the pixel values of the target sampling points;

subtracting the pixel average value, the pixel maximum value and the pixel minimum value in pairs to obtain a difference value set;

and when the difference value set does not contain the difference value smaller than the preset pixel value, determining the corner point to be screened as a target corner point.

6. The method according to claim 1, wherein the determining the corner point to be screened as the target corner point in the case that the target sampling point set conforms to the landmark feature of the target landmark point comprises:

determining a first ratio of the first number and the third number;

determining a second ratio of the second number to the fourth number;

and under the condition that the first ratio meets a first characteristic condition and the second ratio meets a second characteristic condition, determining the corner point to be screened as a target corner point.

7. The method of claim 6, wherein said separately counting a first number of pixels located in a first region, a second number of pixels located in a second region, a third number of pixels located in a third region, and a fourth number of pixels located in a fourth region in said target sampling point set comprises:

determining the median of the pixel values corresponding to the target sampling point set based on the pixel values of the target sampling points;

dividing the target sampling point set based on the median of the pixel values to obtain a first sampling point set and a second sampling point set;

and counting the second number of the pixel points in the second area and the fourth number of the pixel points in the fourth area in the second sampling point set.

8. The method according to claim 1, wherein the determining the corner point to be screened as the target corner point in the case that the target sampling point set conforms to the landmark feature of the target landmark point comprises:

determining a first similarity of the first set of sample points;

determining a second similarity of the second set of sample points;

and under the condition that the first similarity meets a first similarity condition and the second similarity meets a second similarity condition, determining the corner to be screened as a target corner.

9. The method of claim 8, wherein determining the first similarity for the first set of sample points comprises:

determining symmetrical point pairs in the first sampling point set, wherein the symmetrical point pairs are symmetrical about the corner points to be screened;

a first similarity of the first set of sample points is determined based on each of the pairs of symmetric points.

10. An apparatus for corner screening, the apparatus comprising:

11. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 9 when executing the computer program.

12. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 9.