CN110610524A - Camera calibration point coordinate calculation method - Google Patents

Abstract

The invention belongs to the technical field of machine vision, and particularly relates to a camera calibration point coordinate calculation method, which comprises the following steps: s1, for the calibration plate image, obtaining m calibration point image coordinates (x1, y1), (x2, y2). (xm, ym) through a calibration point detection algorithm; s2, randomly selecting one index point image coordinate, and selecting another two index point image coordinates which are closest to the point according to the image coordinate distance to form three initial image coordinates which are marked as (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy 4). Based on the existing coordinates of the images of the calibration points, the row and column coordinates of the calibration points can still be accurately sequenced when the coordinates of the images of the calibration points are incomplete.

Description

Camera calibration point coordinate calculation method

Technical Field

The invention belongs to the technical field of machine vision, and particularly relates to a method for calculating coordinates of a camera calibration point.

Background

In machine vision, camera parameters are often required to be calibrated, currently, the mainstream camera calibration utilizes checkerboard calibration boards or dot calibration boards, calibration points on the calibration boards are arranged in a dot matrix and have row and column position information, a plurality of calibration board images are shot by a camera, calibration points in the calibration board images are detected through an algorithm, because the obtained calibration point image coordinates are arranged in a disordered mode and have no row and column position information and corresponding row and column coordinates, corresponding row and column coordinates need to be calculated through a specific method from the calibration point image coordinates, and finally, the camera parameters are calibrated by utilizing the image coordinates and the row and column coordinates of the calibration points in a fitting mode. Therefore, the sorting of the calibration points in the calibration board image is one of the key techniques for calibration of camera parameters.

In the prior art, envelope calculation is performed by using image coordinate points of calibration points, points located at edge positions in the points are found, the number of rows and columns of the points are calculated, and finally row and column coordinates of each calibration point are obtained. The prior art has the defects that the image coordinate points of the calibration points are required to be arranged in a complete rectangular lattice, the obtained calibration points are often incomplete due to environmental influence and failed detection of the image coordinate of the calibration points, the image coordinate of the calibration point at a certain position is possibly lost to cause failure of row and column coordinates, and the success rate of the conventional sequencing method is low.

Disclosure of Invention

Aiming at the defects of the prior art, the method for calculating the coordinates of the camera calibration points is provided under the conditions that the detection of the calibration points is incomplete and partial calibration points are missing.

In order to achieve the purpose, the invention adopts the following technical scheme:

a method for calculating coordinates of a camera calibration point comprises the following steps:

s1, for the calibration plate image, obtaining m calibration point image coordinates (x1, y1), (x2, y2). (xm, ym) through a calibration point detection algorithm;

s2, randomly selecting one index point image coordinate, and selecting another two index point image coordinates which are closest to the point according to the image coordinate distance to form three initial image coordinates which are marked as (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy 4);

s3, setting three row and column coordinates (px1, py1) ═ 0,0), (px2, py2) ═ 0,1, (px3, py3) ═ 1,0, (px4, py4) ═ 1, 1;

s4, assuming transformation parameters h1, h2, h3, h4, h5, h6, h7, h8, assuming (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy4) and (px1, py1), (px2, py2), (px3, py3), (px4, py4) satisfy the following equations:

solving equations to obtain values of h1, h2, h3, h4, h5, h6, h7 and h 8;

s5, traversing all the coordinates (x) of the index point image₁,y₁),(x₂,y₂)...(x_m,y_m) The transformation coordinates (dx) are calculated according to the following formula₁,dy₁),(dx₂,dy₂)...(dx_m,dy_m)：

S6, for transformation coordinate (dx)₁,dy₁),(dx₂,dy₂)...(dx_m,dy_m) Taking an integer by rounding to obtain an integer coordinate (Dx)₁,Dy₁),(Dx₂,Dy₂)...(Dx_m,Dy_m). And calculating rounding offset s according to₁,s₂,...s_m：

S7, setting an offset threshold t, and screening corresponding rounding offsets S₁,s₂,...s_mInteger coordinate less than t (Dx)₁,Dy₁),(Dx₂,Dy₂)...(Dx_m,Dy_m) Is denoted as (Px)₁,Py₁),(Px₂,Py₂)...(Px_n,Py_n) The number is n;

s8, setting a number threshold value nt, judging whether the current n is larger than or equal to nt, if not, repeating the steps 2 to 7 until the n is larger than or equal to nt;

s9, integer coordinate (Px) obtained by screening₁,Py₁),(Px₂,Py₂)...(Px_n,Py_n) Counting the minimum Px on X and Y coordinates respectively_min,Py_min. Calculating according to the following formula to obtain the sorted row and column coordinates (Qx)₁,Qy₁),(Qx₂,Qy₂)...(Qx_n,Qy_n)：

(Qx_i,Qy_i)＝(Px_i,Py_i)-(Px_min,Py_min)i＝1,2,3...n；

S10, and (Qx)₁,Qy₁),(Qx₂,Qy₂)...(Qx_n,Qy_n) Output as row and column coordinates.

Compared with the prior art, the method has the advantages that: randomly selecting one point of the coordinates of the index point images to be sorted, finding out multiple points adjacent to the point as a point set, assuming the row and column coordinates corresponding to the point set, and calculating a transformation parameter between the two points; and then transforming all the image coordinates of the calibration points, counting the points meeting the offset threshold condition and the transformation coordinates thereof, repeating the process until the number of the calibration points meeting the offset threshold condition is more than the number threshold, and outputting the transformation coordinates thereof as a sequencing result. The beneficial effects are that: the image coordinates of the existing calibration points are used as the basis for sequencing, so that the condition that the image coordinates of the calibration points are incomplete and are not arranged in a rectangular lattice can be overcome.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:

FIG. 1 is a flow chart of a method in an embodiment of the invention;

Detailed Description

As used in the specification and in the claims, certain terms are used to refer to particular components. As one skilled in the art will appreciate, manufacturers may refer to a component by different names. The specification and the claims do not use the difference of names as the way of distinguishing the components, but use the difference of functions of the components as the criterion of distinguishing, such as a pointer, which can be replaced by an iterator. In the following description and in the claims, the terms "include" and "comprise" are used in an open-ended fashion, and thus should be interpreted to mean "include, but not limited to. "substantially" means within an acceptable error range, within which a person skilled in the art can solve the technical problem to substantially achieve the technical result. Further, it is not necessary for certain coefficients or thresholds in the specification and claims to be specific values, but rather values are generally appropriate and some increase or decrease may be possible.

As shown in fig. 1, a method for calculating coordinates of a camera calibration point includes the following steps:

step 1, for a calibration plate image, obtaining m calibration point image coordinates (x1, y1), (x2, y2) through a calibration point detection algorithm, (xm, ym);

step 2, randomly selecting one index point image coordinate, and selecting another two index point image coordinates which are closest to the point according to the image coordinate distance to form three initial image coordinates which are marked as (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy 4);

step 3, setting three row-column coordinates (px1, py1) ═ 0,0), (px2, py2) ═ 0,1, (px3, py3) ═ 1,0, (px4, py4) ═ 1, 1;

step 4, assuming transformation parameters h1, h2, h3, h4, h5, h6, h7, h8, according to the principle of homography matrix transformation, assuming (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy4) and (px1, py1), (px2, py2), (px3, py3), (px4, py4) satisfy the following equations:

solving equations to obtain values of h1, h2, h3, h4, h5, h6, h7 and h 8;

step 5, traversing all the coordinates (x) of the calibration point image₁,y₁),(x₂,y₂)...(x_m,y_m) The transformation coordinates (dx) are calculated according to the following formula₁,dy₁),(dx₂,dy₂)...(dx_m,dy_m)。

Step 6, for the transformation coordinate (dx)₁,dy₁),(dx₂,dy₂)...(dx_m,dy_m) Taking an integer by rounding to obtain an integer coordinate (Dx)₁,Dy₁),(Dx₂,Dy₂)...(Dx_m,Dy_m). And calculating rounding offset s according to₁,s₂,...s_m：

Step 7, setting an offset threshold t, and screening corresponding rounding offsets s₁,s₂,...s_mInteger coordinate less than t (Dx)₁,Dy₁),(Dx₂,Dy₂)...(Dx_m,Dy_m) Is denoted as (Px)₁,Py₁),(Px₂,Py₂)...(Px_n,Py_n) The number is n;

step 8, setting a number threshold value nt, judging whether the current n is greater than or equal to nt, if not, repeating the steps 2 to 7 until the n is greater than or equal to nt;

and 9, counting minimum values Pxmin and Pymin on the X coordinate and the Y coordinate respectively for the integer coordinates (Px1 and Py1), (Px2 and Py2). (Pxn and Pyn) obtained by screening. The sorted row and column coordinates (Qx1, Qy1), (Qx2, Qy2). (Qxn, Qyn) were calculated as follows:

(Qx_i,Qy_i)＝(Px_i,Py_i)-(Px_min,Py_min)i＝1,2,3...n；

step 10, output as row and column coordinates (Qx1, Qy1), (Qx2, Qy 2.) (Qxn, Qyn). Randomly selecting one point of the coordinates of the index point images to be sorted, finding out multiple points adjacent to the point as a point set, assuming the row and column coordinates corresponding to the point set, and calculating a transformation parameter between the two points; and then transforming all the image coordinates of the calibration points, counting the points meeting the offset threshold condition and the transformation coordinates thereof, repeating the process until the number of the calibration points meeting the offset threshold condition is more than the number threshold, and outputting the transformation coordinates thereof as a sequencing result. The image coordinates of the existing calibration points are used as the basis for sequencing, so that the condition that the image coordinates of the calibration points are incomplete and are not arranged in a rectangular lattice can be overcome.

While the foregoing specification illustrates and describes several embodiments of the invention, it is to be understood, as noted above, that the invention is not limited to the forms disclosed herein, but is not intended to be exhaustive of other embodiments, and is capable of use in various other combinations, modifications, and environments and is capable of changes within the scope of the inventive concept as described herein, commensurate with the above teachings, or the skill or knowledge of the relevant art. And that modifications and variations may be effected by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (1)

1. A method for calculating coordinates of a camera calibration point is characterized by comprising the following steps:

s1, for the calibration plate image, obtaining m calibration point image coordinates (x1, y1), (x2, y2). (xm, ym) through a calibration point detection algorithm;

s2, randomly selecting one index point image coordinate, and selecting another two index point image coordinates which are closest to the point according to the image coordinate distance to form three initial image coordinates which are marked as (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy 4);

s3, setting three row and column coordinates (px1, py1) ═ 0,0), (px2, py2) ═ 0,1, (px3, py3) ═ 1,0, (px4, py4) ═ 1, 1;

s4, assuming transformation parameters h1, h2, h3, h4, h5, h6, h7, h8, assuming (qx1, qy1), (qx2, qy2), (qx3, qy3), (qx4, qy4) and (px1, py1), (px2, py2), (px3, py3), (px4, py4) satisfy the following equations:

solving equations to obtain values of h1, h2, h3, h4, h5, h6, h7 and h 8;

s5, traversing all the coordinates (x) of the index point image₁,y₁),(x₂,y₂)...(x_m,y_m) The transformation coordinates (dx) are calculated according to the following formula₁,dy₁),(dx₂,dy₂)...(dx_m,dy_m)：

S6, for transformation coordinate (dx)₁,dy₁),(dx₂,dy₂)...(dx_m,dy_m) Taking an integer by rounding to obtain an integer coordinate (Dx)₁,Dy₁),(Dx₂,Dy₂)...(Dx_m,Dy_m). And calculating rounding offset s according to₁,s₂,...s_m：

S7, setting an offset threshold t, and screening corresponding rounding offsets S₁,s₂,...s_mInteger coordinate less than t (Dx)₁,Dy₁),(Dx₂,Dy₂)...(Dx_m,Dy_m) Is denoted as (Px)₁,Py₁),(Px₂,Py₂)...(Px_n,Py_n) The number is n;

s8, setting a number threshold n_tJudging whether the current n is more than or equal to nt, if not, repeating the steps 2 to 7 until n is more than or equal to n_t；

S9, integer coordinate (Px) obtained by screening₁,Py₁),(Px₂,Py₂)...(Px_n,Py_n) Counting the minimum Px on X and Y coordinates respectively_min,Py_min. Calculating according to the following formula to obtain the sorted row and column coordinates (Qx)₁,Qy₁),(Qx₂,Qy₂)...(Qx_n,Qy_n)：

(Qx_i,Qy_i)＝(Px_i,Py_i)-(Px_min,Py_min) i＝1,2,3...n；

S10, and (Qx)₁,Qy₁),(Qx₂,Qy₂)...(Qx_n,Qy_n) Output as row and column coordinates.