CN112258586A - Calibration method for stereoscopic vision model parameters of single plane mirror - Google Patents
Calibration method for stereoscopic vision model parameters of single plane mirror Download PDFInfo
- Publication number
- CN112258586A CN112258586A CN202011112823.5A CN202011112823A CN112258586A CN 112258586 A CN112258586 A CN 112258586A CN 202011112823 A CN202011112823 A CN 202011112823A CN 112258586 A CN112258586 A CN 112258586A
- Authority
- CN
- China
- Prior art keywords
- model
- parameters
- camera
- stereoscopic vision
- distortion
- 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
- 238000000034 method Methods 0.000 title claims abstract description 45
- 239000011159 matrix material Substances 0.000 claims abstract description 33
- 230000000007 visual effect Effects 0.000 claims abstract description 28
- 238000005259 measurement Methods 0.000 claims abstract description 24
- 238000005457 optimization Methods 0.000 claims abstract description 3
- 238000003384 imaging method Methods 0.000 claims description 19
- 230000003287 optical effect Effects 0.000 claims description 14
- 230000009466 transformation Effects 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims description 2
- 238000013519 translation Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001808 coupling effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 208000013021 vision distortion Diseases 0.000 description 1
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
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/002—Measuring arrangements characterised by the use of optical techniques for measuring two or more coordinates
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Measurement Of Optical Distance (AREA)
Abstract
A calibration method for stereoscopic vision model parameters discloses a method for establishing a stereoscopic vision model and solving model parameters. And introducing a reflection matrix of the plane mirror, establishing visual models of the real camera and the virtual camera, and deducing a three-dimensional measurement model of the single plane mirror stereoscopic vision. And calibrating each parameter of the monocular stereoscopic vision model by adopting a step method, and solving initial values of the parameters in the real camera and the virtual camera in a small range of the image center area. And respectively calibrating the radial distortion coefficient and the eccentric distortion coefficient of the two cameras within the depth of field range. And on the premise of locking the distortion coefficients of the two cameras, respectively carrying out optimization solution on the internal parameters and the external parameters in the respective models, thereby realizing the calibration of the parameters of the stereoscopic vision model of the single plane mirror. The modeling process of the invention is simple, and the model is easy to apply. In addition, the step-by-step calibration method of firstly distortion coefficients and then internal and external parameters avoids the mutual influence between model parameter errors and improves the calibration precision of the parameters of the stereoscopic vision model of the single plane mirror.
Description
Technical Field
The invention belongs to the field of computer vision measurement, and relates to a method for calibrating parameters in a stereoscopic vision model.
Background
The vision measurement technology has the advantages of high real-time performance, strong robustness, excellent measurement precision and three-dimensional full-field measurement, and has been widely concerned by the industry and academia. According to the image processing result and the calibrated camera model parameters, the technology can realize quantitative expression of scene information. The binocular camera is a common configuration form for realizing three-dimensional measurement, however, the measurement mode has the disadvantages of high measurement cost and difficulty in synchronous triggering of the binocular camera. The monocular stereoscopic vision system consisting of the single plane mirror and the single camera effectively solves the pain points. The accurate solving of the imaging model parameters is crucial to improving the visual detection precision of the object to be detected. The model parameters to be solved include internal parameters, external parameters and distortion coefficients. The distortion of the camera is closely related to the position of an object point in a depth of field range, and when short-object-distance and long-focus close-range parameters are adopted for imaging, the imaging distortion of the lens in the depth of field range is particularly serious, so that the distortion becomes a main factor for restricting the improvement of the vision measurement precision. Therefore, the distortion is accurately solved, and the accurate calibration of other parameters in the monocular stereoscopic vision imaging model is realized, so that the method has important significance for improving the vision measurement precision.
The patent number ZL 108253939A of 'variable visual axis monocular stereoscopic vision measuring method' invented by Lianshu et al, university of Tongji invents a method for calibrating parameters of variable visual axis monocular stereoscopic vision, and the method firstly calibrates internal parameters of a monocular camera by adopting methods such as direct linear transformation and the like without considering the condition of a rotating biprism. Then, the rotating double prism is arranged in front of the monocular camera to form a monocular stereoscopic vision system, and the system is calibrated to obtain the relative position relationship between the camera and the rotating prism. The calibration method disclosed by the invention does not consider the mutual influence of the solution errors of the internal parameters, the external parameters and the distortion coefficients in the visual model. In addition, the rotating biprisms need to be removed and introduced during calibration, and the calibration process is complicated. CN 202010157954.9 patent No. CN 202010157954.9, "correction method of image point in monocular stereoscopic vision image", invented a correction method for stereoscopic vision distortion point of single plane mirror, which reduces the visual vertical parallax by making the polar line and the real imaging plane or the virtual imaging plane not intersect at the extreme point, and on this basis, reduces the search range of the matching point in the corresponding viewing angle from the two-dimensional image to the one-dimensional straight line, thus significantly improving the processing accuracy and speed. The invention does not consider the influence of the depth of field position on imaging distortion, and moreover, the disclosure only relates to the calibration of the distortion and does not mention the calibration of other parameters in the visual model.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a calibration method for parameters of a stereoscopic vision model of a single plane mirror. The method comprises the following steps: establishing a monocular stereoscopic vision model and solving parameters in the model, introducing a reflection matrix of a plane mirror into the monocular imaging model in the aspect of establishing a measurement model, and respectively obtaining visual imaging models of a real camera and a virtual camera; on the basis, a three-dimensional measurement model of monocular stereovision is deduced; in the aspect of calibrating model parameters, various parameters of the monocular stereoscopic vision model are calibrated by adopting a step method, wherein the step method is to solve initial values of internal parameters in a real camera model and a virtual camera model in a small range of an image central area. Then, the radial distortion coefficient and the eccentric distortion coefficient of the two cameras in the depth of field range are respectively calibrated. And finally, respectively carrying out optimization solution on the internal parameters and the external parameters in the respective models on the premise of locking the distortion coefficients of the two cameras, thereby realizing the calibration of parameters of the stereoscopic vision model of the single plane mirror.
According to the calibration method for parameters of the stereoscopic vision model of the single plane mirror, the establishment of the monocular stereoscopic vision model and the monocular stereoscopic vision three-dimensional measurement model is realized by introducing the reflection matrix of the plane mirror, the modeling process is simplified, and the application of the model is facilitated. In addition, the step-by-step calibration method of solving the distortion coefficient and then solving the internal parameter and the external parameter avoids the coupling effect between model parameter errors and improves the calibration precision of the stereoscopic model parameters of the single plane mirror. A calibration method for parameters of a stereoscopic vision model of a single plane mirror is as follows:
(1) monocular vision model
The monocular stereoscopic vision system composed of a plane mirror and a camera equivalently comprises two cameras: real cameras and virtual cameras. For real cameras, let the real camera coordinate system or-xryrzrThe lower spatial point is Pr(xr yr zr) Real camera coordinate system and world coordinate system ow-xwywzwThe transformation matrix between, i.e. the extrinsic parameter matrix of the real camera is [ R ]r Tr]。p(ur vr) Is a space point directly on an image surface coordinate system o without a reflectorr-urvrThe visual model of the real camera can be represented by equation (1) as:
wherein s isrIs a scale factor, alpharAnd betarAre respectively defined as ur、vrNormalized focal length in both axial directions, (u)0 v0) Is the intersection point coordinate of the real camera optical axis and the imaging plane.
For the virtual camera, let the virtual camera coordinate system ov-xvyvzvAnd the world coordinate system ow-xwywzwThe transformation matrix between, i.e. the extrinsic parameter matrix of the virtual camera is [ R ]v Tv]. The visual model of the virtual camera can be represented by equation (2):
wherein, p (u)v vv) Virtual image point, Pv(xv yv zv) Is Pr(xr yr zr) For pairs of plane mirrorsThis is called the dot. svIs a scale factor, alphavAnd betavAre respectively defined as uv、vvNormalized focal length in both axial directions.
According to the mirror symmetry property, Pv(xv yv zv) And Pr(xr yr zr) The relationship between them can be expressed by the formula (3):
unfolding to obtain:
wherein (n)x ny nz) Is a unit normal vector of the plane mirror, satisfies nx·nx+ny·ny+nz·nz1. d is the real camera coordinate system or-xryrzrOrigin orPerpendicular distance to the plane. Then equation (2) can be re-expressed as:
wherein,
and
respectively correspond to or-xryrzrAnd ov-xvyvzvA rotation matrix and a translation matrix in between.
(2) Monocular stereoscopic vision three-dimensional measurement model
The formula (1) is arranged to obtain:
obtaining by arranging the formula (2):
as can be seen from formula (3), formula (4), formula (5), formula (6) and formula (7), p (u)r vr) And p (u)v vv) The relationship between them can be expressed by the formula (8):
the coordinates of the space points in the real coordinate system are solved by the following formula:
the formula (9) is the established three-dimensional measurement model of monocular stereoscopic vision.
(3) Visual model parameter calibration
1) Initial value calibration of internal parameter
Considering the influence of distortion, solving the parameter matrix K in the camera by using the Zhang scaling method in a small range of the image central regionrAnd KvAnd obtaining the focal length f of the lens.
2) Distortion coefficient calibration
Manufacturing and assembly defects of the lens cause radial distortion and eccentric distortion, and the lens distortion can be expressed by equation (10):
wherein, (u v) is the coordinate of image point; deltau、ΔvThe distortion produced for an image point is at the horizontal axis u of the imager(uv) And a vertical axis vr(vv) The component of (a);
is the distortion radius; k is a radical of1、k2First two order coefficients, p, of radial distortion, respectively1And p2The first two order coefficients of the eccentric distortion, respectively. Considering the effect of depth of field on lens imaging distortion, the radial distortion coefficient can be expressed by equation (11):
wherein,
is a magnification factor, and
respectively, the focus distance within the depth of field is sn、sm、skIs calculated by the radial distortion coefficient of the ith order on the focal plane. For a real camera in monocular stereoscopic vision, a checkerboard standard object is used as a calibration plate, the calibration plate is kept perpendicular to the optical axis of the real camera, and the calibration plate is placed in the depth of field range along the optical axis so as to be respectively positioned at a focusing distance smAnd skOn the two focal planes. Then, the radial distortion coefficients of each order on the two focusing planes are calculated respectively
And
finally, the focusing distance s can be calculated by the formula (11)nIs the ith radial distortion coefficient on the focusing plane
After the virtual camera is processed by the same method, radial distortion coefficients of each order of the real camera and the virtual camera can be obtained.
The off-center distortion factor that takes into account the effect of the depth of field factor can be expressed by equation (12):
wherein,
and
respectively, the focus distance within the depth of field is snAnd smThe ith order eccentric distortion coefficient on the two pairs of focal planes, and f is the focal length of the lens. For a real camera in monocular stereoscopic vision, a checkerboard standard object is used as a calibration plate, the calibration plate is kept perpendicular to the optical axis of the real camera, and the calibration plate is placed in the depth of field range along the optical axis so as to be in focus at a distance smOn the focal plane of (a). Then, the eccentric distortion coefficient of each order on the focusing plane is calculated
Finally, the focusing distance s can be calculated by the formula (12)nIs the ith radial distortion coefficient on the focusing plane
After the virtual camera is processed by the same method, the eccentric distortion coefficients of each order of the real camera and the virtual camera can be obtained.
3) Internal and external parameter calibration
And 2) respectively obtaining coefficients of each order of radial distortion and eccentric distortion in the visual models of the real camera and the virtual camera. Then, in order to avoid the influence of the solving error of the visual internal and external parameters on the distortion coefficient, the value of the distortion coefficient is respectively fixed for each camera, the distance provided by the checkerboard angular points is taken as the constraint,adopting LM algorithm to optimize and solve the internal parameter matrix and the external parameter matrix to respectively obtain Kr、Kv、[Rr Tr]And [ R ]v Tv]. Finally, a conversion matrix [ R ' T ' between real camera and virtual camera is computed ']。
And further completing the calibration of the parameters of the stereoscopic vision model of the single plane mirror.
The beneficial results of the invention are that the imaging model and the three-dimensional measurement model of the single plane mirror stereoscopic vision are established by means of the reflection matrix of the plane mirror, the measurement principle of the single plane mirror stereoscopic vision is explained, and the model is simple and easy to implement. In addition, the method for calibrating the parameters of the visual model step by step is provided, and the problem that the calibration precision of the parameters of the model is reduced due to the interactive influence of errors is avoided by calibrating the distortion coefficients and then optimizing and solving the internal and external parameters, so that the error solving precision of the stereoscopic vision of the single plane mirror is effectively improved.
Drawings
FIG. 1 is a schematic diagram of a single plane mirror stereovision system.
FIG. 2 is a flow chart of calibration of parameters of a single plane mirror stereo vision model.
In the figure: 1-real camera, 2-plane mirror, 3-virtual camera.
Detailed Description
The following describes the embodiments of the present invention in detail with reference to the technical solutions and the accompanying fig. 1 and 2.
FIG. 1 is a schematic diagram of a single plane mirror stereovision system. FIG. 2 is a flow chart of calibration of parameters of a single plane mirror stereo vision model.
The invention relates to a calibration method for parameters of a stereoscopic vision model of a single plane mirror, which comprises the steps of establishing a vision model and solving parameters of the model. Firstly, introducing a reflection matrix of a plane mirror 2 into a monocular imaging model to respectively obtain visual imaging models of a real camera 1 and a virtual camera 3; on the basis, a three-dimensional reconstruction formula of the space point is deduced; in the aspect of calibrating model parameters, the parameters of the monocular stereoscopic vision model are calibrated by adopting a step-by-step methodFirstly, the step method is to define the initial values of the parameters in the two cameras in a small range of the central area of the image. And then, respectively calibrating the radial distortion coefficient and the eccentric distortion coefficient of the two cameras within the field depth range, finally locking the distortion coefficients of the two cameras, and completing the calibration of the monocular stereoscopic vision model parameters on the basis of optimally solving the respective internal parameters and external parameters of the two cameras by utilizing an LM algorithm. In an embodiment of the invention, the plane mirror 2 is perpendicular to the world coordinate system ow-xwywzwO of (a)wxwzwPlaced in a plane withwxwThe included angle of (a) is 90 deg.. or-xryrzrDistance to plane mirror 2 is 100mm, o of coordinate system of real camera 1rxrzrO of plane and world coordinate systemwxwzwAnd (4) overlapping. Let the real camera coordinate system or-xryrzrThe coordinates of the lower spatial point are
＝[85.82,160.45,0,1]The theoretical focal length of the camera is 17mm, and the camera resolution is 2560 pixels × 2560 pixels. The following detailed description is made of specific embodiments:
1. monocular vision model
The monocular stereoscopic vision system consisting of a flat mirror 2 and a camera comprises two cameras: a real camera 1 and a virtual camera 3. For the real camera 1, let the transformation matrix between the real camera coordinate system and the world coordinate system, i.e. the extrinsic parameter matrix of the real camera 1 be [ Rr Tr]Program code for creating a visual model of a real camera 1 is compiled according to equation (1). For the virtual camera 3, let the virtual camera coordinate system ov-xvyvzvAnd the world coordinate system, i.e. the external parameter matrix of the virtual camera 3 is [ R ]v Tv]The program code of the visual model of the virtual camera 3 is programmed according to equation (2).
Introducing a normal vector of the plane mirror 2, and establishing P according to a formula (3) according to the symmetry property of the plane mirror 2v(xv yv zv) And Pr(xr yr zr) And developing the formula (3) to obtain a formula (4). Then, based on the result of the formula (4), the formula (2) is rearranged to obtain the formula (5), and the code of the content shown in the formula (5) is programmed.
2. Monocular stereoscopic vision three-dimensional measurement model
The formula (1) is arranged to obtain the formula (6), and the formula (2) is arranged to obtain the formula (7). Then, p (u) is established according to formula (3), formula (4), formula (5), formula (6) and formula (7)r vr) And p (u)v vv) The relationship between them, the formula (8) is obtained. And finally, expanding the formula (8) to obtain a formula (9), namely completing the establishment of the monocular stereoscopic vision three-dimensional measurement model, and compiling a code of the content shown in the formula (9) in a program.
3. Visual model parameter calibration
1) Initial value calibration of internal parameter
Under the condition of no distortion, calibrating the parameter matrix K in the camera by using the Zhang calibration methodrAnd KvThe lens distortion code is programmed according to the formula (10), and the intrinsic parameter f of the camera model is solved to be 17.24mm by adopting the Zhang Yong calibration method.
2) Distortion coefficient calibration
In consideration of the influence of the depth factor on the imaging distortion of the lens, a radial distortion coefficient code shown in formula (11) is compiled. For a real camera 1 in monocular stereoscopic vision, a checkerboard standard object is used as a calibration plate, the calibration plate is kept perpendicular to the optical axis of the real camera 1, and the calibration plate is placed in the depth of field range along the optical axis so as to be respectively positioned at a focusing distance sm300mm and sk400mm in two focal planes. Firstly, the focusing distance s is calculatedmThe first two radial distortion coefficients in the 300mm focal plane are
And
is calculated toDistance to focus of skThe first two-order radial distortion coefficients on the 400mm focal plane are
And
the focusing distance s is obtained by calculation according to the formula (11)nFirst two-order radial distortion coefficient on 350mm focal plane
And
calibrating the radial distortion coefficient of the virtual camera 3 by the same method, wherein two focusing distance values used in the experiment are consistent with the two focusing distance values of the real camera 1, and the focusing distance s is obtained by calculating through a formula (11)nFirst two-order radial distortion coefficient on 350mm focal plane
And
compiling a program code corresponding to the eccentric distortion coefficient shown in formula (12), regarding the real camera 1 in monocular stereoscopic vision, taking a checkerboard standard object as a calibration plate, keeping the calibration plate perpendicular to the optical axis of the real camera 1, and placing the calibration plate in the depth of field range along the optical axis so that the calibration plate is in focus at a distance smOn a focusing plane of 300mm, the first two-step eccentric distortion coefficients on the focusing plane are calculated respectively
And
then, the focusing distance s is calculated according to the formula (12)nFirst two steps of deviation of 350mmCoefficients of distortion in the cardioid direction of
And
the eccentric distortion coefficient of the virtual camera 3 is calibrated by the same method, the focusing distance value used in the experiment is also 300mm, and the first two-step eccentric distortion coefficient on the focusing plane with the focusing distance of 350mm is obtained through calculation of a formula (12)
And
through the steps, the radial distortion coefficient and the eccentric distortion coefficient on any focusing plane in the depth of field are solved.
3) Internal and external parameter calibration
By 2) the respective order coefficients of radial distortion and eccentric distortion in the visual model of the real camera 1 and the virtual camera 3 can be obtained respectively. Then, in order to avoid the influence of the solving error of the visual internal and external parameters on the distortion coefficient, the value of the distortion coefficient is respectively fixed for each camera, the distance of the checkered corner points is taken as constraint, the LM algorithm is adopted to optimize and solve the internal parameters and the external parameters in the visual model, and the obtained internal parameter matrix is
Finally, by
And
computing a conversion matrix [ R ' T ' between real Camera 1 and virtual Camera 3 ']，
Finally, the calibration of the parameters of the monocular stereoscopic vision model is completed.
The invention relates to a calibration method for parameters of a stereoscopic vision model of a single plane mirror. An imaging model and a three-dimensional measurement model of the single plane mirror stereoscopic vision are established according to the reflection matrix of the plane mirror, and the model is simple and convenient to use. In addition, the parameters in the visual model are calibrated by a step-by-step method, and the distortion coefficient is calibrated first and then the internal and external parameters are calibrated in the method, so that the interaction influence among model parameter errors is avoided, and the calibration precision of the model parameters is improved.
Claims (1)
1. A calibration method for parameters of a monocular mirror stereoscopic vision model is characterized in that a monocular stereoscopic vision model is built, parameters in the model are solved, and in the aspect of building a measurement model, a reflection matrix of a mirror is introduced into a monocular imaging model to obtain visual imaging models of a real camera and a virtual camera respectively; on the basis, a three-dimensional measurement model of monocular stereovision is deduced; in the aspect of model parameter calibration, various parameters of a monocular stereoscopic vision model are calibrated by adopting a step method, wherein the step method is to solve initial values of internal parameters in a real camera model and a virtual camera model in a small range of an image central area; then, respectively calibrating the radial distortion coefficient and the eccentric distortion coefficient of the two cameras within the field depth range; finally, on the premise of locking distortion coefficients of the two cameras, respectively carrying out optimization solution on internal parameters and external parameters in respective models, and realizing calibration of parameters of the stereoscopic vision model of the single plane mirror; the calibration method for the parameters of the stereoscopic vision model of the single plane mirror comprises the following specific steps:
(1) monocular vision model
The monocular stereoscopic vision system composed of a plane mirror and a camera equivalently comprises two cameras: a real camera and a virtual camera; for real cameras, let the real camera coordinate system or-xryrzrThe lower spatial point is Pr(xr yr zr) Real camera coordinate system and world coordinate system ow-xwywzwThe transformation matrix between, i.e. the extrinsic parameter matrix of the real camera is [ R ]r Tr]；p(ur vr) Is a space point directly on an image surface coordinate system o without a reflectorr-urvrThe visual model of the real camera is expressed by equation (1) as:
wherein s isrIs a scale factor, alpharAnd betarAre respectively defined as ur、vrNormalized focal length in both axial directions, (u)0 v0) Coordinates of an intersection point of a real camera optical axis and an imaging plane;
for the virtual camera, let the virtual camera coordinate system ov-xvyvzvAnd the world coordinate system ow-xwywzwThe transformation matrix between, i.e. the extrinsic parameter matrix of the virtual camera is [ R ]v Tv](ii) a The visual model of the virtual camera is represented by equation (2):
wherein, p (u)v vv) Virtual image point, Pv(xv yv zv) Is Pr(xr yr zr) A point of symmetry about the plane mirror; svIs a scale factor, avAnd betavAre respectively defined as uv、vvNormalized focal length in two axis directions;
according to the mirror symmetry property, Pv(xv yv zv) And Pr(xr yr zr) The relationship between them is expressed by equation (3):
unfolding to obtain:
wherein (n)x ny nz) Is a unit normal vector of the plane mirror, satisfies nx·nx+ny·ny+nz·nz1 is ═ 1; d is the real camera coordinate system or-xryrzrOrigin orA vertical distance to a plane; equation (2) is then re-expressed as:
wherein,
and
respectively correspond to or-xryrzrAnd ov-xvyvzvA rotation matrix and a translation matrix in between;
(2) monocular stereoscopic vision three-dimensional measurement model
The formula (1) is arranged to obtain:
obtaining by arranging the formula (2):
as is known from formula (3), formula (4), formula (5), formula (6) and formula (7), p (u)r vr) And p (u)v vv) In betweenThe relationship is expressed by equation (8) as:
the coordinates of the space points in the real coordinate system are solved by the following formula:
the formula (9) is the established three-dimensional measurement model of monocular stereoscopic vision;
(3) visual model parameter calibration
1) Initial value calibration of internal parameter
Considering the influence of distortion, solving the parameter matrix K in the camera by using the Zhang scaling method in a small range of the image central regionrAnd KvObtaining the focal length f of the lens;
2) distortion coefficient calibration
Manufacturing and assembling defects of the lens may cause radial distortion and eccentric distortion, and the lens distortion is expressed by equation (10):
wherein, (u v) is the coordinate of image point; deltau、ΔvThe distortion produced for an image point is at the horizontal axis u of the imager(uv) And a vertical axis vr(vv) The component of (a);
is the distortion radius; k is a radical of1、k2First two order coefficients, p, of radial distortion, respectively1And p2The first two order coefficients of the eccentric distortion respectively; considering the effect of the depth of field on the lens imaging distortion, the radial distortion coefficient is expressed by equation (11):
wherein,
is a magnification factor, and
respectively, the focus distance within the depth of field is sn、sm、skThe ith order radial distortion coefficient on the focusing plane; for a real camera in monocular stereoscopic vision, a checkerboard standard object is used as a calibration plate, the calibration plate is kept perpendicular to the optical axis of the real camera, and the calibration plate is placed in the depth of field range along the optical axis so as to be respectively positioned at a focusing distance smAnd skOn two focusing planes; then, the radial distortion coefficients of each order on the two focusing planes are calculated respectively
And
i is 1, 2; finally, the focusing distance s is calculated by the formula (11)nIs the ith radial distortion coefficient on the focusing plane
After the virtual camera is processed by the same method, radial distortion coefficients of each order of the real camera and the virtual camera are obtained;
the off-center distortion coefficient considering the effect of the depth of field factor is expressed by equation (12):
wherein,
and
respectively, the focus distance within the depth of field is snAnd smThe ith order eccentric distortion coefficient on the two pairs of focal planes, and f is the focal length of the lens; for a real camera in monocular stereoscopic vision, a checkerboard standard object is used as a calibration plate, the calibration plate is kept perpendicular to the optical axis of the real camera, and the calibration plate is placed in the depth of field range along the optical axis so as to be in focus at a distance smOn the focusing plane of (a); then, the eccentric distortion coefficient of each order on the focusing plane is calculated
i is 1, 2; finally, the focusing distance s can be calculated by the formula (12)nIs the ith radial distortion coefficient on the focusing plane
After the virtual camera is processed by the same method, the eccentric distortion coefficients of each order of the real camera and the virtual camera are obtained;
3) internal and external parameter calibration
Respectively obtaining coefficients of each order of radial distortion and eccentric distortion in the visual models of the real camera and the virtual camera through the step 2); then, in order to avoid the influence of the solving error of the visual internal and external parameters on the distortion coefficient, the value of the distortion coefficient is respectively fixed for each camera, the distance provided by the checkered corner points is used as constraint, the LM algorithm is adopted to optimize and solve the internal parameter matrix and the external parameter matrix, and K is respectively obtainedr、Kv、[Rr Tr]And [ R ]v Tv](ii) a Finally, a conversion matrix [ R ' T ' between real camera and virtual camera is computed ']；
And further completing the calibration of the parameters of the stereoscopic vision model of the single plane mirror.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011112823.5A CN112258586B (en) | 2020-10-16 | 2020-10-16 | Calibration method for stereoscopic vision model parameters of single plane mirror |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011112823.5A CN112258586B (en) | 2020-10-16 | 2020-10-16 | Calibration method for stereoscopic vision model parameters of single plane mirror |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112258586A true CN112258586A (en) | 2021-01-22 |
| CN112258586B CN112258586B (en) | 2022-10-21 |
Family
ID=74243796
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011112823.5A Active CN112258586B (en) | 2020-10-16 | 2020-10-16 | Calibration method for stereoscopic vision model parameters of single plane mirror |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112258586B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113365043A (en) * | 2021-06-02 | 2021-09-07 | 北京理工大学 | Variable-scale panoramic imaging synchronous switching reflector device and panoramic expansion method |
| CN114509048A (en) * | 2022-01-20 | 2022-05-17 | 中科视捷(南京)科技有限公司 | Monocular camera-based overhead transmission line space three-dimensional information acquisition method and system |
| CN114705122A (en) * | 2022-04-13 | 2022-07-05 | 成都飞机工业（集团）有限责任公司 | Large-field stereoscopic vision calibration method |
| CN115641382A (en) * | 2022-10-21 | 2023-01-24 | 哈尔滨工业大学 | External parameter calibration method for orthogonal stereoscopic vision structure |
| CN116883517A (en) * | 2023-09-07 | 2023-10-13 | 华东交通大学 | Camera parameter calibration method without overlapping view fields based on plane mirror |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102980513A (en) * | 2012-11-02 | 2013-03-20 | 浙江工业大学 | Monocular panoramic three-dimensional vision sensor focusing on objects |
| US20170256042A1 (en) * | 2016-03-07 | 2017-09-07 | Lateral Reality Kft. | Methods and computer program products for calibrating stereo imaging systems by using a planar mirror |
| CN108416812A (en) * | 2018-03-20 | 2018-08-17 | 中国电子科技集团公司第五十四研究所 | A kind of scaling method of single camera mirror image binocular vision system |
| CN108830906A (en) * | 2018-05-31 | 2018-11-16 | 福州大学 | A kind of camera parameters automatic calibration method based on virtual Binocular Vision Principle |
| CN111578860A (en) * | 2020-04-22 | 2020-08-25 | 清华大学 | Three-dimensional shape measuring method based on reflector and monocular vision |
-
2020
- 2020-10-16 CN CN202011112823.5A patent/CN112258586B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102980513A (en) * | 2012-11-02 | 2013-03-20 | 浙江工业大学 | Monocular panoramic three-dimensional vision sensor focusing on objects |
| US20170256042A1 (en) * | 2016-03-07 | 2017-09-07 | Lateral Reality Kft. | Methods and computer program products for calibrating stereo imaging systems by using a planar mirror |
| CN108416812A (en) * | 2018-03-20 | 2018-08-17 | 中国电子科技集团公司第五十四研究所 | A kind of scaling method of single camera mirror image binocular vision system |
| CN108830906A (en) * | 2018-05-31 | 2018-11-16 | 福州大学 | A kind of camera parameters automatic calibration method based on virtual Binocular Vision Principle |
| CN111578860A (en) * | 2020-04-22 | 2020-08-25 | 清华大学 | Three-dimensional shape measuring method based on reflector and monocular vision |
Non-Patent Citations (4)
| Title |
|---|
| FUQIANG ZHOU等: "A novel way of understanding for calibrating stereo vision sensor constructed by a single camera and mirrors", 《MEASUREMENT》 * |
| S.BARONE等: "catadioptric stere-vision system using a spherocal mirror", 《SCIENCEDIRECT》 * |
| 郭伟青: "基于镜面成像技术的单目全景立体感知三维重构技术研究", 《中国博士学位论文全文数据库信息科技辑》 * |
| 郭伟青等: "基于镜面成像技术的三维立体视觉测量与重构综述", 《计算机科学》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113365043A (en) * | 2021-06-02 | 2021-09-07 | 北京理工大学 | Variable-scale panoramic imaging synchronous switching reflector device and panoramic expansion method |
| CN114509048A (en) * | 2022-01-20 | 2022-05-17 | 中科视捷(南京)科技有限公司 | Monocular camera-based overhead transmission line space three-dimensional information acquisition method and system |
| CN114509048B (en) * | 2022-01-20 | 2023-11-07 | 中科视捷(南京)科技有限公司 | Overhead transmission line space three-dimensional information acquisition method and system based on monocular camera |
| CN114705122A (en) * | 2022-04-13 | 2022-07-05 | 成都飞机工业（集团）有限责任公司 | Large-field stereoscopic vision calibration method |
| CN115641382A (en) * | 2022-10-21 | 2023-01-24 | 哈尔滨工业大学 | External parameter calibration method for orthogonal stereoscopic vision structure |
| CN115641382B (en) * | 2022-10-21 | 2023-09-08 | 哈尔滨工业大学 | External parameter calibration method for orthogonal stereoscopic vision structure |
| CN116883517A (en) * | 2023-09-07 | 2023-10-13 | 华东交通大学 | Camera parameter calibration method without overlapping view fields based on plane mirror |
| CN116883517B (en) * | 2023-09-07 | 2023-11-10 | 华东交通大学 | Camera parameter calibration method without overlapping view fields based on plane mirror |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112258586B (en) | 2022-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112258586B (en) | Calibration method for stereoscopic vision model parameters of single plane mirror | |
| CN105698699B (en) | A kind of Binocular vision photogrammetry method based on time rotating shaft constraint | |
| CN101581569B (en) | Calibrating method of structural parameters of binocular visual sensing system | |
| CN107367229B (en) | Free binocular stereo vision rotating shaft parameter calibration method | |
| WO2018076154A1 (en) | Spatial positioning calibration of fisheye camera-based panoramic video generating method | |
| CN108416812B (en) | Calibration method of single-camera mirror image binocular vision system | |
| CN105931222B (en) | The method for realizing high-precision camera calibration with low precision two dimensional surface target | |
| CN109579695B (en) | Part measuring method based on heterogeneous stereoscopic vision | |
| CN105043250B (en) | A kind of double-visual angle data alignment method based on 1 common indicium points | |
| CN101354796B (en) | Omnidirectional stereo vision three-dimensional rebuilding method based on Taylor series model | |
| CN109099838B (en) | Three-dimensional information measuring method of monocular stereoscopic vision system based on refraction and reflection | |
| CN110264528A (en) | Quick self-calibration method for fisheye lens binocular camera | |
| CN110874854B (en) | Camera binocular photogrammetry method based on small baseline condition | |
| CN114705122A (en) | Large-field stereoscopic vision calibration method | |
| CN113776785B (en) | Three-dimensional light path analysis method of monocular stereoscopic vision system | |
| CN114066983A (en) | Intelligent supplementary scanning method based on two-axis rotary table and computer readable storage medium | |
| CN107560554A (en) | A kind of three-dimensional information vision measuring method based on relay lens | |
| WO2022134939A1 (en) | Data splicing and system calibration method for human body digital measurement device | |
| CN107909543A (en) | A kind of flake binocular vision Stereo matching space-location method | |
| CN110728745B (en) | Underwater binocular stereoscopic vision three-dimensional reconstruction method based on multilayer refraction image model | |
| CN111429571B (en) | Rapid stereo matching method based on spatio-temporal image information joint correlation | |
| Li et al. | A virtual binocular line-structured light measurement method based on a plane mirror | |
| CN113763480A (en) | Multi-lens panoramic camera combination calibration method | |
| CN116804537A (en) | Binocular range finding system and method | |
| CN115713545A (en) | Bionic binocular vision tracking device and method driven by optical wedge set |
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 |