CN109191527A - A kind of alignment method and device based on minimum range deviation - Google Patents
A kind of alignment method and device based on minimum range deviation Download PDFInfo
- Publication number
- CN109191527A CN109191527A CN201811358794.3A CN201811358794A CN109191527A CN 109191527 A CN109191527 A CN 109191527A CN 201811358794 A CN201811358794 A CN 201811358794A CN 109191527 A CN109191527 A CN 109191527A
- Authority
- CN
- China
- Prior art keywords
- array
- coordinate
- subject table
- product
- range deviation
- 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
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
- G06T7/74—Determining position or orientation of objects or cameras using feature-based methods involving reference images or patches
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
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)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the present application shows a kind of alignment method and device based on minimum range deviation, firstly, the alignment method for minimizing range deviation is directly started with using aligning product specification as target, the contraposition result of calculating more meets the evaluation criteria of production；Secondly, the alignment method shown in the embodiment of the present application, can efficiently solve non-contraposition problem placed in the middle, the contraposition scene of irregular shape product suitable for actual production.
Description
Technical field
The present invention relates to technical field of machine vision, in particular to it is a kind of based on minimize range deviation alignment method and
Device.
Background technique
In industrial circles such as electronic semi-conductor, touch screen, solar energy, automobile and parts, generally require two or more
A components carry out contraposition installation.The precision of contraposition installation directly determines the quality of product, in the modern industrial production process, one
As the contrapositions of components is realized using machine vision alignment system.
As shown in Figure 1, for a kind of current structural schematic diagram of common machine vision alignment system, the machine vision pair
Position system includes target platform 110, subject table 120, target image acquisition device 130 and object images acquisition device 140；Mesh
Mark object 150 is placed on target platform 110, and real-time objects 160 are placed in the subject table 120, the target object
150 and the real-time objects 160 be provided with contraposition mark 170, in actual industrial processes, for example, it is described in real time it is right
As 160 can be mobile phone liquid crystal screen, the target object 150 can be backlight module, pass through the machine vision alignment system
Realize the fitting of the two；The target image acquisition device 130 and the object images acquisition device 140 are described right for absorbing
Bit identification 170.The process of contraposition fitting includes: the contraposition of 140 Image Acquisition of object images acquisition device intake real-time objects 160
Mark 170, and according to the mapping relations of predetermined object images plane and subject table plane, real-time objects are calculated
The subject table coordinate of contraposition mark 170 in 160；Equally, target image acquisition device 130 absorbs the contraposition of target object 150
Mark 170, and according to the mapping relations of predetermined target image plane and target platform plane, target object is calculated
The target platform coordinate of contraposition mark 170 in 150；Real-time objects 160 are sent into the target platform 110 by subject table 120
Surface, and according to the subject table coordinate and target platform coordinate of corresponding contraposition mark 170, regulating object platform 120
Position, the real-time objects 160 also carry out position adjustment under the drive of the subject table 120, to complete real-time objects
160 are bonded with target object 150.During above-mentioned contraposition fitting, predefines object images plane and subject table is flat
Face and target image plane and the basis that the mapping relations of target platform plane are that contraposition is bonded, are referred to as demarcated；The calibration
Levels of precision be the key that contraposition fitting accuracy.
Currently, the scaling method of common machine vision alignment system is mainly nine-point circle shape target scaling method, including
Following steps: the target disc that 9 positions determine is placed in the target platform 110, is acquired and is filled by the object images
Set the object images coordinate in the 140 acquisitions target disc center of circle；It is mobile that subject table 120 is manually operated, keeps the object flat
The datum mark of platform 120 successively passes through the center of circle of 9 target discs, after observation determines that the datum mark is overlapped with the center of circle, note
Record the subject table coordinate of the datum mark；Pass through the object images coordinate difference in the adjacent center of circle and the subject table of datum mark
Coordinate difference calculates the mapping relations for determining object images plane and subject table plane, completes calibration.However, as product is more
Sample, contraposition scene is also to become increasingly complex.It is different from the workpiece that the product that tradition aligns between two parties is regular shape, when to shape
Shape or the irregular product of material are aligned, and require contraposition after specific range on the basis of apart from specification when, traditional residence
Middle alignment mode will be not suitable for such scene.
Summary of the invention
Goal of the invention of the invention is to provide a kind of alignment method and device based on minimum range deviation, to solve
Scaling method low precision in the prior art and problem poor for applicability.
The embodiment of the present application first aspect shows a kind of based on the alignment method for minimizing range deviation, the method packet
It includes:
Object images acquisition device obtains benchmark image on target platform, establishes coordinate system on the benchmark image, obtains
The coordinate of multiple characteristic quantities of product in subject table is taken, the benchmark image includes: reference characteristic amount, and, reference line；
According to the coordinate of the characteristic quantity, and, reference line calculates characteristic quantity number at a distance from reference line
Group, according to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
The mobile preset bias of the subject table is controlled, obtains the seat of multiple characteristic quantities of product in subject table again
Mark, calculates characteristic quantity array at a distance from reference line；
The range deviation quadratic sum apart from array and the reference distance array is calculated, it is flat according to the range deviation
Side and determining target offset amount, the target offset amount are to align result.
Selectable, described the step of controlling the subject table movement preset bias, includes:
X, Y, θ coordinate direction that the subject table is controlled along the subject table move fixed range, and/or think, rotation
Turn fixed angle.
Selectable, the characteristic quantity includes: characteristic point, and/or, straight line.
Selectable, described the step of obtaining the coordinate of multiple characteristic quantities of product in subject table, includes:
Obtain the image of product in the subject table；
According to the image of the product, the characteristic quantity of the product is positioned, the characteristic quantity includes: the placed in the middle right of institute's product
Site or straight line placed in the middle.
It is selectable, range deviation quadratic sum of the calculating apart from array and the reference distance array, according to described
Range deviation quadratic sum determines that the step of target offset amount includes:
It calculates separately apart from array and reference distance array range deviation quadratic sum；
Inverse iteration is carried out by gradient descent algorithm and calculates solution, if range deviation quadratic sum reaches threshold value or reaches
Maximum number of iterations, offset result at this time are target offset amount.
It is selectable, it is described to calculate the range deviation quadratic sum apart from array, according to the range deviation quadratic sum
The step of determining target offset amount include:
It calculates separately apart from array and reference distance array range deviation quadratic sum；
Determine that generating the corresponding offset of minimum range sum of square of deviations is target offset amount.
It is selectable, described the step of obtaining the coordinate of multiple characteristic quantities of product in subject table specifically:
The coordinate of characteristics of objects amount is obtained by framing algorithm.
The embodiment of the present application second aspect shows a kind of based on the alignment device for minimizing range deviation, described device packet
It includes:
Acquiring unit, is used for initialization offset amount, and object images acquisition device obtains the seat of benchmark image on target platform
Mark obtains the coordinate of multiple characteristic quantities of product in subject table；
Computing unit obtains benchmark image on target platform for object images acquisition device, on the benchmark image
Coordinate system to be established, the coordinate of multiple characteristic quantities of product in subject table is obtained, the benchmark image includes: reference characteristic amount,
And reference line；
Array computing unit, for the coordinate according to the characteristic quantity, and, reference line, calculate the characteristic quantity with
Reference line apart from array, according to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
Control unit obtains product in subject table for controlling the mobile preset bias of the subject table again
The coordinate of multiple characteristic quantities calculates characteristic quantity array at a distance from reference line；
Determination unit, for calculating the range deviation quadratic sum apart from array and the reference distance array, according to
The range deviation quadratic sum determines that target offset amount, the target offset amount are to align result.
Selectable, the determination unit includes:
First computing unit, for calculating separately apart from array and reference distance array range deviation quadratic sum；
Unit is solved, solution is calculated for carrying out inverse iteration by gradient descent algorithm, if range deviation quadratic sum reaches
To threshold value or reach maximum number of iterations, offset result at this time is target offset amount.
Selectable, the determination unit includes:
Second computing unit, for calculating separately apart from array and reference distance array range deviation quadratic sum；
First determination unit, for determining that generating the corresponding offset of minimum range sum of square of deviations is target offset amount.
From the above technical scheme, the embodiment of the present application show it is a kind of based on minimize range deviation alignment method and
Device, firstly, the alignment method for minimizing range deviation is directly started with using aligning product specification as target, the contraposition result of calculating
More meet the evaluation criteria of production；Secondly, the alignment method shown in the embodiment of the present application, can efficiently solve non-contraposition placed in the middle
Problem, the contraposition scene of irregular shape product suitable for actual production.Alignment method of the present invention specifically includes that production
The positioning of product characteristic quantity, characteristic quantity offset iterations and range deviation quadratic sum calculate three parts.Characteristics of image dotted line positions
The necessary characteristic point of the algorithm is obtained according to framing algorithm, and/or, the coordinate representation of straight line.Characteristic quantity offset iterations are
Refer in optimizing distance contraposition, offset can the variation of continuous iteration, and needed after each iteration by characteristic point by newest inclined
Shifting amount is deviated, and makes a little constantly to level off to reference distance to the distance of respective straight.Minimizing range difference and calculating is that iteration is inclined
The calculating process of shifting amount, when all the points after offset to respective straight distance are and corresponding reference distance deviation is sufficiently small,
Then offset at this time is final contraposition result.
Detailed description of the invention
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without creative efforts, can also obtain according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the structural schematic diagram that a kind of common machine vision alignment system exemplified is preferably implemented according to one；
Fig. 2 is that a kind of structural frames based on the alignment method for minimizing range deviation exemplified are preferably implemented according to one
Figure；
Fig. 3 sheet is the schematic diagram that the subject table reference-calibrating position exemplified is preferably implemented according to one；
Fig. 4 sheet is the schematic diagram that the benchmark image exemplified is preferably implemented according to one, and, the signal of benchmark image coordinate
Figure；
Fig. 5 sheet is the schematic diagram that the θ coordinate exemplified is preferably implemented according to one；
Fig. 6 sheet is to be preferably implemented to exemplify the schematic diagram of product in subject table according to one；
Fig. 7 sheet is to be preferably implemented to exemplify the postrotational schematic diagram of product according to one；
Fig. 8 sheet is to be preferably implemented to exemplify the schematic diagram of product translation according to one；
Fig. 9 sheet is that the schematic diagram after production interchange is shown according to another preferred embodiment.
Specific embodiment
Below in conjunction with the attached drawing in the technical solution shown in the embodiment of the present invention, to the technology shown in the embodiment of the present invention
Technical solution in scheme carries out clear, complete description, it is clear that described embodiment is only that present invention a part is implemented
Example, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not making creativeness
Every other embodiment obtained, shall fall within the protection scope of the present invention under the premise of labour.
Machine vision is a kind of automatic acquisition analysis image, to obtain one scenery of description or control the data of certain movement
Technology.During modern industry automated production, machine vision is just becoming a kind of raising production efficiency and is guaranteeing product matter
The key technology of amount.Alignment system based on machine vision, that is, machine vision alignment system refers to and utilizes CCD (Charge
Coupled Device, charge coupled cell) image acquiring devices such as video camera are as imaging sensor, at integrated use image
The system that the technologies such as reason, motion algorithm carry out non-contact two dimension or three-dimensional coordinate alignment, it has merged electronics based on optics
Technology, computer technology and image processing techniques have many advantages, such as that precision is high, high-efficient, high degree of automation, low cost.Its
In, image processing techniques is the key technology of machine vision alignment system, including picture smooth treatment, image sharpening, image pair
Than degree enhancing algorithm etc., analyzed by the color character etc. to acquisition target image, obtain acquisition target profile information,
The data such as location information.
Machine vision alignment system is widely used in the preparation and detection process of LCD and semiconductor, generally comprises CCD and takes the photograph
Camera, three-dimensional mobile platform and step motor control unit, image processing unit, computer system control unit, result output
And feedback unit etc..Such as in the assembling process of mobile phone liquid crystal screen, machine vision alignment system by mobile phone liquid crystal screen and
Acquisition, filtering, separation and the identification of the contraposition mark of backlight module, are obtained the departure between image tagged position, are controlled with this
Subject table movement processed, completion mobile phone liquid crystal screen and backlight module precisely align.In the actual production process, machine vision pair
For position system according to composite factors such as whole Machine Design, movement process, cost control, being generally designed as subject table can be with
The platform of multiple freedom degree accurate movements is carried out, and target platform is generally fixed static platform.
In order to guarantee the precision of machine vision alignment system, it is necessary to the machine vision alignment system before production
It is demarcated.The process of calibration, the plane of delineation and subject table plane and CCD for exactly establishing CCD camera shooting image
The plane of delineation of machine shooting and the coordinate mapping relations of target platform plane.The levels of precision of calibration directly determines aligning accuracy,
And then the quality of product is influenced, it is only needed in the actual production process certainly in installation CCD camera or CCD camera position for the first time
It sets and is just demarcated when changing.The scaling method and dress of machine vision alignment system provided in an embodiment of the present invention
It sets, it is mobile by control object platform, obtain the object images changes in coordinates and subject table coordinate of datum mark in subject table
Variation accurately calculates the calibration completed to subject table；Then the multiple marks for the real-time objects being set in subject table are obtained
Remember point image, according to the calibration result of subject table, calculates the subject table grid deviation obtained between adjacent marker point；Finally
Real-time objects are fitted into target platform, obtain the target image grid deviation between adjacent marker point, and in conjunction with the label
The subject table grid deviation of point calculates the calibration result for determining target platform, completes target platform calibration.Entire calibrated
Cheng Zhong can complete the calibration to static target platform by the calibration to moveable subject table, have very high degree of precision
And applicability.
Referring to Fig. 2, the embodiment of the present application shows a kind of alignment method based on minimum range deviation, the method packet
It includes:
S101 object images acquisition device obtains benchmark image on target platform, establishes coordinate on the benchmark image
System obtains the coordinate of multiple characteristic quantities of product in subject table, and the benchmark image includes: reference characteristic amount, and, benchmark
Straight line.
In the technical solution shown in the embodiment of the present invention, the object images acquisition device can image for industrial CCD
Machine or industry CMOS (Complementary Metal Oxide Semiconductor compensates metal-oxide semiconductor (MOS))
Video camera adjusts focal length, the light of video camera according to the size of subject table and video camera to the distance etc. of the subject table
The parameters such as circle, so that can clearly obtain the product of subject table datum mark in camera field of view.Fig. 3 is that the present invention is implemented
The schematic diagram for the subject table reference-calibrating position that example provides, wherein institute, 11 be subject table, and 12 be product；By adjusting taking the photograph
Camera enables the visual field of the video camera clearly to obtain the image of product 12 in the subject table 11, and guarantees institute
It states and product 11 is located in the visual field by paracentral position.Certainly, the position of product 12 can be in the subject table 11
For any point in the subject table, as long as the product 12 immobilizes and is easy relative to the subject table 11
It is obtained by the object images acquisition device, size and position of the embodiment of the present invention to 11 datum mark of subject table
With no restrictions, those skilled in the art can arbitrarily select according to actual needs.
In the technical solution shown in the embodiment of the present invention, the image for the product that benchmark image acquisition device obtains, and,
Benchmark image is flat image.
Then the coordinate system of benchmark image is established on benchmark image, as shown in figure 4, the coordinate system of the benchmark image can
To establish coordinate system by origin of an apex angle of visual field benchmark image, the X-axis and Y-axis of the object images coordinate system are put down respectively
Row is in the outer edge sideline of the benchmark image, and X-axis is to be positive to the right, and Y-axis to be positive upwards；The benchmark image coordinate system,
Coordinate system can also be established as origin using any one vertex of benchmark image.In general, the X-axis of the benchmark image coordinate system and
Y-axis is respectively parallel to the sideline of the subject table, and X-axis is to be positive to the right, and Y-axis to be positive upwards.
It is worth noting that, the application shows technical solution in real time, in addition to being related to the translation of product during contraposition
Outside, the rotation of product is further related to, therefore, the technical solution shown in the embodiment of the present application, benchmark image coordinate system further includes that θ is sat
Mark.It is determined referring to Fig. 5, being rotated using origin as rotation center using the subject table, rotates clockwise and be positive, rotated counterclockwise
It is negative.Certainly, described in those skilled in the art can be established with any preset coordinate origin and corresponding change in coordinate axis direction
Object images coordinate system and the subject table coordinate system, above-mentioned coordinate system are also not necessarily limited to rectangular coordinate system, or other
Coordinate system, such as polar coordinate system etc..
Wherein, in subject table the coordinate of the characteristic quantity of product acquisition methods:
The image for obtaining product in the subject table obtains the characteristic quantity of image, the spy by framing algorithm
Sign amount includes: characteristic point, and/or, straight line.
Product image characteristic point and straight line are obtained by framing algorithm.For the product of a triangle, triangle
Apex angle be characterized a little, the side of triangle is straight line.It in practical applications can be according to the shape of product according to framing method
Carry out the characteristic quantity of positioning product.
S102 according to the coordinate of the characteristic quantity, and, reference line calculates the characteristic quantity at a distance from reference line
Array, according to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
Firstly, according to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
It can be illustrated according to production specification, obtain the standard of product apart from array；
It is also possible to obtain the reference characteristic amount of benchmark image by framing algorithm, the reference characteristic amount includes:
Reference characteristic point and/or reference line.
Referring to Fig. 4, triangle in figure is benchmark image, the rectangle positioned at the right, and, following rectangle is close to three
Angular side is respectively reference line, according to production specification illustrate benchmark image reference distance array be L1 ', L2 ',
L3’}。
Then, according to the coordinate of the characteristic quantity, and, reference line, calculate the characteristic quantity and reference line away from
From array.
Specifically, wherein A, B, C are respectively referring to Fig. 6, Fig. 6 is the original state of triangle object in subject table
Three characteristic quantities of the object.
Coordinate system is constructed previously according to, benchmark image, using the X-axis of coordinate system and Y-axis as the first reference line and the
Two reference lines；
In a coordinate system, characteristic quantity A under original state, the coordinate that 3 points of B, C be respectively (X10, Y10), (X20, Y20),
(X30, Y30).
Corresponding relationship between characteristic quantity and reference line, Ke Yiwei, determine characteristic quantity to the shortest distance of reference line be institute
State the corresponding reference line of characteristic quantity.
The corresponding reference line of C in Fig. 6, B is X-axis, and the corresponding reference line of A is Y-axis.
Characteristic quantity A is at a distance from reference line (Y-axis) are as follows: X10=L10, characteristic quantity B are at a distance from reference line (X-axis)
Are as follows: Y20=L20, characteristic quantity C are at a distance from reference line (X-axis) are as follows: Y30=L30.
Corresponding product apart from array be { L10, L20, L30 }.
Initialization offset amount.The initial translation length and rotation angle of characteristic point are taken, generally value is 0；
S103 controls the mobile preset bias of the subject table, obtains multiple characteristic quantities of product in subject table again
Coordinate, calculate characteristic quantity array at a distance from reference line；
Control object platform is mobile to move fixed range along the direction X, Y, θ of the subject table, and it is flat to obtain object again
The coordinate of multiple characteristic quantities of product on platform.
As shown in fig. 7, object is rotated θ angle on θ, the seat of multiple characteristic quantities of product in subject table is obtained again
Mark, the coordinate of characteristic quantity A1, B1, C1 are respectively (X11, Y11), (X21, Y21), (X31, Y31).
Characteristic quantity A1 is at a distance from reference line (Y-axis) are as follows: X11=L11, characteristic quantity B1 and reference line (X-axis) away from
From are as follows: Y21=L21, characteristic quantity C1 are at a distance from reference line (X-axis) are as follows: Y31=L31.
Second group is { L11, L21, L31 } apart from array, wherein second group is θ apart from the corresponding offset of array.
Please refer to Fig. 8 be subject table respectively along X and Y-direction move fixed range after reference point location schematic diagram,
In technical solution shown in the embodiment of the present invention, subject table moves fixed range Δ Xd along the negative direction of subject table X-axis,
Subject table moves fixed range Δ Yd, corresponding characteristic quantity A3, B3 along the positive direction of subject table Y-axis, C3's
Coordinate is respectively (X12, Y12), (X22, Y22), (X32, Y32).
Characteristic quantity A2 is at a distance from reference line (Y-axis) are as follows: X12=L12, characteristic quantity B1 and reference line (X-axis) away from
From are as follows: Y21=L22, characteristic quantity C1 are at a distance from reference line (X-axis) are as follows: Y31=L32.
S104 calculates the range deviation quadratic sum apart from array and the reference distance array, inclined according to the distance
Poor quadratic sum determines that target offset amount, the target offset amount are to align result.
Reference distance array is respectively { L1 ', L2 ', L3 ' }；
First group { L10, L20, L30 }.
Second group is { L11, L21, L31 } apart from array；
The calculation formula for optimizing range deviation can indicate are as follows:
It can calculate separately apart from array and reference distance array range deviation quadratic sum；
Inverse iteration is carried out by gradient descent algorithm and calculates solution, if range deviation quadratic sum reaches threshold value or reaches
Maximum number of iterations, offset result at this time are target offset amount.
It can also be to calculate separately apart from array and reference distance array range deviation quadratic sum；
Determine that generating the corresponding offset of minimum range sum of square of deviations is target offset amount.
Rotating the angle θ is target offset amount.
Preferably, it is aligned on one side for gauged distance with straight line come compatible by adding the point for participating in aligning between two parties, another both sides
For the contraposition scene aligned between two parties.
Specifically, obtaining the image of product in the subject table；
According to the image of the product, the characteristic quantity of the product is positioned, the characteristic quantity includes: the placed in the middle right of institute's product
Site or straight line placed in the middle.
Referring to Fig. 9, L4 and L5 is the actual range for participating in center ofthe contraposition in figure, after contraposition placed in the middle, two distances at
It is equal and be equal to L '；And L6 participates in gauged distance contraposition, the distance levels off to gauged distance L6 ' after contraposition.
The embodiment of the present application second aspect shows a kind of based on the alignment device for minimizing range deviation, described device packet
It includes:
Acquiring unit, is used for initialization offset amount, and object images acquisition device obtains the seat of benchmark image on target platform
Mark obtains the coordinate of multiple characteristic quantities of product in subject table；
Computing unit obtains benchmark image on target platform for object images acquisition device, on the benchmark image
Coordinate system to be established, the coordinate of multiple characteristic quantities of product in subject table is obtained, the benchmark image includes: reference characteristic amount,
And reference line；
Array computing unit, for the coordinate according to the characteristic quantity, and, reference line, calculate the characteristic quantity with
Reference line apart from array, according to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
Control unit obtains product in subject table for controlling the mobile preset bias of the subject table again
The coordinate of multiple characteristic quantities calculates characteristic quantity array at a distance from reference line；
Determination unit, for calculating the range deviation quadratic sum apart from array and the reference distance array, according to
The range deviation quadratic sum determines that target offset amount, the target offset amount are to align result.
Selectable, the determination unit includes:
First computing unit, for calculating separately apart from array and reference distance array range deviation quadratic sum；
Unit is solved, solution is calculated for carrying out inverse iteration by gradient descent algorithm, if range deviation quadratic sum reaches
To threshold value or reach maximum number of iterations, offset result at this time is target offset amount.
Selectable, the determination unit includes:
Second computing unit, for calculating separately apart from array and reference distance array range deviation quadratic sum；
First determination unit, for determining that generating the corresponding offset of minimum range sum of square of deviations is target offset amount
From the above technical scheme, the embodiment of the present application show it is a kind of based on minimize range deviation alignment method and
Device, firstly, the alignment method for minimizing range deviation is directly started with using aligning product specification as target, the contraposition result of calculating
More meet the evaluation criteria of production；Secondly, the alignment method shown in the embodiment of the present application, can efficiently solve non-contraposition placed in the middle
Problem, the contraposition scene of irregular shape product suitable for actual production.Alignment method of the present invention specifically includes that production
The positioning of product characteristic quantity, characteristic quantity offset iterations and range deviation quadratic sum calculate three parts.Characteristics of image dotted line positions
The necessary characteristic point of the algorithm is obtained according to framing algorithm, and/or, the coordinate representation of straight line.Characteristic quantity offset iterations are
Refer in optimizing distance contraposition, offset can the variation of continuous iteration, and needed after each iteration by characteristic point by newest inclined
Shifting amount is deviated, and makes a little constantly to level off to reference distance to the distance of respective straight.Minimizing range difference and calculating is that iteration is inclined
The calculating process of shifting amount, when all the points after offset to respective straight distance are and corresponding reference distance deviation is sufficiently small,
Then offset at this time is final contraposition result.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to of the invention its
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the invention, these modifications, purposes or
Person's adaptive change follows general principle of the invention and including the undocumented common knowledge in the art of the present invention
Or conventional techniques.The description and examples are only to be considered as illustrative, and true scope and spirit of the invention are by following
Claim is pointed out.
It should be understood that the present invention is not limited to the precise structure already described above and shown in the accompanying drawings, and
And various modifications and changes may be made without departing from the scope thereof.The scope of the present invention is limited only by the attached claims.
Claims (10)
1. a kind of based on the alignment method for minimizing range deviation, which is characterized in that the described method includes:
Object images acquisition device obtains benchmark image on target platform, and coordinate system, acquisition pair are established on the benchmark image
As the coordinate of multiple characteristic quantities of product on platform, the benchmark image includes: reference characteristic amount, and, reference line；
According to the coordinate of the characteristic quantity, and, reference line calculates characteristic quantity array at a distance from reference line, root
According to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
The mobile preset bias of the subject table is controlled, obtains the coordinate of multiple characteristic quantities of product in subject table again,
Calculate characteristic quantity array at a distance from reference line；
The range deviation quadratic sum apart from array and the reference distance array is calculated, according to the range deviation quadratic sum
Determine that target offset amount, the target offset amount are to align result.
2. the method according to claim 1, wherein the control subject table moves preset bias
Step includes:
X, Y, θ coordinate direction that the subject table is controlled along the subject table move fixed range, and/or think, rotation is solid
Fixed angle.
3. the method according to claim 1, wherein the characteristic quantity includes: characteristic point, and/or, straight line.
4. according to the method described in claim 3, it is characterized in that, described obtain multiple characteristic quantities of product in subject table
The step of coordinate includes:
Obtain the image of product in the subject table；
According to the image of the product, the characteristic quantity of the product is positioned, the characteristic quantity includes: the contraposition placed in the middle of institute's product
Point or straight line placed in the middle.
5. method according to claim 1-4, which is characterized in that the calculating is apart from array and the stand-off
Range deviation quadratic sum from array, the step of determining target offset amount according to the range deviation quadratic sum include:
It calculates separately apart from array and reference distance array range deviation quadratic sum；
Inverse iteration is carried out by gradient descent algorithm and calculates solution, if range deviation quadratic sum reaches threshold value or reaches maximum
The number of iterations, offset result at this time are target offset amount.
6. method according to claim 1-4, which is characterized in that the calculating distance apart from array is inclined
Poor quadratic sum, the step of determining target offset amount according to the range deviation quadratic sum include:
It calculates separately apart from array and reference distance array range deviation quadratic sum；
Determine that generating the corresponding offset of minimum range sum of square of deviations is target offset amount.
7. method according to claim 1-4, which is characterized in that product is multiple in the acquisition subject table
The step of coordinate of characteristic quantity specifically:
The coordinate of characteristics of objects amount is obtained by framing algorithm.
8. a kind of based on the alignment device for minimizing range deviation, which is characterized in that described device includes:
Acquiring unit, is used for initialization offset amount, and object images acquisition device obtains the coordinate of benchmark image on target platform, obtains
Take the coordinate of multiple characteristic quantities of product in subject table；
Computing unit obtains benchmark image on target platform for object images acquisition device, establishes on the benchmark image
Coordinate system obtains the coordinate of multiple characteristic quantities of product in subject table, and the benchmark image includes: reference characteristic amount, and,
Reference line；
Array computing unit, for the coordinate according to the characteristic quantity, and, reference line calculates the characteristic quantity and benchmark
Straight line apart from array, according to the reference characteristic amount, and, reference line calculating benchmark is apart from array；
Control unit obtains the multiple of product in subject table for controlling the mobile preset bias of the subject table again
The coordinate of characteristic quantity calculates characteristic quantity array at a distance from reference line；
Determination unit, for calculating the range deviation quadratic sum apart from array and the reference distance array, according to described
Range deviation quadratic sum determines that target offset amount, the target offset amount are to align result.
9. device according to claim 8, which is characterized in that the determination unit includes:
First computing unit, for calculating separately apart from array and reference distance array range deviation quadratic sum；
Unit is solved, solution is calculated for carrying out inverse iteration by gradient descent algorithm, if range deviation quadratic sum reaches threshold
It is worth or reaches maximum number of iterations, offset result at this time is target offset amount.
10. device according to claim 8, which is characterized in that the determination unit includes:
Second computing unit, for calculating separately apart from array and reference distance array range deviation quadratic sum；
First determination unit, for determining that generating the corresponding offset of minimum range sum of square of deviations is target offset amount.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811358794.3A CN109191527B (en) | 2018-11-15 | 2018-11-15 | Alignment method and device based on minimum distance deviation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811358794.3A CN109191527B (en) | 2018-11-15 | 2018-11-15 | Alignment method and device based on minimum distance deviation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109191527A true CN109191527A (en) | 2019-01-11 |
| CN109191527B CN109191527B (en) | 2021-06-11 |
Family
ID=64939460
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811358794.3A Active CN109191527B (en) | 2018-11-15 | 2018-11-15 | Alignment method and device based on minimum distance deviation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109191527B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110928235A (en) * | 2019-11-18 | 2020-03-27 | 广东利元亨智能装备股份有限公司 | Workpiece bonding method and device, electronic equipment and workpiece bonding system |
| CN112845115A (en) * | 2020-12-31 | 2021-05-28 | 华芯智造微电子(重庆)有限公司 | Efficient automatic sorting method |
| CN112845187A (en) * | 2020-12-31 | 2021-05-28 | 华芯智造微电子(重庆)有限公司 | Automatic sorting machine |
| CN112976559A (en) * | 2021-02-05 | 2021-06-18 | 广东科升智能装备有限公司 | Alignment method and film pasting method based on linear array CCD |
| CN115355822A (en) * | 2022-10-19 | 2022-11-18 | 成都新西旺自动化科技有限公司 | Special-shaped alignment calculation method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090214166A1 (en) * | 2005-12-30 | 2009-08-27 | Wei-Ping Huang | Positioning optical fibers |
| CN101680756A (en) * | 2008-02-12 | 2010-03-24 | 松下电器产业株式会社 | Compound eye imaging device, distance measurement device, parallax calculation method and distance measurement method |
| CN102037800A (en) * | 2008-03-25 | 2011-04-27 | 伊利诺斯工具制品有限公司 | Method and apparatus for dispensing material on a substrate |
| CN102183205A (en) * | 2011-01-19 | 2011-09-14 | 北京航空航天大学 | Method for matching optimal assembly poses of large-sized parts |
| CN105427288A (en) * | 2015-11-10 | 2016-03-23 | 凌云光技术集团有限责任公司 | Calibration method and device of machine vision alignment system |
| CN106127722A (en) * | 2016-05-03 | 2016-11-16 | 深圳视觉龙智能传感器有限公司 | The demarcation of polyphaser and para-position applying method |
-
2018
- 2018-11-15 CN CN201811358794.3A patent/CN109191527B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090214166A1 (en) * | 2005-12-30 | 2009-08-27 | Wei-Ping Huang | Positioning optical fibers |
| CN101680756A (en) * | 2008-02-12 | 2010-03-24 | 松下电器产业株式会社 | Compound eye imaging device, distance measurement device, parallax calculation method and distance measurement method |
| CN102037800A (en) * | 2008-03-25 | 2011-04-27 | 伊利诺斯工具制品有限公司 | Method and apparatus for dispensing material on a substrate |
| CN102183205A (en) * | 2011-01-19 | 2011-09-14 | 北京航空航天大学 | Method for matching optimal assembly poses of large-sized parts |
| CN105427288A (en) * | 2015-11-10 | 2016-03-23 | 凌云光技术集团有限责任公司 | Calibration method and device of machine vision alignment system |
| CN106127722A (en) * | 2016-05-03 | 2016-11-16 | 深圳视觉龙智能传感器有限公司 | The demarcation of polyphaser and para-position applying method |
Non-Patent Citations (3)
| Title |
|---|
| SANG HO PARK 等: "A Line Feature Matching Technique Based on an Eigenvector Approach", 《COMPUTER VISION AND IMAGE UNDERSTANDING》 * |
| 于海霞 等: "视觉检测技术在矫直机在线检测系统中的应用", 《大连工业大学学报》 * |
| 卢金燕 等: "基于多传感器的大口径器件自动对准策略", 《自动化学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110928235A (en) * | 2019-11-18 | 2020-03-27 | 广东利元亨智能装备股份有限公司 | Workpiece bonding method and device, electronic equipment and workpiece bonding system |
| CN110928235B (en) * | 2019-11-18 | 2021-04-06 | 广东利元亨智能装备股份有限公司 | Workpiece bonding method and device, electronic equipment and workpiece bonding system |
| CN112845115A (en) * | 2020-12-31 | 2021-05-28 | 华芯智造微电子(重庆)有限公司 | Efficient automatic sorting method |
| CN112845187A (en) * | 2020-12-31 | 2021-05-28 | 华芯智造微电子(重庆)有限公司 | Automatic sorting machine |
| CN112976559A (en) * | 2021-02-05 | 2021-06-18 | 广东科升智能装备有限公司 | Alignment method and film pasting method based on linear array CCD |
| CN112976559B (en) * | 2021-02-05 | 2022-11-25 | 广东科升智能装备有限公司 | Alignment method and film pasting method based on linear array CCD |
| CN115355822A (en) * | 2022-10-19 | 2022-11-18 | 成都新西旺自动化科技有限公司 | Special-shaped alignment calculation method and system |
| CN115355822B (en) * | 2022-10-19 | 2023-01-17 | 成都新西旺自动化科技有限公司 | Special-shaped alignment calculation method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109191527B (en) | 2021-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109191527A (en) | A kind of alignment method and device based on minimum range deviation | |
| CN111536902B (en) | Galvanometer scanning system calibration method based on double checkerboards | |
| CN108921901B (en) | Large-view-field camera calibration method based on precise two-axis turntable and laser tracker | |
| CN105427288B (en) | A kind of scaling method and device of machine vision alignment system | |
| CN106529587B (en) | Vision course recognition methods based on object detection | |
| CN109859272B (en) | Automatic focusing binocular camera calibration method and device | |
| Zhang et al. | A robust and rapid camera calibration method by one captured image | |
| CN107680156B (en) | Three-dimensional reconstruction method based on polarization information | |
| CN110858899B (en) | Method and system for measuring optical axis center and field angle of camera movement | |
| CN107339935B (en) | Target space intersection measuring method for full-view scanning measuring system | |
| CN102768762B (en) | Digital camera calibration method targeted to shield tunnel defect digital radiography detection and device thereof | |
| CN102221331B (en) | Measuring method based on asymmetric binocular stereovision technology | |
| CN106871787A (en) | Large space line scanning imagery method for three-dimensional measurement | |
| CN205333067U (en) | Laser photogrammetric survey system | |
| CN108805940B (en) | Method for tracking and positioning zoom camera in zooming process | |
| CN109544643A (en) | A kind of camera review bearing calibration and device | |
| CN103795935B (en) | A kind of camera shooting type multi-target orientation method and device based on image rectification | |
| CN109656033A (en) | A kind of method and device for distinguishing liquid crystal display dust and defect | |
| CN101487702A (en) | Binocular vision based traffic accident on-site photogrammetric survey method | |
| WO2020113978A1 (en) | Method for calculating center position of hole located on plane | |
| KR101430970B1 (en) | Alligning method of display panel | |
| CN105809685A (en) | Single-concentric circle image-based camera calibration method | |
| CN206583440U (en) | A kind of projected image sighting distance detecting system | |
| CN113781581B (en) | Depth of field distortion model calibration method based on target loose attitude constraint | |
| CN108898585A (en) | A kind of axial workpiece detection method and its device |
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 | ||
| CB02 | Change of applicant information |
Address after: 100094 Beijing city Haidian District Cui Hunan loop 13 Hospital No. 7 Building 7 room 701 Applicant after: Lingyunguang Technology Co., Ltd Address before: 100094 Beijing city Haidian District Cui Hunan loop 13 Hospital No. 7 Building 7 room 701 Applicant before: LUSTER LIGHTTECH GROUP Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |