CN113506347A - Camera internal reference processing method and system based on single picture - Google Patents
Camera internal reference processing method and system based on single picture Download PDFInfo
- Publication number
- CN113506347A CN113506347A CN202110732249.1A CN202110732249A CN113506347A CN 113506347 A CN113506347 A CN 113506347A CN 202110732249 A CN202110732249 A CN 202110732249A CN 113506347 A CN113506347 A CN 113506347A
- Authority
- CN
- China
- Prior art keywords
- calibration plate
- calibration
- concentric
- pattern
- camera
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
- G06T7/75—Determining position or orientation of objects or cameras using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30244—Camera pose
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The application discloses a camera internal reference processing method and system based on a single picture, wherein the method comprises the following steps: obtaining a calibration image; identifying at least one concentric circle on three planes in the calibration image to obtain a two-dimensional image coordinate; selecting three straight lines outwards from the center of a concentric circle of each plane in the three planes in the calibration image, wherein each straight line in the three straight lines extends from the center of the circle to two sides and is intersected with an inner circular ring and an outer circular ring of the concentric circle of the plane, an equation is respectively established for two intersection points of each straight line passing through the center of the circle and the outer circular ring, and three equations are established in all three straight lines in each concentric circle; and solving nine equations corresponding to the three concentric circles to obtain the internal parameters of the camera. By the method and the device, the problem that a single photo only has measurement information of the space point and cannot acquire the world coordinate of the space point but needs to calculate the internal parameters is solved, and the flexibility and the precision of calibration are improved.
Description
Technical Field
The application relates to the technical field of camera calibration, in particular to a camera internal reference processing method and system based on a single picture.
Background
In computer vision, three-dimensional reconstruction with two-dimensional images requires camera calibration to extract three-dimensional structural information from the two-dimensional images. The camera calibration aims at determining the internal and external parameters and the distortion coefficient of the camera, thereby laying a foundation for computer vision. The camera calibration method not only needs to solve the internal and external parameters and distortion coefficients of the camera very accurately, but also meets the actual requirements of various application scenes. The camera calibration is not simply applied to three-dimensional reconstruction, and has many applications in various scenes, such as robot navigation, industrial control, medical diagnosis and the like, and application scenes in the 5G era will be more vigorous.
According to the dimension of the calibration object, the calibration method can be roughly divided into a camera calibration based on a three-dimensional calibration object, a camera calibration based on a two-dimensional calibration object, a camera calibration based on a one-dimensional calibration object and a camera self-calibration method.
The traditional calibration method based on the three-dimensional calibration object is to use a camera to shoot the three-dimensional calibration object with the known precise geometric shape, the method can realize camera calibration by only one image, the calibration precision is high, but the calibration device is complex and expensive, and is difficult to realize under general conditions. The calibration method based on the two-dimensional calibration object uses a two-dimensional plane template with a set pattern for calibration, the method can also obtain higher calibration precision, does not need the motion information of the plane template, is simple and easy to use, has stronger applicability, but at least needs more than three images with different angles, and calibrates based on the unchanged internal parameters. The calibration method based on the one-dimensional calibration object comprises a plurality of collinear points with known mutual distances or proportions, the one-dimensional calibration object under different poses is shot by a camera to obtain a plurality of images, the one-dimensional calibration object has the advantages of simple manufacture and capability of adapting to the situation of a plurality of cameras, but the calibration method needs more than 6 images, the calibration precision is low based on the camera calibration method of the plane template, and meanwhile, the one-dimensional calibration object needs to be controlled to do specific motion. The self-calibration method does not need any calibration object, but has good texture characteristics and low precision depending on a scene, so that the wide application of the method is limited.
Calibration methods based on two-dimensional and one-dimensional methods all require a camera to take a plurality of pictures, internal parameters calculated by different pictures are inconsistent due to the focusing effect of the camera in the moving process of the camera or a calibration object, and the deviation of the internal parameters of the camera can compensate the error of external parameters of the camera, so that calibration errors are caused, and the calibration methods cannot be used in scenes needing to calculate the attitude information of the camera.
Disclosure of Invention
The embodiment of the application provides a camera internal parameter processing method and system based on a single photo, and aims to at least solve the problem that the single photo only has measurement information of a space point and cannot obtain world coordinates of the space point but needs to calculate internal parameters in the prior art.
According to one aspect of the application, a camera internal reference processing method based on a single photo is provided, and comprises the following steps: obtaining a calibration image, wherein the calibration image is a photo obtained by shooting a calibration plate, the calibration plate is of a three-dimensional structure formed by three planes, and each plane of the three planes is drawn with at least one concentric circle or three line segments passing through known distances between two adjacent points of the same point; identifying at least one concentric circle on three planes in the calibration image to obtain a two-dimensional image coordinate; selecting three straight lines outwards from the center of a concentric circle of each plane in the three planes in the calibration image, wherein each straight line in the three straight lines extends from the center of the circle to two sides and is intersected with an outer circular ring of the concentric circle of the plane, an equation is respectively established for two intersection points of each straight line passing through the center of the circle and the outer circular ring, three equations are established for the three straight lines in each concentric circle, and the radiuses of the inner circular ring and the outer circular ring of the concentric circle in the real world are obtained through measurement; or, three equations are established for three line segments on the same plane; solving nine equations corresponding to three line segments of the three concentric circles or three line segments of the three planes, wherein the three line segments have known distances from two adjacent points of the same point, to obtain the internal parameters of the camera, wherein the unknown numbers in the nine equations comprise the internal parameters of the camera and the depth from the concentric circles to the plane of the camera.
Further, the intrinsic parameters of the camera include: the scale factor of the image coordinate system, the image principal point coordinates and the deflection parameter.
Further, the calibration board comprises a main calibration board and an auxiliary calibration board, and before the calibration image is acquired, the method further comprises: the calibration plate is formed by using a main calibration plate and an auxiliary calibration plate, the auxiliary calibration plate is folded and is placed in a three-dimensional mode with the main calibration plate, at least one concentric circle pattern is arranged on each plane in a three-dimensional space formed by two planes formed after the auxiliary calibration plate is folded and one plane of the main calibration plate, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least two concentric ring patterns, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate can be different in the same size.
Further, the calibration board comprises a main calibration board and an auxiliary calibration board, and before the calibration image is acquired, the method further comprises: the calibration plate is formed by using a main calibration plate and two auxiliary calibration plates, and each plane in a three-dimensional space formed by three calibration plates in a three-dimensional way is provided with at least one concentric circle pattern, wherein the pattern drawn on the main calibration plate comprises: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: the main calibration plate and the auxiliary calibration plate are provided with at least one concentric ring pattern, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate are the same, and the sizes of the concentric ring patterns are the same or different.
Further, the inner circle pattern of the concentric circles is divided into N parts with equal areas by straight lines starting from the center of the circle; and/or the rectangle is divided into N parts by a straight line from the center of the rectangle, wherein N is larger than or equal to 3, half of the N parts are drawn into a first color, the other half of the N parts are drawn into a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
Further, N is taken to be 4.
According to another aspect of the present application, there is provided a single-photograph-based camera internal reference processing system, including: the system comprises a calibration board and software, wherein the software is used for executing the camera internal reference processing method based on the single picture.
Further, the calibration plate includes: a main calibration plate and an auxiliary calibration plate, wherein the calibration plate is formed by placing the folded auxiliary calibration plate and the main calibration plate in a three-dimensional manner, at least one concentric circle pattern is arranged on each plane in a three-dimensional space formed by two planes formed by folding the auxiliary calibration plate and one plane of the main calibration plate, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least two concentric ring patterns, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate are the same; alternatively, the calibration plate comprises: the three calibration plates are arranged in a three-dimensional mode to form the calibration plate, each plane in the three-dimensional space of the calibration plate is provided with at least one concentric circle pattern, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: the main calibration plate and the auxiliary calibration plate are provided with at least one concentric ring pattern, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate are the same, and the sizes of the concentric ring patterns are the same or different.
Further, the inner circle pattern of the concentric circles is divided into N parts with equal areas by straight lines starting from the center of the circle; and/or the rectangle is divided into N parts by a straight line from the center of the rectangle, wherein N is larger than or equal to 3, half of the N parts are drawn into a first color, the other half of the N parts are drawn into a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
Further, N is taken to be 4.
In the embodiment of the application, a calibration image is obtained, wherein the calibration image is a picture obtained by shooting a calibration plate, the calibration plate is of a three-dimensional structure formed by three planes, and each plane of the three planes is drawn with at least one concentric circle or three line segments passing through the known distance between two adjacent points at the same point; identifying at least one concentric circle on three planes in the calibration image to obtain a two-dimensional image coordinate; selecting three straight lines outwards from the center of a concentric circle of each plane in the three planes in the calibration image, wherein each straight line in the three straight lines extends from the center of the circle to two sides and is intersected with an outer circular ring of the concentric circle of the plane, an equation is respectively established for two intersection points of each straight line passing through the center of the circle and the outer circular ring, three equations are established for the three straight lines in each concentric circle, and the radiuses of the inner circular ring and the outer circular ring of the concentric circle in the real world are obtained through measurement; or, three equations are established for three line segments on the same plane; solving nine equations corresponding to three line segments of the three concentric circles or three line segments of the three planes, wherein the three line segments have known distances from two adjacent points of the same point, to obtain the internal parameters of the camera, wherein the unknown numbers in the nine equations comprise the internal parameters of the camera and the depth from the concentric circles to the plane of the camera. The method and the device solve the problem that in the prior art, a single photo only has measurement information of the space point and cannot acquire the world coordinate of the space point but needs to calculate the internal parameters, so that the calibration precision is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:
FIG. 1 is a flow chart of a method for processing camera parameters based on a single picture according to an embodiment of the present application;
FIG. 2(a) (b) is a schematic illustration of a calibration plate according to an embodiment of the present application;
FIG. 3 is a schematic illustration of concentric circles and straight lines according to an embodiment of the present application;
FIG. 4 is a schematic view of a camera mounting and calibration plate position according to an embodiment of the present application; and
FIG. 5 is a detailed flow chart of a camera calibration implementation according to an embodiment of the present application;
FIG. 6 is a schematic diagram of an alternative to a secondary calibration plate according to an embodiment of the present application.
Detailed Description
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
The present embodiment provides a design of a camera calibration template and a new camera calibration method matched with the same. Under the condition that the focal length of the camera may be continuously and unknowingly changed, there is no good method for determining five internal parameters of the camera based on a single picture when only pixel information exists, and the embodiment realizes the calibration of the five internal parameters by using a newly designed calibration board and a matched calibration method, so that the external parameters can be further determined by the internal parameters to realize the whole calibration. The calibration method comprises the steps that a main calibration plate and a folded concentric auxiliary plate are combined to form a calibration mode which is not coplanar in pairs, the calibration information of a single image is utilized to realize camera calibration, only one concentric pattern of each of three surfaces needs to be identified in the calibration process, and other patterns are designed to calculate external parameters and improve the calibration precision. The calibration has flexibility, robustness, low cost and practicability.
In the present embodiment, a method for processing camera internal reference based on a single photo is provided, and fig. 1 is a flowchart of a method for processing camera internal reference based on a single photo according to an embodiment of the present application, as shown in fig. 1, the method includes the following steps:
step S102, obtaining a calibration image, wherein the calibration image is a picture obtained by shooting a calibration plate, the calibration plate is of a three-dimensional structure formed by three planes, and each plane of the three planes is drawn with at least one concentric circle or three line segments (as shown in figure 6) passing through the known distance between two adjacent points of the same point;
step S104, identifying at least one concentric circle on three planes in the calibration image to obtain a two-dimensional image coordinate;
step S106, selecting three straight lines outwards from the center of a concentric circle of each plane in the three planes in the calibration image, wherein each straight line in the three straight lines extends from the center of the circle to two sides and is intersected with an outer circular ring of the concentric circle of the plane, an equation is respectively established for two intersection points of each straight line passing through the center of the circle and the outer circular ring, three equations are established in all three straight lines in each concentric circle, and the radiuses of the inner circular ring and the outer circular ring of the concentric circle in the real world are obtained through measurement; or, three equations are established for three line segments on the same plane;
step S108, solving nine equations corresponding to the three concentric circles or three line segments which comprise known distances between two adjacent points of the same point on three planes to obtain the internal parameters of the camera, wherein the unknowns in the nine equations comprise the internal parameters of the camera and the depth from the concentric circles to the plane of the camera.
In the above steps, at least three concentric circles exist in one picture, wherein each concentric circle has three straight lines, each straight line has three points, and since the actual diameters of the inner ring and the outer ring of the concentric circles can be measured, sufficient equations can be established according to the values, and the unknowns in the equations are parameters required for calibration. The camera calibration can be realized by only using one picture through the steps, so that the problems in the prior art are solved.
There are a number of intrinsic parameters of the camera, which may include, for example: the scale factor of the image coordinate system, the image principal point coordinates and the deflection parameter.
There are various ways of constructing the calibration board, for example, the calibration board includes a main calibration board and an auxiliary calibration board, the calibration board is constructed by using one main calibration board and one auxiliary calibration board, the auxiliary calibration board is folded and placed in three-dimensional with the one main calibration board, each plane in a three-dimensional space formed by two planes formed by folding the one auxiliary calibration board and one plane of the one main calibration board has at least one concentric circle pattern, wherein the pattern drawn on the main calibration board includes: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least two concentric ring patterns, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate are the same, and the sizes of the concentric ring patterns can be different or the same.
For another example, the calibration board includes a main calibration board and an auxiliary calibration board, one main calibration board and two auxiliary calibration boards are used to form the calibration board, and at least one concentric circle pattern is formed on each plane in a three-dimensional space formed by three calibration boards in a three-dimensional placement manner, where the pattern drawn on the main calibration board includes: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least one concentric ring pattern, the concentric ring pattern on the main calibration plate and the concentric ring pattern on the auxiliary calibration plate are the same, and the size of the patterns can be different or the same.
For better identification of the concentric circles, in an alternative embodiment, the inner circle pattern of the concentric circles is divided into N (for example, N is 4) parts with equal areas by straight lines from the center of the circle; and/or the rectangle is divided into N parts by a straight line from the center of the rectangle, wherein N is larger than or equal to 3, half of the N parts are drawn into a first color, the other half of the N parts are drawn into a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
In this embodiment, a system for processing camera internal reference based on a single photo is provided, which includes: the system comprises a calibration board and software, wherein the software is used for executing the camera internal reference processing method based on the single picture.
In an alternative embodiment, the calibration plate may include: a main calibration plate and an auxiliary calibration plate, wherein the calibration plate is formed by placing the folded auxiliary calibration plate and the main calibration plate in a three-dimensional manner, at least one concentric circle pattern is arranged on each plane in a three-dimensional space formed by two planes formed by folding the auxiliary calibration plate and one plane of the main calibration plate, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least two concentric ring patterns, wherein the concentric ring patterns on the main calibration plate and the auxiliary calibration plate are the same, and the sizes of the concentric ring patterns can be different or the same; alternatively, the calibration plate comprises: the three calibration plates are arranged in a three-dimensional mode to form the calibration plate, each plane in the three-dimensional space of the calibration plate is provided with at least one concentric circle pattern, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and a pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least one concentric ring pattern, the concentric ring pattern on the main calibration plate and the concentric ring pattern on the auxiliary calibration plate are the same, and the size of the patterns can be different or the same.
For better identification of the concentric circles, in an alternative embodiment, the inner circle pattern of the concentric circles is divided into N (N is 4) parts with equal areas by straight lines from the center of the circle; and/or the rectangle is divided into N parts by a straight line from the center of the rectangle, wherein N is larger than or equal to 3, half of the N parts are drawn into a first color, the other half of the N parts are drawn into a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
In a preferred embodiment, a more complex calibration plate (also referred to as a calibration template) may also be used, which is described below.
In the present preferred embodiment, there is provided a camera calibration plate, including: main calibration board and supplementary calibration board, wherein, the pattern of drawing on the main calibration board includes: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least one pattern of concentric rings.
Through the calibration plate, the concentric circular ring patterns can be used for calibration, and at least one pair of rectangular patterns is used, so that the plane where the concentric circular ring patterns are located is favorably confirmed. In the case of using the camera calibration plate, calibration can be completed by taking only one picture because there are at least three concentric circle patterns in this picture.
When the camera calibration plate is used, a main calibration plate and an auxiliary calibration plate can be used as three-dimensional structures, and a plurality of methods for manufacturing the three-dimensional structures are available, for example, the number of the auxiliary calibration plates is 1, and the number of the concentric circular patterns on the auxiliary calibration plate is at least two, wherein when the auxiliary calibration plate is folded, at least one concentric circular pattern is respectively arranged on two surfaces of the auxiliary calibration plate which is folded; for another example, the number of the auxiliary calibration plates is at least 2, and each auxiliary calibration plate is provided with at least one pattern of the concentric rings. In practical use, the two modes can be flexibly selected, or more calibration plates can be used for calibration.
As an alternative embodiment, two sets of calibration plates may be used, one set being 1 secondary calibration plate and the other set being 2 secondary calibration plates, as is the case with the two examples above. Setting the two groups of calibration plates at the positions spaced by a preset distance, taking a picture by a camera, taking the two groups of calibration plates, calculating the internal parameters of the camera by using the geometric measurement of the three non-coplanar concentric circles and the sub-pixel coordinates of the corresponding characteristic points, and calculating the external parameters by using the characteristic points of the main calibration plate and the matched world coordinates. In order to make the patterns of the calibration plates easier to identify and calculate, the patterns on the main calibration plate and/or the auxiliary calibration plate are distributed according to a plurality of rows, wherein the patterns of each row are circular, or the patterns of each row are rectangular. The different rows of the main calibration plates are separated by straight lines, and one of the auxiliary calibration plates is separated by a line of rows and columns from each of the concentric circles. The two types of lines are referred to as first lines and/or second lines, which are described separately below:
a first straight line is drawn between the main calibration plate and/or the auxiliary calibration plate part or all the rows, wherein the first straight line is used for spacing the adjacent rows; the patterns on the main calibration plate and/or the auxiliary calibration plate are aligned according to columns;
and second straight lines are drawn between partial or all columns on the main calibration plate and/or the auxiliary calibration plate, wherein the second straight lines are used for separating adjacent columns, and the length of each second straight line can separate all elements or partial elements of the adjacent columns.
As another alternative, the inner part of the inner ring pattern of concentric circles and/or the matrix may also be treated: the inner circle pattern of the concentric circles is divided into N parts with equal areas by straight lines starting from the circle center; and/or the rectangle is divided into N parts (for example, 4) by a straight line from the center of the rectangle, wherein N is greater than or equal to 3, half of the N parts are drawn as a first color, the other half of the N parts are drawn as a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
When the rectangle is divided into four parts, there can be two cases: the rectangle is divided into four parts by two diagonal lines, or two middle lines respectively connecting middle points of opposite sides of the rectangle are divided into four parts with equal areas, wherein the rectangle comprises at least one of four types (only three of the four types are adopted in the embodiment):
the type I is divided into four parts by diagonal lines, and the colors are distributed alternately; type two, divided into four parts by diagonal line, the color is opposite to type one; the type III is divided into four parts by the two middle lines, and the colors are distributed at intervals; type four, divided into four parts by the two middle lines, the color is opposite to type three.
In the rectangles distributed on the main calibration plate and the auxiliary calibration plate in rows, when the rectangles are seen from two sides of the main calibration plate to the middle, the ordered combinations of each pair of rectangular patterns are different, so that the world coordinates corresponding to the central points of the patterns can be matched by identifying the pair of patterns.
In the camera calibration template based on the concentric rings provided by this embodiment, calibration patterns on the calibration plate do not need to be completely shot, and only concentric circles on three different planes need to be shot in one image, so that the internal parameters and the distortion coefficient of the camera can be calculated, fig. 2 is a schematic diagram of the calibration plate according to the embodiment of the present application, as shown in fig. 2, (a) in fig. 2 is a main calibration plate, and (b) in fig. 2 is an auxiliary calibration plate, and finally, the centers of any pair of rectangular patterns on the main calibration plate in (a) in fig. 2 can be identified to calculate the external parameters. In order to obtain a clear image, the existing camera basically has an automatic focusing function, so that the calibration of a plurality of photos is not suitable any more, in order to avoid calibration errors caused by the change of internal parameters of the camera due to focusing when the camera takes different photos, the calibration of one photo becomes a necessary trend, but the existing method for determining the internal parameters of one image needs to reduce the internal parameters for calibration, so that the difference exists between the calibration and the actual problem, in order to accurately calculate the internal and external parameters and distortion coefficients of the camera under the condition of five internal parameters, the embodiment adopts the calibration mode shown in figure 3 to finish the calibration of the camera by only taking one photo, a calibration template can be directly printed by a laser printer, equipment is simple and easy to obtain, and one calibration plate is controlled to be on a flat ground, and the other calibration plate is arbitrarily placed along with the camera at a certain angle with the calibration plate, the invention relates to a calibration method of a three-dimensional calibration object, which is characterized in that the distortion coefficient is obtained by using the existing method for calculating the distortion coefficient by concentric circles, then the internal parameters can be accurately calculated only by knowing the radius of the concentric circles, and the complete calibration of the distortion parameter and the internal and external parameters is realized by combining the traditional method for calculating the external parameters by the known internal parameters. Because the method only uses a single image, the camera calibration is very quick and can be applied to a high-speed real-time calibration scene.
In the embodiment, considering that the concentric circle pattern has the advantage of being capable of detecting the projection center, a calibration plate pattern (see fig. 2) and a calibration method for identifying and positioning other marks by using the concentric circle as a main calibration mode are developed by using a part of methods in a document of one-dimensional calibration of Zhangzhen friend 2004. The calibration plate pattern designed by the embodiment is composed of concentric rings, rectangles and straight lines, the inner circle pattern of each rectangle and each concentric ring is divided into four parts from a center point, the pattern center extraction by using sub-pixel angular points or saddle points is more accurate by using a mode of black and white blocks, the rectangles are divided into three types of patterns, the advantages of easy distinguishing and paired matching are achieved, and the calibration can be carried out by local identification. The embodiment provides the following calibration template and calibration method:
1 common camera imaging model
The imaging principle of a common camera is a pinhole imaging model, so that one three-dimensional point
Projecting onto two-dimensional image points
The corresponding formula is as follows:
s is a scale factor, (R, t) is an extrinsic parameter, R is a 3 × 3 rotation matrix, t is a translation matrix, K is an intrinsic parameter matrix, α, β are scale factors of the image (u, v) coordinate system, (u, v)0,v0) The coordinates of the principal point of the image are adopted, gamma is a deflection parameter, and the purpose of camera calibration is to determine the 5 intrinsic parameters.
The Zhang Zhengyou in 2004 carried out camera calibration based on one-dimensional calibration object, and he stringed three toy beads on a stick to form a one-dimensional object, the distance between the three beads was known, the fixed bottom point freely moved the stick, and shot 6 and more than 6 pictures with a video camera, and solved 6 unknowns with 6 and more than equations, thereby solved the inner parameter closed solution of the camera calibration model. Because the calibration method of the present embodiment is based on the calibration method, and is limited by space, the present embodiment briefly outlines the form of closed solution proposed by the rightful friend, as follows:
order to
Since B is a symmetric matrix, it can be converted into: b ═ B11 B12 B22 B13 B23 B33]TLet h be [ h ]1 h2h3]T，
Then is formed by
Can derive vTx＝L2Here, the
When N images of a one-dimensional calibration object are observed, it can be obtained by superimposing N equations: vx ═ L21, where V ═ V1,...,vN]T，1＝[1,...,1]T. Then using least square solution to obtain x-L2(VTV)-1VT1, solving the internal parameters and the depth z from the circle center to the plane of the camera by using singular value decompositionA。
2 Concentric circular ring template for camera calibration and calibration method design
Camera calibration of one-dimensional objects is based on algebraic constraints of absolute quadratic curve images, and in some degenerate or critical motion situations, the solution is not unique and the calibration algorithm will fail. The camera calibration is carried out based on the Zhangyingyou one-dimensional calibration object, under the condition that the focal length of the camera is possibly continuously and unknowingly changed, the estimation problem of the camera pose is considered in the embodiment, meanwhile, the precision of the camera calibration depends on the detection of the control points to a great extent, compared with the concentric circle pattern, the calibration pattern based on the quadrangle and the triangle is more easily influenced by image noise, distortion and resolution, and the circular pattern has the advantage of being capable of detecting a projection center, so that a calibration plate pattern and a calibration mode which take the concentric circle as a calibration mode are developed, and the calibration plate is shown in figure 2. The pattern of the calibration plate in (a) in fig. 2 is composed of concentric circles, rectangles and straight lines, and the inner circle pattern of each rectangle and concentric circle is divided into four parts from the center point, the pattern center is more accurately extracted by sub-pixel angular points or saddle points in a mode of black and white blocks being alternated, and the rectangles and concentric circles are divided into four types of patterns, so that the patterns are easy to distinguish and match. Because the embodiment recognizes the concentric circles to determine the coordinates of the two-dimensional image thereof and accurately knows the world coordinates of the concentric circles, the rectangles are divided into three different types of patterns by using the straight lines as boundary lines, and the specific recognized concentric circle can be determined by matching every two rectangles to perform calibration. The pattern can be printed on a laser printer and a good pattern for calibration should first provide sufficient geometric constraints and then be easily detected and identified. The present embodiment uses concentric circular planar patterns to perform the work of the present embodiment, having more features than other patterns. In addition, the technology only needs a camera to shoot a calibration plate on three different planes, and the calibration can be completed only by identifying the patterns on part of the calibration plate. The theoretical derivation of this calibration method is described below.
The camera calibration by the one-dimensional object also comprises the depth z from the unknown center of the circle to the camera planeA1In combination with 5 intrinsic parameters, the total number of unknowns is 6, so at least 6 equations are required. The camera calibration based on one photo of the three-dimensional object of the embodiment needs to further shoot the depth z from the center of the concentric circle to the plane of the camera on the other two planesA2，zA3There are 8 unknowns, and this embodiment needs to establish 8 equations or more to solve. In two concentric circlesWhen the radius of the calibration plate is known, it is verified that the concentric circles of each calibration plate plane can only provide 3 uncorrelated equations, so that the total number of equations provided by the three planes is 9, and the calibration condition is satisfied. By analogy with the processing method of Zhangyingyou, the 9 equations can be expressed as follows:
because the expression of b is composed of 5 intrinsic parameters, an
The above equation can be converted into the following equation system:
let yi＝α2b(i+1),i＝1,2,…,5,
j is 1,2,3, the above equation can be simplified as:
remember the formula as
Can be calculated by 8 unknowns and 9 equations
It can be easily found that when the relation is established for the first concentric circle, x in Zhangnyou 2004 is related to y
Therefore, the following formula can be used for calculating the internal parameters and the camera plane from the center of the circleDepth of face
Finally, the external parameters of the camera can be calculated by a PnP method through the combination of the central coordinate value of any pair of rectangular patterns of the recognized main calibration plate and the corresponding world coordinate. The camera calibration can be realized through the calibration template and the calibration method. The model avoids the error of the camera external parameters compensated by the change of the parameters in a plurality of pictures taken by the camera.
The following describes implementations of the present application with reference to specific embodiments.
FIG. 3 is a schematic diagram of concentric circles and straight lines, such as the center O of the concentric circles in FIG. 3, according to an embodiment of the present applicationiThe two circles are set to have radii of 15mm and 8mm, respectively, which may be set to other values, with respect to the fixed points corresponding to the one-dimensional calibration object, and thus the distance between any two points among the three points of fig. 2 is known. From the center O of the concentric circleiStarting from which three straight lines A are respectively drawniCross the two sides of the excircle at point Ai ，BiThus, similar to the one-dimensional calibration object fixed with a point rotating three angles, the analogy of zhangying friend one-dimensional calibration can obtain three equations, and the calibration method of the present embodiment is to obtain three concentric circles on different planes, fig. 4 is a schematic diagram of the camera installation and calibration plate position according to the embodiment of the present application, as shown in fig. 4, the calibration method is schematic diagram, the main calibration plate and the auxiliary calibration plate are in three-dimensional world coordinates, the camera can shoot at three planes, the camera and the auxiliary calibration plate rotate at three angles arbitrarily, in actual operation, the calibration plate in (a) in fig. 2 can be attached to a flat ground, fixed, folded from the middle by hand in the middle by using the calibration plate in (b) in fig. 2, and the camera can shoot a photo by rotatingThe calibration plate held by hand can rotate along with the camera in the process, so that three concentric circles (one concentric circle in each plane) on different calibration plate planes can be better shot from each picture, and 9 equations can be obtained from one picture to meet the calibration condition of the camera.
Fig. 5 is a specific flowchart for implementing camera calibration according to an embodiment of the present application, and as shown in fig. 5, the flowchart includes the following steps: designing a calibration template according to requirements; placing the calibration template in three different planes according to the mode; shooting a specified number of concentric rings on the calibration template by using a camera to be calibrated to obtain a calibration image; identifying concentric circles on three planes on the calibration image to obtain two-dimensional image coordinates of the calibration image, and identifying two-dimensional graph coordinates of any two pattern centers on the main calibration plate; and calculating the internal and external parameters and distortion coefficients of the camera by a formula to finish the calibration of the camera.
In this embodiment, an electronic device is provided, comprising a memory in which a computer program is stored and a processor configured to run the computer program to perform the method in the above embodiments.
These computer programs may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks, and corresponding steps may be implemented by different modules.
The programs described above may be run on a processor or may also be stored in memory (or referred to as computer-readable media), which includes both non-transitory and non-transitory, removable and non-removable media, that implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (10)
1. A camera internal reference processing method based on a single photo is characterized by comprising the following steps:
obtaining a calibration image, wherein the calibration image is a photo obtained by shooting a calibration plate, the calibration plate is of a three-dimensional structure formed by three planes, and each plane of the three planes is drawn with at least one concentric circle or three line segments passing through known distances between two adjacent points of the same point;
identifying at least one concentric circle on three planes in the calibration image to obtain a two-dimensional image coordinate;
selecting three straight lines outwards from the center of a concentric circle of each plane in the three planes in the calibration image, wherein each straight line in the three straight lines extends from the center of the circle to two sides and is intersected with an inner circular ring and an outer circular ring of the concentric circle of the plane, an equation is respectively established for the intersection point of each straight line passing through the center of the circle and the outer circular ring, three equations are established in all three straight lines in each concentric circle, and the radius of the outer circular ring of the concentric circle in the real world is obtained through measurement; or, three equations are established for three line segments on the same plane;
solving nine equations corresponding to three line segments of known adjacent two-point distance passing through the same point on the three concentric circles or three planes to obtain the internal parameters of the camera, wherein the unknown numbers in the nine equations comprise the internal parameters of the camera and the depth from the concentric circles to the plane of the camera.
2. The method of claim 1, wherein the intrinsic parameters of the camera comprise: the scale factor of the image coordinate system, the image principal point coordinates and the deflection parameter.
3. The method of claim 1, wherein the calibration plate comprises a primary calibration plate and a secondary calibration plate, the method further comprising, prior to acquiring the calibration image:
the calibration plate is formed by using a main calibration plate and an auxiliary calibration plate, the auxiliary calibration plate is folded and is placed in a three-dimensional mode with the main calibration plate, at least one concentric circle pattern is arranged on each plane in a three-dimensional space formed by two planes formed after the auxiliary calibration plate is folded and one plane of the main calibration plate, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least two concentric ring patterns, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate can be different in the same size.
4. The method of claim 1, wherein the calibration plate comprises a primary calibration plate and a secondary calibration plate, the method further comprising, prior to acquiring the calibration image:
the calibration plate is formed by using a main calibration plate and two auxiliary calibration plates, and each plane in a three-dimensional space formed by three calibration plates in a three-dimensional way is provided with at least one concentric circle pattern, wherein the pattern drawn on the main calibration plate comprises: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: the main calibration plate and the auxiliary calibration plate are provided with at least one concentric ring pattern, the concentric ring patterns on the main calibration plate and the auxiliary calibration plate are the same, and the sizes of the concentric ring patterns are the same or different.
5. The method according to any one of claims 1 to 4,
the inner circle pattern of the concentric circles is divided into N parts with equal areas by straight lines starting from the circle center; and/or the rectangle is divided into N parts by a straight line from the center of the rectangle, wherein N is larger than or equal to 3, half of the N parts are drawn into a first color, the other half of the N parts are drawn into a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
6. The method of claim 5, wherein N is taken to be 4.
7. A single-picture-based camera reference processing system, comprising: a calibration board and software, wherein,
the software is adapted to perform the method as claimed in claim 1 or 2.
8. The system of claim 7,
the calibration plate includes: a main calibration plate and an auxiliary calibration plate, wherein the calibration plate is formed by placing the folded auxiliary calibration plate and the main calibration plate in a three-dimensional manner, at least one concentric circle pattern is arranged on each plane in a three-dimensional space formed by two planes formed by folding the auxiliary calibration plate and one plane of the main calibration plate, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: the concentric circle patterns on the main calibration plate and the auxiliary calibration plate are the same, and the sizes of the concentric circle patterns are the same or different; alternatively, the first and second electrodes may be,
the calibration plate includes: the three calibration plates are arranged in a three-dimensional mode to form the calibration plate, each plane in the three-dimensional space of the calibration plate is provided with at least one concentric circle pattern, and the patterns drawn on the main calibration plate comprise: at least one concentric circular pattern and at least one pair of rectangular patterns; the pattern drawn on the auxiliary calibration plate comprises: at least one concentric ring pattern, the concentric ring pattern on the main calibration plate and the auxiliary calibration plate can be different in the same size.
9. The system of claim 7 or 8,
the inner circle pattern of the concentric circles is divided into N parts with equal areas by straight lines starting from the circle center; and/or the rectangle is divided into N parts by a straight line from the center of the rectangle, wherein N is larger than or equal to 3, half of the N parts are drawn into a first color, the other half of the N parts are drawn into a second color, the first color is different from the second color, and the colors of any two adjacent parts in the N parts are different.
10. The system of claim 9, wherein N is taken to be 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110732249.1A CN113506347B (en) | 2021-06-29 | 2021-06-29 | Camera internal reference processing method and system based on single picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110732249.1A CN113506347B (en) | 2021-06-29 | 2021-06-29 | Camera internal reference processing method and system based on single picture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113506347A true CN113506347A (en) | 2021-10-15 |
| CN113506347B CN113506347B (en) | 2022-07-26 |
Family
ID=78011323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110732249.1A Active CN113506347B (en) | 2021-06-29 | 2021-06-29 | Camera internal reference processing method and system based on single picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113506347B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116329768A (en) * | 2023-05-22 | 2023-06-27 | 广东麦特维逊医学研究发展有限公司 | Calibration method of laser marking system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020186897A1 (en) * | 2001-04-02 | 2002-12-12 | Korea Advanced Institute Of Science And Technology | Camera calibration system using planar concentric circles and method thereof |
| CN102982551A (en) * | 2012-12-14 | 2013-03-20 | 云南大学 | Method for solving intrinsic parameters of parabolic catadioptric camera linearly by utilizing three unparallel straight lines in space |
| CN109242915A (en) * | 2018-09-29 | 2019-01-18 | 合肥工业大学 | Multicamera system scaling method based on multi-face solid target |
| CN111429532A (en) * | 2020-04-30 | 2020-07-17 | 南京大学 | Method for improving camera calibration accuracy by utilizing multi-plane calibration plate |
-
2021
- 2021-06-29 CN CN202110732249.1A patent/CN113506347B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020186897A1 (en) * | 2001-04-02 | 2002-12-12 | Korea Advanced Institute Of Science And Technology | Camera calibration system using planar concentric circles and method thereof |
| CN102982551A (en) * | 2012-12-14 | 2013-03-20 | 云南大学 | Method for solving intrinsic parameters of parabolic catadioptric camera linearly by utilizing three unparallel straight lines in space |
| CN109242915A (en) * | 2018-09-29 | 2019-01-18 | 合肥工业大学 | Multicamera system scaling method based on multi-face solid target |
| CN111429532A (en) * | 2020-04-30 | 2020-07-17 | 南京大学 | Method for improving camera calibration accuracy by utilizing multi-plane calibration plate |
Non-Patent Citations (2)
| Title |
|---|
| QIAN CHEN,ET AL.: "Camera Calibration with Two Arbitrary Coplanar Circles", 《COMPUTER VISION-ECCV 2004》 * |
| 禹旺勋: "基于圆和圆环点的摄像机线性标定", 《价值工程》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116329768A (en) * | 2023-05-22 | 2023-06-27 | 广东麦特维逊医学研究发展有限公司 | Calibration method of laser marking system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113506347B (en) | 2022-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110689579B (en) | Rapid monocular vision pose measurement method and measurement system based on cooperative target | |
| CN110246185B (en) | Image processing method, device, system, storage medium and calibration system | |
| CN110956660B (en) | Positioning method, robot, and computer storage medium | |
| Kim et al. | A camera calibration method using concentric circles for vision applications | |
| CN106887023A (en) | For scaling board and its scaling method and calibration system that binocular camera is demarcated | |
| Henrique Brito et al. | Radial distortion self-calibration | |
| CN105066962B (en) | A kind of high-precision photogrammetric apparatus of the big angle of visual field of multiresolution | |
| CN113841384B (en) | Calibration device, chart for calibration and calibration method | |
| CN102376089A (en) | Target correction method and system | |
| CN113920205B (en) | Calibration method of non-coaxial camera | |
| CN109255818B (en) | Novel target and extraction method of sub-pixel level angular points thereof | |
| CN115457147A (en) | Camera calibration method, electronic device and storage medium | |
| JP2016218815A (en) | Calibration device and method for line sensor camera | |
| CN110136048B (en) | Image registration method and system, storage medium and terminal | |
| CN114299156A (en) | Method for calibrating and unifying coordinates of multiple cameras in non-overlapping area | |
| CN113506347B (en) | Camera internal reference processing method and system based on single picture | |
| CN111699513B (en) | Calibration plate, internal parameter calibration method, machine vision system and storage device | |
| CN114549660A (en) | Multi-camera calibration method, device and equipment based on cylindrical self-identification marker | |
| JP2006098065A (en) | Calibration device and method, and three-dimensional modelling device and system capable of using the same | |
| CN111179347B (en) | Positioning method, positioning equipment and storage medium based on regional characteristics | |
| CN110458951B (en) | Modeling data acquisition method and related device for power grid pole tower | |
| CN111968182A (en) | Calibration method for binocular camera nonlinear model parameters | |
| US20230245346A1 (en) | Method and device for corner marking and parameter calibration | |
| CN116205993A (en) | Double-telecentric lens high-precision calibration method for 3D AOI | |
| JP4548228B2 (en) | Image data creation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |