CN102252611A - Geometric positioning method - Google Patents
Geometric positioning method Download PDFInfo
- Publication number
- CN102252611A CN102252611A CN201110118014XA CN201110118014A CN102252611A CN 102252611 A CN102252611 A CN 102252611A CN 201110118014X A CN201110118014X A CN 201110118014XA CN 201110118014 A CN201110118014 A CN 201110118014A CN 102252611 A CN102252611 A CN 102252611A
- Authority
- CN
- China
- Prior art keywords
- sample
- geometric properties
- tested
- invariants
- angle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses a method for accurately positioning an object at high speed based on optical geometrical characteristics of an image. The method comprises the following steps of: collecting a sample, performing edge extraction on the shape of a standard sample by a sobel operator, establishing a mathematical model by using seven invariants of Hu and describing the sample, collecting the pattern of the sample to be tested by an industrial camera, comparing the standard sample and the sample to be tested, and calculating and positioning. The algorithm is to compare the contours of images and is not related to a gray value of a pixel, so that the method is not influenced by the change of radiance; and the seven invariants of Hu can calculate object scaling and an angle, so that the seven invariants are not influenced by scaling and the change of angle. During calculation, a pyramid algorithm and an multi media extension (MMX) acceleration algorithm are used, so that the algorithms are high in computation speed, and low in error and are suitable for positioning of automatic production equipment, such as surface mount technologies (SMT), light-emitting diodes (LED), printed circuit boards (PCB), solar plate diving machines and the like.
Description
[technical field]
The present invention relates to the automated production equipment of high-accuracy processing and manufacturing industry, relate in particular to the location of automated production equipments such as SMT, LED, PCB, sun power board separator.
[background technology]
In the modern production manufacturing industry, making electronics industry is a ring important in the industrial chain, along with the electronic product technology constantly develops, electronic product is constantly compact and lightening, and function is more and more, from strength to strength, technically, smaller volume, chip integration is more and more higher on the one hand to be structural design with the aspect, the density of chip significantly improves, this accuracy requirement to automated production equipment just begins to improve, simultaneously, and along with electronic product, LCD, the price of LED product is more and more lower, and production efficiency becomes the height of decision enterprise profitability, and this speed and degree of reliability to automated production equipment requires also very high, and for the instrument of chip mounter and board separator and so on, most important is exactly the accuracy of location.
The method that is used to locate on the industrial automation production equipment is a lot, and modal have a mechanical clamp location, and photoelectric sensor location, pressure transducer location, optical imagery location comprise based on the location of gray scale with based on how much location.In early days because computer speed is slow, data-handling capacity a little less than, generally the localization method of Cai Yonging is that locate with sensor mechanical clamp location.The main shortcoming of these two kinds of localization methods is that bearing accuracy is low, and speed is slow, and partly also needs to rely on manually-operated, can only be semi-automatic.Later enhancing in 2000 along with computer process ability, and image algorithm is constantly perfect, increasing industrial automation equipment has adopted the method for optical imagery location, especially abroad on the high-end automation equipment, for example, the semiconductor crystal wafer production equipment, the PCB production equipment, the SMT production equipment, LCD panel manufacturing equipment, solar panels production equipment etc.It is consistent that differentiation on this automatic technology is obtained information with the mankind, human information more than 90% obtains by eyes, and the information that obtains by sense of touch is less than 1%, and the collection of optical imagery is like being that machine is obtaining information with " eyes ", and location algorithm is exactly that machine is in the process with " brain " thinking.And the optical imagery localization method is divided into gray scale location and geometry location, is the position that the size of correlativity of the gray-scale value of gray-scale value by analyzing template image and object under test image is come judgment object based on the location algorithm of gray scale.Because the gray-scale value close relation of this algorithm and each picture element of image, so it is very responsive for the brightness variation of illumination.And workpiece is to the absorption of light in the middle of actual production, and reflection can not be in full accord, therefore can influence the accuracy rate and the percent of pass of location; The gray scale algorithm is the similarity of coming calculation template and object under test by autocorrelation function in addition, so if object under test has certain convergent-divergent dimensionally or certain angular deflection is arranged, coefficient of autocorrelation can sharply reduce.And in the practical application in industry, the size of workpiece can not be in full accord, and the angle of putting also can not be always zero, and this also can influence locating accuracy and percent of pass.
[summary of the invention]
Thereby the present invention is directed to above problem has proposed a kind ofly to be applicable to automation equipment especially, has overcome object under test is difficult to the convergent-divergent avoided or certain angle on object under test deflection, has solved location based on gray scale to the very responsive a kind of geometry location method that influences the problem of precision of light.This method can reach suitable device for location technology high speed, efficient, high-precision requirement.
In order to realize above effect, the technical scheme that the present invention takes is: whole flow processs of described geometry location method, comprise links such as setting localization criteria, collection testing data, contrast, calculating it is characterized in that concrete disposal route may further comprise the steps,
(1), gathers the image of master sample with industrial digital camera;
(2), do edge extracting, this shape can be arbitrarily by the shape of sobel operator extraction master sample;
(3), with seven invariants of Hu shape, angle and the locus of sample are described out with following mathematical model:
I(A,B)=sum
i=1..7abs(1/m
A i?-?1/m
B i),
M wherein
A i=sign (h
A i) log (h
A i);
m
B i=sign(h
B i)·log(h
B i);
h
A i, h
B iThe Hu square of-A and B; This model can be used as the geometric properties of sample;
(4), the cross correlation of the geometric properties of all samples is extracted, as template;
(5), gather the image of testing sample with industrial digital camera;
(6), with seven invariants of Hu shape, angle and the locus of sample to be tested are described out with aforementioned mathematical model;
(7), the template of master sample is compared with the geometric properties of the template of sample to be tested, and by pyramid algorith and MMX software accelerating algorithm, draw volume coordinate, the anglec of rotation and the scaling of the highest geometric properties of similarity, the result of gained is exactly the required result in location
Described pyramid algorith is meant 1980, the hand-held laplacian pyramid algorithm that proposes of Peter and Ted Adelson, the compression performance that this algorithm is resolved thought and sale owing to many resolutions of its advanced person is widely used in image compression encoding, described MMX software accelerating algorithm is actual to be a multimedia extension to the Pentium instruction set, its core concept is to adapt to the low precision of multimedia era, the data flow characteristics of big data quantity, utilize the data path of Pentium 64 bit widths fully, the a plurality of data of parallel processing in an instruction cycle, draw the volume coordinate of the highest geometric properties of similarity, the anglec of rotation and scaling, the result of gained is exactly the required result in location.
Described sobel operator is the mutual relationship of geometric properties, or common feature on master sample and the sample to be tested, such as mark or the distance between center of gravity and the angle of geometric properties, location.
Beneficial effect of the present invention: described sobel operator itself is exactly one of operator in the Flame Image Process, and mainly as rim detection, technically, it is a discreteness difference operator, is used for the approximate value of gradient of arithmograph image brightness function.Use this operator in any point of image, will produce corresponding gradient vector or its method vector, the masterplate edge of image is extracted, then edge image is converted into geometric profile image set (Contour), and the profile collection is rearranged by the weight size according to the HU invariant, with mathematical model template is described out, with pyramid algorith template and image outline are carried out correlativity scanning comparison at last, and rearrange by the zone in big young pathbreaker's image of correlativity score.The position of score maximum is exactly the result of location.Because this algorithm is that the profile of image is compared, thus with it doesn't matter its influence that not changed by luminance brightness of the gray-scale value of pixel; So seven invariants of Hu can calculate the influence that it is not changed by scaling and angle to object convergent-divergent and angle.
[description of drawings]
Fig. 1 is the process flow diagram of geometry location method of the present invention.
[embodiment]
In order to make purpose of the present invention, technical scheme and advantage clearer,, the present invention is further elaborated below in conjunction with drawings and Examples.
Whole flow processs of described geometry location method comprise links such as setting localization criteria, collection testing data, contrast, calculating it is characterized in that concrete disposal route may further comprise the steps,
(1), gathers the image of master sample with industrial digital camera;
(2), do edge extracting, this shape can be arbitrarily by the shape of sobel operator extraction master sample;
(3), with seven invariants of Hu shape, angle and the locus of sample are described out with following mathematical model:
I(A,B)=sum
i=1..7abs(1/m
A i?-?1/m
B i),
M wherein
A i=sign (h
A i) log (h
A i);
m
B i=sign(h
B i)·log(h
B i);
h
A i, h
B iThe Hu square of-A and B; This model can be used as the geometric properties of sample;
(4), the cross correlation of the geometric properties of all samples is extracted, as template;
(5), gather the image of testing sample with industrial digital camera;
(6), with seven invariants of Hu shape, angle and the locus of sample to be tested are described out with aforementioned mathematical model;
(7), the template of master sample is compared with the geometric properties of the template of sample to be tested, and by pyramid algorith and MMX software accelerating algorithm, draw volume coordinate, the anglec of rotation and the scaling of the highest geometric properties of similarity, the result of gained is exactly the required result in location
Described sobel operator is the mutual relationship of geometric properties, or common feature on master sample and the sample to be tested, such as mark or the distance between center of gravity and the angle of geometric properties, location.
As the flow process among Fig. 1, take the image of master sample with industrial digital camera, basis as later contrast, utilize the sobel operator master sample to be extracted the edge of master sample, and the Hu invariant of basis of calculation sample is as template, then for determinand, also carry out industrial digital and take the image that head is taken determinand, utilize the sobel operator to extract the determinand edge equally, for seven invariants of determinand edge calculations determinand Hu that extract as template, comparing link at last, is that the Hu invariant data that obtain under two group models are compared, and draws in the testing image in how much the highest edge coordinates of template matches degree.
In concrete practical operation, we for the cruciform MARK point on the pcb board as master sample, at first selected object as master sample is taken, extract on it criss-cross geometric properties as template, the model of setting up with the Hu invariant calculates it, shooting at several objects under test is next extracted in the process of feature, takes following scheme: object under test one: change the brightness to its illumination; Object under test two: rotation pcb board; Object under test three: on pcb board, add impurity; The pcb board image of gathering above-mentioned situation then respectively is as sample to be tested, utilize above-mentioned algorithm computation to go out the locus that cruciform mark is ordered, the anglec of rotation and scaling, with the coordinate contrast that the result and the actual cruciform mark that calculate are ordered, error is in a pixel.
The repetitive positioning accuracy that experimental results show that this algorithm is in 1 pixel.Because used pyramid and MMX accelerating algorithm, the non-product of the computing velocity of this algorithm is fast.Use the computing machine of 2.8GHz dominant frequency, for the non-constant of picture quality, also can be controlled in the 30ms its positioning time between 10ms~15ms its positioning time, and being provided with the influence of its speed of template size is very little.
This technology is being incorporated in the board separator able to programme, described board separator is the optical, mechanical and electronic integration automatic equipment, it is defined in the CONFIG.SYS by the cutting route of the pcb board that the imageing sensor that is installed on the three-dimensional platform will cut, and the guiding milling cutter cuts accurately to plate.Be the automation equipment of cutting yoke plate PCB on the SMT production line, be mainly used in cutting apart and boring, the especially mobile phone of high integration of circuit board, the cutting apart of yoke plates such as MP3, it has replaced people's work point plate, improves the quality of products, and reduces scrappage.This equipment is mainly used on the SMT electronics paster production line.
Claims (2)
1. geometry location method, its whole flow processs comprise to be set localization criteria, gathers links such as testing data, contrast, calculating, it is characterized in that concrete disposal route may further comprise the steps,
(1), gathers the image of master sample with industrial digital camera;
(2), do edge extracting, this shape can be arbitrarily by the shape of sobel operator extraction master sample;
(3), with seven invariants of Hu shape, angle and the locus of sample are described out with following mathematical model:
I(A,B)=sum
i=1..7abs(1/m
A i?-?1/m
B i),
M wherein
A i=sign (h
A i) log (h
A i);
m
B i=sign(h
B i)·log(h
B i);
h
A i, h
B iThe Hu square of-A and B; This model can be used as the geometric properties of sample;
(4), the cross correlation of the geometric properties of all samples is extracted, as template;
(5), gather the image of testing sample with industrial digital camera;
(6), with seven invariants of Hu shape, angle and the locus of sample to be tested are described out with aforementioned mathematical model;
(7), the template of master sample is compared with the geometric properties of the template of sample to be tested, and by pyramid algorith and MMX software accelerating algorithm, draw volume coordinate, the anglec of rotation and the scaling of the highest geometric properties of similarity, the result of gained is exactly the required result in location.
2. according to the described geometry location method of claim 1, it is characterized in that, described sobel operator is the mutual relationship of geometric properties, or common feature on master sample and the sample to be tested, distance and angle between the geometric properties of sample to be tested, the mark of location or center of gravity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110118014XA CN102252611B (en) | 2011-05-09 | 2011-05-09 | Geometric positioning method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110118014XA CN102252611B (en) | 2011-05-09 | 2011-05-09 | Geometric positioning method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102252611A true CN102252611A (en) | 2011-11-23 |
CN102252611B CN102252611B (en) | 2013-11-27 |
Family
ID=44980035
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201110118014XA Expired - Fee Related CN102252611B (en) | 2011-05-09 | 2011-05-09 | Geometric positioning method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102252611B (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102607421A (en) * | 2012-04-12 | 2012-07-25 | 苏州天准精密技术有限公司 | Large-view-field image measuring method and equipment |
CN103217431A (en) * | 2012-01-19 | 2013-07-24 | 昆山思拓机器有限公司 | Image detection method of SMT screen plate |
CN105744824A (en) * | 2014-12-10 | 2016-07-06 | 安徽海创自动控制设备有限公司 | Automatic paster mounting device |
CN107250715A (en) * | 2015-03-27 | 2017-10-13 | 三菱电机株式会社 | Detection means |
CN107578431A (en) * | 2017-07-31 | 2018-01-12 | 深圳市海思科自动化技术有限公司 | A kind of Mark points visual identity method |
CN107944494A (en) * | 2017-11-30 | 2018-04-20 | 华中科技大学 | A kind of Mark point visual identitys and localization method based on not bending moment |
CN108458655A (en) * | 2017-02-22 | 2018-08-28 | 上海理工大学 | Support the data configurableization monitoring system and method for vision measurement |
CN109631763A (en) * | 2019-02-01 | 2019-04-16 | 东莞中科蓝海智能视觉科技有限公司 | Irregular part detects localization method |
CN109631803A (en) * | 2019-02-01 | 2019-04-16 | 东莞中科蓝海智能视觉科技有限公司 | Part angle detects adjusting method |
CN109919199A (en) * | 2019-02-13 | 2019-06-21 | 东南大学 | The detection method of Wind turbines abnormal data based on image procossing |
CN110826627A (en) * | 2019-11-06 | 2020-02-21 | 广东三维家信息科技有限公司 | Image similarity measuring method and device and electronic equipment |
CN111430289A (en) * | 2020-05-07 | 2020-07-17 | 上海果纳半导体技术有限公司 | Wafer positioning and calibrating device and wafer positioning and calibrating method |
CN112579810A (en) * | 2019-09-30 | 2021-03-30 | 深圳市嘉立创科技发展有限公司 | Printed circuit board classification method and device, computer equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6603563B1 (en) * | 2000-04-05 | 2003-08-05 | Accu-Sort Systems, Inc. | Apparatus for determining measurements of an object utilizing negative imaging |
EP1519144A1 (en) * | 2003-09-29 | 2005-03-30 | Nederlandse Organisatie voor toegepast-natuurwetenschappelijk onderzoek TNO | Free-form optical surface measuring apparatus and method |
CN101144703A (en) * | 2007-10-15 | 2008-03-19 | 陕西科技大学 | Article geometrical size measuring device and method based on multi-source image fusion |
-
2011
- 2011-05-09 CN CN201110118014XA patent/CN102252611B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6603563B1 (en) * | 2000-04-05 | 2003-08-05 | Accu-Sort Systems, Inc. | Apparatus for determining measurements of an object utilizing negative imaging |
EP1519144A1 (en) * | 2003-09-29 | 2005-03-30 | Nederlandse Organisatie voor toegepast-natuurwetenschappelijk onderzoek TNO | Free-form optical surface measuring apparatus and method |
CN101144703A (en) * | 2007-10-15 | 2008-03-19 | 陕西科技大学 | Article geometrical size measuring device and method based on multi-source image fusion |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103217431A (en) * | 2012-01-19 | 2013-07-24 | 昆山思拓机器有限公司 | Image detection method of SMT screen plate |
CN102607421A (en) * | 2012-04-12 | 2012-07-25 | 苏州天准精密技术有限公司 | Large-view-field image measuring method and equipment |
CN105744824A (en) * | 2014-12-10 | 2016-07-06 | 安徽海创自动控制设备有限公司 | Automatic paster mounting device |
CN107250715B (en) * | 2015-03-27 | 2019-09-17 | 三菱电机株式会社 | Detection device |
CN107250715A (en) * | 2015-03-27 | 2017-10-13 | 三菱电机株式会社 | Detection means |
CN108458655A (en) * | 2017-02-22 | 2018-08-28 | 上海理工大学 | Support the data configurableization monitoring system and method for vision measurement |
CN107578431A (en) * | 2017-07-31 | 2018-01-12 | 深圳市海思科自动化技术有限公司 | A kind of Mark points visual identity method |
CN107944494A (en) * | 2017-11-30 | 2018-04-20 | 华中科技大学 | A kind of Mark point visual identitys and localization method based on not bending moment |
CN107944494B (en) * | 2017-11-30 | 2020-09-18 | 华中科技大学 | Mark point visual identification and positioning method based on invariant moment |
CN109631803A (en) * | 2019-02-01 | 2019-04-16 | 东莞中科蓝海智能视觉科技有限公司 | Part angle detects adjusting method |
CN109631763A (en) * | 2019-02-01 | 2019-04-16 | 东莞中科蓝海智能视觉科技有限公司 | Irregular part detects localization method |
CN109919199A (en) * | 2019-02-13 | 2019-06-21 | 东南大学 | The detection method of Wind turbines abnormal data based on image procossing |
CN112579810A (en) * | 2019-09-30 | 2021-03-30 | 深圳市嘉立创科技发展有限公司 | Printed circuit board classification method and device, computer equipment and storage medium |
CN112579810B (en) * | 2019-09-30 | 2023-10-27 | 深圳市嘉立创科技发展有限公司 | Printed circuit board classification method, device, computer equipment and storage medium |
CN110826627A (en) * | 2019-11-06 | 2020-02-21 | 广东三维家信息科技有限公司 | Image similarity measuring method and device and electronic equipment |
CN111430289A (en) * | 2020-05-07 | 2020-07-17 | 上海果纳半导体技术有限公司 | Wafer positioning and calibrating device and wafer positioning and calibrating method |
Also Published As
Publication number | Publication date |
---|---|
CN102252611B (en) | 2013-11-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102252611B (en) | Geometric positioning method | |
CN105784716B (en) | Friction plate Quality Inspection System | |
US20120197427A1 (en) | Apparatus For Cutting Amd/or Etching Articles Comprising A Flat Surface On Which Designs And/or Writings Are Reproduced And A method For Actuating The Apparatus | |
CN106935683B (en) | A kind of positioning of solar battery sheet SPEED VISION and correction system and its method | |
CN107588723B (en) | Circular mark leak source detection method on a kind of High-speed target based on two-step method | |
CN108489394A (en) | A kind of large-scale sheet metal works almost T-stable automatic detection device and method | |
CN105841641A (en) | Laser triangulation method-based three-dimensional measuring instrument and flatness detection method | |
CN111681222A (en) | Automatic detection and judgment method for edge breakage of tool mark and scribing machine applied by same | |
CN113222982A (en) | Wafer surface defect detection method and system based on improved YOLO network | |
CN107891012B (en) | Pearl size and circularity sorting device based on equivalent algorithm | |
CN107101576A (en) | A kind of part method for comprehensive detection and system | |
CN112614105A (en) | Depth network-based 3D point cloud welding spot defect detection method | |
CN107895166B (en) | Method for realizing target robust recognition based on feature descriptor by geometric hash method | |
CN208042989U (en) | A kind of large-scale sheet metal works almost T-stable automatic detection device | |
CN106503735B (en) | The recognition methods of wiring board item number and device | |
CN114549412A (en) | Method, device and system for detecting quality of workpiece | |
CN102419157A (en) | Micro-depth-dimension automatic image measuring system | |
CN113109364B (en) | Method and device for detecting chip defects | |
CN114066893A (en) | Method, device and system for detecting quality of workpiece | |
CN110097042A (en) | A kind of the character machining system and detection method of the shaver shell with radian | |
CN107990825B (en) | High-precision position measuring device and method based on priori data correction | |
CN206075118U (en) | A kind of control system for auto parts and components processing | |
CN110930359A (en) | Method and system for detecting automobile shifting fork | |
CN115791801A (en) | 3D glass on-line monitoring platform based on machine vision | |
CN206864487U (en) | A kind of solar battery sheet SPEED VISION positioning and correction system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
C17 | Cessation of patent right | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131127 Termination date: 20140509 |