CN108364318A - A kind of planar dimension monocular measuring method for eliminating protective glass refractive effect - Google Patents
A kind of planar dimension monocular measuring method for eliminating protective glass refractive effect Download PDFInfo
- Publication number
- CN108364318A CN108364318A CN201810049946.5A CN201810049946A CN108364318A CN 108364318 A CN108364318 A CN 108364318A CN 201810049946 A CN201810049946 A CN 201810049946A CN 108364318 A CN108364318 A CN 108364318A
- Authority
- CN
- China
- Prior art keywords
- flat shape
- shape characteristic
- camera
- characteristic point
- planar dimension
- 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
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Data Mining & Analysis (AREA)
- Computational Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The present invention relates to a kind of planar dimension monocular measuring methods for eliminating protective glass refractive effect, the method is on the basis of analyzing glass refraction phenomenon, by the inner parameter and camera of calibration for cameras relative to the external parameter for measuring plane, calculate the glass refraction influence matrix and coordinate conversion matrix of each flat shape characteristic point, establish the non-linear camera imaging model based on pinhole imaging system, and then according to the non-linear camera imaging model foundation planar dimension detection model based on pinhole imaging system, the world coordinates for the measurement plane inner plane Feature Points for eliminating protective glass refractive effect is finally calculated using the planar dimension detection model.A kind of planar dimension monocular measuring method of elimination protective glass refractive effect of the present invention, calculating is simple, easy to implement, belongs to steel rolling automation field.
Description
Technical field
The invention belongs to steel rolling automation field, more particularly to a kind of planar dimension for eliminating protective glass refractive effect
Monocular measuring method.
Background technology
Vision measurement is the measurement and positioning that computer vision is applied to geometric dimension, using image as acquisition information
Means not only have the advantages that non-contact measurement, also have the uniquenesses such as at low cost, easy to operate, maneuverability, real-time
Advantage, thus it is widely used in industry spot.
Especially in industrial circles such as forging, rollings, need to carry out its size during the metal reprocessing after heating
It measures, to ensure product size quality, since these processing method conditions are extremely severe, leads to the hot geometric parameter of on-line measurement
It is very difficult.Vision measurement has the characteristics that non-contact, speed is fast, with high accuracy, therefore is widely used in these fields.It is forging
Field, document 1 (Bi Chao, high-temperature forging vision measurement technical research, University Of Tianjin's Master's thesis, 2012) surround high-temperature forging ruler
Very little vision measurement technology has carried out raising picture quality, the calibration of Pixel Dimensions equivalent, sub-pixel edge extracting and camera mark
Surely equal research work.View-based access control model measuring principle has built symmetrical binocular vision system, by establishing the mathematical model of system,
Corresponding three-dimensional world coordinate is acquired, and then obtains the three-dimensional dimension of forging.It is complete in order to measure intermediate base in rolling field
Long lateral bending information, document 2 (AVH Ollikkala, TP Kananen, AJM Holappa, Camera-based
Curvature Measurement of a Large Incandescent Object, Optical Measurement
Systems for Industrial Inspection VIII,2013,8788(1):87883D) use an area array CCD phase
Machine is tiltedly mounted on above Coil Box, is imaged to the intermediate base overall length after roughing by once shooting, and by assisting swashing
Light converts the intermediate base flat image in figure phase to actual size.But due to the severe working condition such as high temperature, dust, steam
Presence, be safeguard industries camera, casing protection device used to industrial camera, while in order to ensure the bat of industrial CCD camera
Condition is taken the photograph, protective glass can be installed in camera front end, to achieve the purpose that protect camera and shooting.
During being measured to article geometrical size using machine vision technique, use camera calibration technology can be with
The inside and outside parameter of camera system is obtained, and then the conversion pass of image coordinate and actual coordinate can be established by national forest park in Xiaokeng
System, completes the measurement of geometric dimension.But due to the presence of protective glass, the refraction of light can be caused, in this case, then
The transformational relation that coordinate is characterized using simple national forest park in Xiaokeng will produce nonlinearity erron, increase the inaccuracy subsequently measured
Property, bring error larger or even wrong measurement result for production.
Invention content
To solve the above problems, the present invention provides a kind of planar dimension monocular side eliminating protective glass refractive effect
Method.The present invention is achieved by the following technical solutions:
A kind of planar dimension monocular measuring method for eliminating protective glass refractive effect, the method is according to no glass conditions
Under the external parameter of camera internal parameter obtained by calibrating and camera relative measurement plane, that whether there is or not each flat shapes when glass is special
It is special that each flat shape is calculated in the glass refraction influence matrix of the world coordinates spacing, each flat shape characteristic point of levying point
Levy the coordinate conversion matrix of point;
The non-linear camera based on pinhole imaging system is established according to the coordinate conversion matrix of each flat shape characteristic point
Imaging model;
And then according to the non-linear camera imaging model foundation planar dimension detection model based on pinhole imaging system, finally
The measurement plane inner plane shape feature for eliminating protective glass refractive effect is calculated using the planar dimension detection model
The world coordinates of point.
Further, it the described method comprises the following steps:
Step S1 demarcates camera, before installing protective glass to obtain the inner parameter M of camerainIt is opposite with camera
In the external parameter M for measuring planeex;
Step S1 calibration process the specific steps are:
Step S1.1, image obtain:It obtains the image of multiple gridiron pattern scaling board different positions and poses and measures in plane
Gridiron pattern scaling board image;
Step S1.2 calculates the subpixel coordinates of angle point:The chessboard case marker is extracted using Harris Corner Detection Algorithms
Angle point in the image of fixed board different positions and pose and the gridiron pattern scaling board image in measurement plane, and use shade of gray feature
Method carries out sub-pix calculating, obtains the subpixel coordinates (u of accurate angle pointM,vN):
Harris Corner Detection Algorithms judge whether a point is angle point by Harris receptance functions value, described
Harris receptance function formula are:
R(uM,vN)=Det (M)-k*Trace2(M) (1)
Wherein, R (uM,vN) it is receptance function value;Det (M) is the determinant of matrix M, and M is a two-dimensional matrix, can table
It is shown asIuThe single order shade of gray of U axis directions, I are fastened for image pixel coordinatesvFor image slices
The single order shade of gray of V axis directions on plain coordinate system;Trace (M) is the mark of matrix M;K is an empirical, and value is
0.05;
It using formula (2), is calculated using shade of gray characteristic method, by successive ignition, is finally obtained more accurate
Angular-point sub-pixel coordinate (uM,vN), the formula (2) is:
Wherein,For the shade of gray value for calculating at point;For matrixTransposed matrix;For image original
Vectors of the point O to angle point;For image origin O to the vector for calculating point;
Step S1.3 calculates the inner parameter M of camerainWith camera relative to the external parameter M for measuring planeex:According to step
Angular-point sub-pixel coordinate obtained by rapid S1.2, the inner parameter M of camera is calculated using Zhang Zhengyou camera calibration methodsinAnd camera
Relative to the external parameter M for measuring planeex:
Camera internal parameter Min:
Wherein, OcPoint is lens optical center, crosses OcThe image plane normal of point is O, O with image plane intersection pointcO is focal length f,
Unit mm;Of- UV is image pixel coordinates system, it is using the point in the pixel planes upper left corner as coordinate origin Of, the seat of each pixel
Mark (u, v) is columns and line number of the pixel in array respectively, and O point image pixel coordinates are (u0,v0);O-XY is image object
Coordinate system is managed, it is using O as origin, X, and Y-axis is parallel to image pixel coordinates system U, V axis, unit mm;Dx, dy are each pixel
Physical size in X-axis and Y direction, unit mm/ pixels;Oc-XcYcZcFor camera coordinates system, it is with OcFor origin, Xc,Yc
Axis is respectively parallel to corresponding image physical coordinates system X, and the optic axis of Y-axis, camera is ZcAxis, unit mm;
To measure the angle point O in the gridiron pattern scaling board upper left corner in planewFor origin, world coordinate system O is establishedw-XwYwZw,
Xw,YwAxis is respectively parallel to the scaling board X-comers transverse direction line and longitudinal line, ZwIt is upward perpendicular to calibration plane,
Then camera is relative to the external parameter M for measuring planeex:
Wherein, spin matrixTranslation vector T=[tx ty tz]T, spin matrix R interior elements r1,r2,
r3,r4,r5,r6,r7,r8,r9Dimensionless, translation vector T interior elements tx,ty,tzUnit mm;
Step S2, after installing protective glass, when tested flat shape is in measured zone, image shot by camera uses
Corner Detection Algorithm or contour detecting algorithm extract flat shape characteristic point, record the pixel coordinate of the flat shape characteristic point
(u, v) data;
Of- UV is image pixel coordinates system, it is using the point in the pixel planes upper left corner as coordinate origin Of, each pixel
Coordinate (u, v) is columns and line number of the pixel in array respectively;
Step S3, for pixel coordinate (u, v) data of the step S2 flat shape characteristic point obtained, calculating has,
World coordinates separation delta without the flat shape characteristic point under the influence of glass2, Δ2It calculates as shown in following formula (5):
Wherein, d is thickness of glass, unit mm;α is the incidence angle of the corresponding flat shape characteristic point, thenF is the focal length of camera lens, unit mm;Dx, dy are each pixel in X-axis
With the physical size in Y direction, unit mm/ pixels;N is glass refraction, dimensionless.
Step S4, according to pixel coordinate (u, the v) data and step S3 of the step S2 flat shape characteristic points obtained
The world coordinates separation delta with and without flat shape characteristic point under the influence of glass being calculated2, glass is calculated using formula (6)
Refractive effect matrix
Wherein, flat shape characteristic point in image physical coordinates system with the angle of X-axis(u0,v0)
For image physical coordinates system origin pixel coordinate, (u, v) is the pixel coordinate of flat shape characteristic point;u≥u0When, take+Δ2cos
γ;U < u0When, take-Δ2cosγ;v≥v0When, take+Δ2sinγ;V < v0When, take-Δ2sinγ;
Step S5, according to the camera internal parameter Min, the camera relative measurement plane external parameter MexWith it is described
The glass refraction influence matrix of each flat shape characteristic pointThe coordinate for each flat shape characteristic point being calculated turns
Change matrixEstablish the non-linear camera imaging model based on pinhole imaging system;
The non-linear camera imaging model based on pinhole imaging system is the planar shaped measured in world coordinate system in plane
World coordinates (the x of shape characteristic pointw,yw,zw) pixel coordinate (u, v) to form corresponding flat shape characteristic point is projected with it
Transformational relation, as shown in following formula (7):
Wherein, world coordinate system Ow-XwYwZwTo measure the angle point O in the gridiron pattern scaling board upper left corner in planewFor origin,
Xw,YwAxis is respectively parallel to the scaling board X-comers transverse direction line and longitudinal line, ZwIt is upward perpendicular to calibration plane;For the coordinate conversion matrix of each flat shape characteristic point;zcFor zoom factor, unit mm;
The coordinate conversion matrix of each flat shape characteristic pointCalculating such as following formula (8) shown in:
Wherein, transition matrixElementDimensionless;ElementUnit mm;
Step S6, z in the non-linear camera imaging model based on pinhole imaging systemwWhen=0, planar dimension detection is established
Model, the planar dimension detection model calculate the corresponding world coordinate system of the flat shape characteristic point pixel coordinate (u, v)
World coordinates (the x of flat shape characteristic point in middle measurement planew,yw), as shown in following formula (9):
Wherein,
For the element in the coordinate conversion matrix of each flat shape characteristic point;Element Dimensionless;ElementUnit mm.
The beneficial effects of the present invention are:On the basis of obtaining the uncalibrated image of unshielded glass, build through the invention
The planar dimension detection model of the vertical non-linear camera imaging model and proposition based on pinhole imaging system.The present invention is flat by calculating
The transition matrix of the pixel coordinate of face Feature Points world coordinates corresponding with image pixel coordinates, acquires characteristic point and is eliminating
World coordinates under the influence of light refraction achievees the purpose that eliminating protective glass has an impact planar dimension vision measurement, is
Subsequent measurement provides accurate planar dimension data.
Description of the drawings
Fig. 1 is the schematic diagram of X-comers in the embodiment of the present invention.
Fig. 2 is the accurate schematic diagram of sub-pix based on shade of gray characteristic method in the embodiment of the present invention.
Fig. 3 is world coordinate system, camera coordinates system, image pixel coordinates system and image physical coordinates in the embodiment of the present invention
The coordinate relation schematic diagram of system.
Fig. 4 is the schematic diagram influenced on imaging relations after installation glass in the embodiment of the present invention.
Fig. 5 is the flow chart for the planar dimension monocular measuring method that the embodiment of the present invention eliminates protective glass refractive effect.
Specific implementation mode
To keep the technical problem to be solved in the present invention, technical solution and advantage clearer, below in conjunction with attached drawing and tool
Body embodiment is described in detail.It should be appreciated that specific embodiment described herein is used only for explaining the present invention, and do not have to
It is of the invention in limiting.
On the contrary, the present invention covers any replacement done in the spirit and scope of the present invention being defined by the claims, repaiies
Change, equivalent method and scheme.Further, in order to make the public have a better understanding the present invention, below to the thin of the present invention
It is detailed to describe some specific detail sections in section description.Part without these details for a person skilled in the art
Description can also understand the present invention completely.
Embodiment 1
The present embodiment provides a kind of planar dimension monocular measuring methods for eliminating protective glass refractive effect, such as Fig. 1-Fig. 5
It is shown, it the described method comprises the following steps:
Step S1 demarcates camera, before installing protective glass to obtain the inner parameter M of camerainIt is opposite with camera
In the external parameter M for measuring planeex;
The specific of step S1 calibration process be step by step:
S1.1, image obtain:It obtains the image of multiple gridiron pattern scaling board different positions and poses and demarcates the gridiron pattern of plane
Scaling board image;
In the present embodiment, it obtains the image of 20 gridiron pattern scaling board different positions and poses and 1 is located at and measures in plane
Gridiron pattern scaling board image;
S1.2 calculates the subpixel coordinates of angle point:The gridiron pattern scaling board is extracted using Harris Corner Detection Algorithms
The image of different positions and pose and measure angle point in gridiron pattern scaling board image in plane, and using shade of gray characteristic method into
Row sub-pix calculates, and obtains the subpixel coordinates (u of accurate angle pointM,vN):
As shown in Figure 1, angle point judges to use formula (1) as Harris receptance functions:
R(uM,vN)=Det (M)-k*Trace2(M) (1)
Wherein, R (uM,vN) it is receptance function value;Det (M) is the determinant of matrix M, and M is a two-dimensional matrix, can table
It is shown asIuThe single order shade of gray of U axis directions, I are fastened for image pixel coordinatesvFor image slices
The single order shade of gray of V axis directions on plain coordinate system;Trace (M) is the mark of matrix M;K is an empirical, and value is
0.05;
In the present embodiment, wherein the coordinate of angle point is on a gridiron pattern scaling board:
(2007,820) | (2064,996) | (2122,1174) | (2180,1355) |
(2239,1537) | (1830,866) | (1886,1044) | (1943,1223) |
(2002,1405) | (2060,1588) | (1650,913) | (1706,1092) |
(1762,1273) | (1820,1456) | (1878,1642) | (1467,961) |
(1522,1141) | (1578,1323) | (1635,1509) | (1692,1696) |
(1283,1008) | (1337,1190) | (1392,1374) | (1448,1561) |
(1505,1750) | (1094,1058) | (1148,1241) | (1202,1426) |
(1258,1615) | (1314,1805) | (904,1107) | (957,1291) |
(1010,1479) | (1065,1668) | (1120,1861) |
As shown in Fig. 2, the sub-pix exact algorithm based on shade of gray characteristic method is iterated using formula (2), finally
Obtain more accurate angular-point sub-pixel coordinate (uM,vN):
Wherein,For the shade of gray value for calculating at point;For matrixTransposed matrix;For image original
Vectors of the point O to angle point;For image origin O to the vector for calculating point;
In the present embodiment, wherein the subpixel coordinate of angle point is on a gridiron pattern scaling board:
S1.3 calculates the inner parameter M of camerainWith camera relative to the external parameter M for measuring planeex:According to step
Angular-point sub-pixel coordinate obtained by S1.2 calculates the inner parameter M of camera using Zhang Zhengyou camera calibration methodsinWith camera relative to
Measure the external parameter M of planeex;
Camera internal parameter Min:
Wherein, OcPoint is lens optical center, crosses OcThe image plane normal of point is O, O with image plane intersection pointcO is focal length f,
Unit mm;Of- UV is image pixel coordinates system, it is using the point in the pixel planes upper left corner as coordinate origin Of, the seat of each pixel
Mark (u, v) is columns and line number of the pixel in array respectively, and O point image pixel coordinates are (u0,v0);O-XY is image object
Coordinate system is managed, it is using O as origin, X, and Y-axis is parallel to image pixel coordinates system U, V axis, unit mm;Dx, dy are each pixel
Physical size in X-axis and Y direction, unit mm/ pixels;Oc-XcYcZcFor camera coordinates system, it is with OcFor origin, Xc,Yc
Axis is parallel to corresponding image physical coordinates system X, and the optic axis of Y-axis, camera is ZcAxis, unit mm;
To measure the angle point O in the gridiron pattern scaling board upper left corner in planewFor origin, world coordinate system O is establishedw-XwYwZw,
Xw,YwAxis is respectively parallel to the scaling board X-comers transverse direction line and longitudinal line, ZwIt is upward perpendicular to calibration plane,
Then camera is relative to the external parameter M for measuring planeex:
Wherein, spin matrixTranslation vector T=[tx ty tz]T, spin matrix R interior elements r1,r2,
r3,r4,r5,r6,r7,r8,r9Dimensionless, translation vector T interior elements tx,ty,tzUnit mm;
In the present embodiment, as shown in figure 3, by Zhang Zhengyou standardizations, can be in the hope of the inner parameter of cameraCamera relative to measure plane external parameter be
Step S2, after installing protective glass, when tested flat shape is in measured zone, image shot by camera uses
Corner Detection Algorithm or contour detecting algorithm extract flat shape characteristic point, are recorded in the spy of flat shape described in pixel coordinate system
Levy pixel coordinate (u, v) data of point;
In the present embodiment, 35 flat shape characteristic points are extracted from captured image, pixel coordinate is:
Step S3, for pixel coordinate (u, v) data of the step S2 flat shape characteristic points obtained, using formula
(5) the world coordinates separation delta with and without flat shape characteristic point under the influence of glass is calculated2, Δ2It calculates as shown in following formula (5):
Wherein, d is thickness of glass, unit mm;α is the incidence angle of the corresponding flat shape characteristic point, thenF is the focal length of camera lens, unit mm;Dx, dy are each pixel in X-axis
With the physical size in Y direction, unit mm/ pixels;N is glass refraction, dimensionless.
In the present embodiment, thickness of glass d=4mm, glass refraction n=1.61, the corresponding incidence of flat shape characteristic point
Angle α is:
Number | α | Number | α |
1 | 5.2408 | 19 | 2.2846 |
2 | 4.0201 | 20 | 3.4665 |
3 | 2.9105 | 21 | 5.0539 |
4 | 2.2709 | 22 | 6.4099 |
5 | 2.6332 | 23 | 5.4080 |
6 | 3.8187 | 24 | 4.5954 |
7 | 5.3755 | 25 | 4.1584 |
8 | 4.9349 | 26 | 4.3026 |
9 | 3.5230 | 27 | 5.0619 |
10 | 2.0394 | 28 | 6.2858 |
11 | 0.4916 | 29 | 7.8037 |
12 | 1.1707 | 30 | 7.0557 |
13 | 2.8961 | 31 | 6.5066 |
14 | 4.7097 | 32 | 6.2628 |
15 | 5.3640 | 33 | 6.4172 |
16 | 4.0951 | 34 | 7.0050 |
17 | 2.8936 | 35 | 7.9824 |
18 | 2.0670 |
Thus, the corresponding world coordinates separation delta with and without under the influence of glass of flat shape characteristic point2For:
Number | Δ2 | Number | Δ2 |
1 | 0.1396 | 19 | 0.0605 |
2 | 0.1068 | 20 | 0.0920 |
3 | 0.0772 | 21 | 0.1346 |
4 | 0.0601 | 22 | 0.1713 |
5 | 0.0698 | 23 | 0.1441 |
6 | 0.1014 | 24 | 0.1222 |
7 | 0.1432 | 25 | 0.1105 |
8 | 0.1313 | 26 | 0.1143 |
9 | 0.0935 | 27 | 0.1348 |
10 | 0.0540 | 28 | 0.1679 |
11 | 0.0130 | 29 | 0.2096 |
12 | 0.0310 | 30 | 0.1890 |
13 | 0.0768 | 31 | 0.1740 |
14 | 0.1253 | 32 | 0.1673 |
15 | 0.1429 | 33 | 0.1715 |
16 | 0.1088 | 34 | 0.1876 |
17 | 0.0767 | 35 | 0.2146 |
18 | 0.0547 |
Step S4, according to pixel coordinate (u, the v) data and step S3 of the step S2 flat shape characteristic points obtained
The world coordinates separation delta with and without flat shape characteristic point under the influence of glass calculated2, glass refraction is calculated using formula (6)
Influence matrix
Wherein, flat shape characteristic point in image physical coordinates system with the angle of X-axis(u0,v0)
For image physical coordinates system origin pixel coordinate, (u, v) is the pixel coordinate of flat shape characteristic point;u≥u0When, take+Δ2cos
γ;U < u0When, take-Δ2cosγ;v≥v0When, take+Δ2sinγ;V < v0When, take-Δ2sinγ;
In the present embodiment, the corresponding γ of flat shape characteristic point is:
Step S5, according to the camera internal parameter Min, the camera relative measurement plane external parameter MexWith it is described
The glass refraction influence matrix of each flat shape characteristic pointThe coordinate for each flat shape characteristic point being calculated turns
Change matrixEstablish the non-linear camera imaging model based on pinhole imaging system;
The non-linear camera imaging model based on pinhole imaging system is the planar shaped measured in world coordinate system in plane
World coordinates (the x of shape characteristic pointw,yw,zw) pixel coordinate (u, v) to form corresponding flat shape characteristic point is projected with it
Transformational relation, as shown in following formula (7):
Wherein, world coordinate system Ow-XwYwZwTo measure the angle point O in the gridiron pattern scaling board upper left corner in planewFor origin,
Xw,YwAxis is respectively parallel to the scaling board X-comers transverse direction line and longitudinal line, ZwIt is upward perpendicular to calibration plane;For the coordinate conversion matrix of each flat shape characteristic point;zcFor zoom factor, unit mm;
The coordinate conversion matrix of each flat shape characteristic pointCalculating such as following formula (8) shown in:
Wherein, transition matrixElementDimensionless;ElementUnit mm;
Step S6:Z in the non-linear camera imaging model based on pinhole imaging systemwWhen=0, planar dimension detection is established
Model, the planar dimension detection model calculate the corresponding world coordinate system of the flat shape characteristic point pixel coordinate (u, v)
World coordinates (the x of flat shape characteristic point in middle measurement planew,yw), as shown in following formula (9):
Wherein,
For the element in the coordinate conversion matrix of each flat shape characteristic point;Element Dimensionless;ElementUnit mm.
In the present embodiment, the flat shape characteristic point being calculated according to the planar dimension detection model is in world coordinates
In system measure plane on world coordinates be:
The present embodiment on the basis of obtaining the uncalibrated image of unshielded glass, establish through the invention based on aperture at
The non-linear camera imaging model of picture and the planar dimension detection model of proposition.The present invention passes through Calculation Plane Feature Points
The transition matrix of pixel coordinate world coordinates corresponding with image pixel coordinates acquires characteristic point under the influence of eliminating light refraction
World coordinates, achieve the purpose that eliminate protective glass planar dimension vision measurement is had an impact, for it is subsequent measure provide
Accurate planar dimension data.In the present embodiment, the characteristic point world coordinates that the technical method provided through the invention acquires
Mean error be 0.02613mm improve accuracy of detection compared with the mean error 0.15027mm of conventional method.
Claims (8)
1. it is a kind of eliminate protective glass refractive effect planar dimension monocular measuring method, which is characterized in that the method according to
Without the external parameter of camera internal parameter obtained by calibrating and camera relative measurement plane under glass conditions, whether there is or not each when glass
The world coordinates spacing of flat shape characteristic point, each flat shape characteristic point glass refraction influence matrix be calculated it is each
The coordinate conversion matrix of flat shape characteristic point;
The non-linear camera imaging based on pinhole imaging system is established according to the coordinate conversion matrix of each flat shape characteristic point
Model;
It is final to utilize and then according to the non-linear camera imaging model foundation planar dimension detection model based on pinhole imaging system
The measurement plane inner plane Feature Points for eliminating protective glass refractive effect are calculated in the planar dimension detection model
World coordinates.
2. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 1, special
Sign is, the described method comprises the following steps:
Step S1 demarcates camera, before installing protective glass to obtain the inner parameter M of camerainWith camera relative to survey
Measure the external parameter M of planeex;
Step S2, after installing protective glass, when tested flat shape is in measured zone, image shot by camera uses angle point
Detection algorithm or contour detecting algorithm extract flat shape characteristic point, record the flat shape characteristic point pixel coordinate (u,
V) data;
Of- UV is image pixel coordinates system, it is using the point in the pixel planes upper left corner as coordinate origin Of, the coordinate of each pixel
(u, v) is columns and line number of the pixel in array respectively;
Step S3 is calculated for pixel coordinate (u, v) data of the step S2 flat shape characteristic points obtained with and without glass
The world coordinates separation delta of flat shape characteristic point under the influence of glass2;
Step S4 is calculated according to pixel coordinate (u, the v) data of the step S2 flat shape characteristic points obtained and step S3
The obtained world coordinates separation delta with and without flat shape characteristic point under the influence of glass2, calculate glass refraction influence matrix
Step S5, according to the camera internal parameter M of calibrationin, camera relative measurement plane external parameter MexWith each planar shaped
The glass refraction influence matrix of shape characteristic pointThe coordinate conversion matrix for each flat shape characteristic point being calculatedEstablish the non-linear camera imaging model based on pinhole imaging system;
Step S6, z in the non-linear camera imaging model based on pinhole imaging systemwWhen=0, planar dimension detection model is established,
The planar dimension detection model calculates to be measured in the corresponding world coordinate system of the flat shape characteristic point pixel coordinate (u, v)
World coordinates (the x of flat shape characteristic point in planew,yw)。
3. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 2, feature
It is, the particular content of the step S1 is:
Step S1.1, image obtain:It obtains the image of multiple gridiron pattern scaling board different positions and poses and measures the chessboard in plane
Case marker fixed board image;
Step S1.2 calculates the subpixel coordinates of angle point:The gridiron pattern scaling board is extracted using Harris Corner Detection Algorithms
The image of different positions and pose and measure angle point in gridiron pattern scaling board image in plane, and using shade of gray characteristic method into
Row sub-pix calculates, and obtains the subpixel coordinates (u of accurate angle pointM,vN);
Step S1.3 calculates the inner parameter M of camerainWith camera relative to the external parameter M for measuring planeex:According to step
The inner parameter M of camera is calculated using Zhang Zhengyou camera calibration methods for angular-point sub-pixel coordinate obtained by S1.2inWith camera phase
For measuring the external parameter M of planeex。
4. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 3, feature
It is, the world coordinates separation delta described in step S3 with and without flat shape characteristic point under the influence of glass2Calculating such as following formula
(5) shown in:
Wherein, d is thickness of glass, unit mm;α is the incidence angle of corresponding flat shape characteristic point, thenF is the focal length of camera lens, unit mm;Dx, dy are each pixel in X-axis
With the physical size in Y direction, unit mm/ pixels;N is glass refraction, dimensionless.
5. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 4, feature
It is, glass refraction influence matrix described in step S4Calculating formula such as formula (6) shown in:
Wherein, flat shape characteristic point in image physical coordinates system with the angle of X-axis(u0,v0) it is figure
As physical coordinates system origin pixel coordinate, (u, v) is the pixel coordinate of flat shape characteristic point;u≥u0When, take+Δ2cosγ;u
< u0When, take-Δ2cosγ;v≥v0When, take+Δ2sinγ;V < v0When, take-Δ2sinγ。
6. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 5, feature
It is, the non-linear camera imaging model based on pinhole imaging system described in step S5 is that putting down in plane is measured in world coordinate system
World coordinates (the x of face Feature Pointsw,yw,zw) projected with it to be formed corresponding flat shape characteristic point pixel coordinate (u,
V) transformational relation, as shown in following formula (7):
Wherein,For the coordinate conversion matrix of each flat shape characteristic point;zcFor zoom factor, unit mm;Ow-XwYwZwWith
Measure the angle point O in the gridiron pattern scaling board upper left corner in planewFor origin, Xw,YwAxis is respectively parallel to scaling board gridiron pattern angle
The lateral line of point and longitudinal line, ZwIt is upward perpendicular to calibration plane.
7. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 6, feature
It is, the coordinate conversion matrix of each flat shape characteristic pointCalculating such as following formula (8) shown in:
Wherein, transition matrixElementDimensionless;ElementUnit mm.
8. a kind of planar dimension monocular measuring method for eliminating protective glass refractive effect according to claim 7, feature
It is, shown in planar dimension detection model such as following formula described in step S6 (9):
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810049946.5A CN108364318B (en) | 2018-01-18 | 2018-01-18 | Plane size monocular measurement method for eliminating refraction influence of protective glass |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810049946.5A CN108364318B (en) | 2018-01-18 | 2018-01-18 | Plane size monocular measurement method for eliminating refraction influence of protective glass |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108364318A true CN108364318A (en) | 2018-08-03 |
CN108364318B CN108364318B (en) | 2020-06-26 |
Family
ID=63006723
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810049946.5A Active CN108364318B (en) | 2018-01-18 | 2018-01-18 | Plane size monocular measurement method for eliminating refraction influence of protective glass |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108364318B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111161339A (en) * | 2019-11-18 | 2020-05-15 | 珠海随变科技有限公司 | Distance measuring method, device, equipment and computer readable medium |
CN113345027A (en) * | 2021-05-24 | 2021-09-03 | 武汉港迪智能技术有限公司 | Gantry crane area protection method based on dome camera |
CN114913243A (en) * | 2022-05-30 | 2022-08-16 | 歌尔股份有限公司 | Distortion detection method and device for optical material and medium |
CN116734727A (en) * | 2022-10-08 | 2023-09-12 | 荣耀终端有限公司 | Positioning method and device |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103364171A (en) * | 2013-07-15 | 2013-10-23 | 中国航天空气动力技术研究院 | Video measuring system and measuring method for model gestures in high-speed wind tunnel |
-
2018
- 2018-01-18 CN CN201810049946.5A patent/CN108364318B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103364171A (en) * | 2013-07-15 | 2013-10-23 | 中国航天空气动力技术研究院 | Video measuring system and measuring method for model gestures in high-speed wind tunnel |
Non-Patent Citations (3)
Title |
---|
ZHENG GONG 等: "Flexible method of refraction correction in vision measurement systems with multiple glass ports", 《OPTICS EXPRESS》 * |
刘玉宝: "线结构光三维测量关键技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
王俊 等: "多介质下空间目标的视觉测量", 《计算机应用》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111161339A (en) * | 2019-11-18 | 2020-05-15 | 珠海随变科技有限公司 | Distance measuring method, device, equipment and computer readable medium |
CN111161339B (en) * | 2019-11-18 | 2020-11-27 | 珠海随变科技有限公司 | Distance measuring method, device, equipment and computer readable medium |
CN113345027A (en) * | 2021-05-24 | 2021-09-03 | 武汉港迪智能技术有限公司 | Gantry crane area protection method based on dome camera |
CN113345027B (en) * | 2021-05-24 | 2022-05-06 | 武汉港迪智能技术有限公司 | Gantry crane area protection method based on dome camera |
CN114913243A (en) * | 2022-05-30 | 2022-08-16 | 歌尔股份有限公司 | Distortion detection method and device for optical material and medium |
CN116734727A (en) * | 2022-10-08 | 2023-09-12 | 荣耀终端有限公司 | Positioning method and device |
Also Published As
Publication number | Publication date |
---|---|
CN108364318B (en) | 2020-06-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103471531B (en) | The online non-contact measurement method of axial workpiece linearity | |
CN105894499B (en) | A kind of space object three-dimensional information rapid detection method based on binocular vision | |
CN103714535B (en) | Binocular vision measurement system camera parameter online adjustment method | |
CN108662987B (en) | Calibration method of 2D camera type laser measuring head | |
CN110189314B (en) | Automobile instrument panel image positioning method based on machine vision | |
Liu et al. | An improved online dimensional measurement method of large hot cylindrical forging | |
CN109579695B (en) | Part measuring method based on heterogeneous stereoscopic vision | |
CN108364318A (en) | A kind of planar dimension monocular measuring method for eliminating protective glass refractive effect | |
CN106500625B (en) | A kind of telecentricity stereo vision measurement method | |
CN105327950A (en) | Hot rolling intermediate billet camber detecting device and detecting method considering thickness change | |
CN109974618A (en) | The overall calibration method of multisensor vision measurement system | |
Liu et al. | Measuring method for micro-diameter based on structured-light vision technology | |
CN103697811A (en) | Method of obtaining three-dimensional coordinates of profile of object through combining camera and structural light source | |
CN109506629B (en) | Method for calibrating rotation center of underwater nuclear fuel assembly detection device | |
Wang et al. | Corners positioning for binocular ultra-wide angle long-wave infrared camera calibration | |
CN111998834B (en) | Crack monitoring method and system | |
CN106447729A (en) | 2 dimensional digital image related compensation method based on transformation of coordinates and 2 dimensional optical digital image related extensometer | |
CN117314986A (en) | Unmanned aerial vehicle cross-mode power distribution equipment inspection image registration method based on semantic segmentation | |
JPH04181106A (en) | Calibration device of position dimension measuring device | |
CN116091488B (en) | Displacement testing method and displacement testing system for engine swing test | |
CN112489141B (en) | Production line calibration method and device for single-board single-image strip relay lens of vehicle-mounted camera | |
CN114963981A (en) | Monocular vision-based cylindrical part butt joint non-contact measurement method | |
Fan et al. | Dimensional inspecting system of shaft parts based on machine vision | |
Danyu et al. | A simple calibration method for line-structured light vision sensor based on planar target of different positions | |
Hurník et al. | Multi-view camera system for measurement of heavy forgings |
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 |