CN107609473A - A kind of 3D visual identifying systems and its recognition methods - Google Patents
A kind of 3D visual identifying systems and its recognition methods Download PDFInfo
- Publication number
- CN107609473A CN107609473A CN201710662010.5A CN201710662010A CN107609473A CN 107609473 A CN107609473 A CN 107609473A CN 201710662010 A CN201710662010 A CN 201710662010A CN 107609473 A CN107609473 A CN 107609473A
- Authority
- CN
- China
- Prior art keywords
- component
- identified
- posture
- detection site
- visual identifying
- 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.)
- Pending
Links
Landscapes
- Length Measuring Devices By Optical Means (AREA)
Abstract
The present invention relates to a kind of 3D visual identifying systems, the method being identified using the system, and the purposes that workpiece is identified in pipeline automation production using the system.By passing through 3D outline identifications and pose adjustment successively to the identified object such as workpiece, so as to obtain the image information with some detection sites corresponding to workpiece, high-speed camera is recycled to shoot each detection site successively, the length between corresponding detection site is obtained, as treats the reference information of quality inspection.The present invention can greatly improve the automatization level of work piece production, avoid artificial quality inspection efficiency is low and human error, while be with good expansibility, can easily be configured, maintenance and upgrade.
Description
Technical field
The present invention relates to a kind of 3D visual identifying systems, the method being identified using the system, and use the system
The purposes that workpiece is identified in pipeline automation production.
Background technology
Production line balance drastically increases industrial production efficiency, helps to realize the standardization of product, is wide at present
For the mode of production of use.Traditional production line balance is a kind of " passive type " production, is passed workpiece production line machinery
Send, the link such as each processing, inspection, packaging is realized by the worker for being distributed in streamline both sides.With the hair of automatic technology
Existing, increasing streamline introduces all kinds of automation mechanized operations and judgment technology, so as to reduce artificial input, ensure product matter
Amount.
As the basis of pipeline automation, the identification to workpiece is the step of starting one of automatic flow, and has much difficulty
The step of spending.3D Visual identification technologies, also known as Machine Vision Recognition Technology, base is established for the industrial applications of the step
Plinth.The general principle of 3D Visual identification technologies is to be handled, obtained by gathering the image of workpiece, and by certain algorithm
The physical messages such as workpiece profile, appearance, texture, form are obtained, and these physical messages are digitized, modelling, so as to be counted
The data signal that calculation machine can be identified and handled so that the equipment such as manipulator can identify and locate according to these data signals
Science and engineering part, so as to reach the purpose of automated production (CN106251353A, CN106003093A, CN105021124A).
A variety of methods for realizing 3D visual identitys are disclosed in the prior art, such as axis of workpiece is obtained by dual camera
Feature (Wei Xian etc., coal mine machinery, 2013,34 (8), 72-73), 3D workpiece surface appearances are obtained by optical grating projection
(CN102305601A) etc..One of research and development focus of prior art is exactly to be to provide all kinds of different 3D visual identifying systems;
But such technology is often keen to provide new identification design or mathematical modeling, without how deeply considering by 3D visions
The result of identification is applied in specific industrial production.On the other hand, prior art also puts forth effort to improve the precision of 3D visual identitys
(CN104915957A);But part involved in actual industrial production has typical 3D shape characteristics more, with complexity
Equipment realizes that high-precision 3D identifications might not meet industrial production needs.
In actual industrial production, for the workpiece through production line balance or transmission, often there is higher size requirement;In order to
Manufacture draws the product met the size requirements, it is necessary to manually carry out quality inspection.On the one hand, artificial participation reduces oneself of streamline
Dynamicization degree;On the other hand, artificial quality inspection can not possibly realize the detection one by one for whole workpiece, often bring what yield rate reduced
Problem.
The content of the invention
It is an object of the present invention to solve problems of the prior art, with realize automation for workpiece and
The high size detection of the degree of accuracy;Therefore, traditional 3D Visual identification technologies and automation size detection technology are combined, carry
For a kind of 3D visual identifying systems, and use system progress workpiece identification and the method for size detection.
To solve aforementioned technical problem, the present invention proposes following 3D visual identifying systems, and the system includes following
Component:
(1) 3D structures structure component, the 3D structural elements system are used for the 3D structural informations for obtaining identified object, and
The 3D structural informations are transmitted to computation module in digitized form;
(2) posture corrects component, the identified object that the posture correcting system provides according to 3D structural elements systems
3D structural informations, change the position and attitude of identified object, comply with default standardization posture;
(3) camera assembly, the camera assembly at a slow speed are used to detect identified object in the position of transit area at a slow speed, and
Positional information is sent to computation module in digitized form;
(4) high-speed camera component, one group of detection site of the identified object of high-speed camera component positioning, and be somebody's turn to do
The time interval that group detection site is taken;
(5) computation module, the computation module are used for the calculating of each component input signals in the 3D visual identifying systems,
And technical result is inputted;
(6) identified object-order is delivered to each detection means by transfer assembly, the transfer assembly.
The 3D structures structure component is not particularly limited, as long as disclosed in the prior art can be to being identified component
The equipment that is identified of 3D information and/or method can be used.
For example, the 3D structures structure component includes dual camera, the edge of the identified object of dual camera collection
Information, and the marginal information is sent to the computation module, known by calculating identified object axis feature with obtaining
The attitude information of other component.Specifically, the solution equation of the practise physiognomy centre of form of the image on three dimension scale is
Radii sequence Radium (i) (i=1,2,3 ...) then is tried to achieve, and the correspondence for calculating h values both sides semidiameter is absolute
Value, finally try to achieve the feature axis of identified object.
For another example the 3D structures structure component includes Grid Projection device and image collecting device, the Grid Projection
Device is collected warp mesh image by image collecting device and is delivered to described to identified subject surface projection gridding image
Computation module, the attitude information of identified component is calculated.
The posture, which corrects component, includes a manipulator, and the manipulator is acted on the identified component, leads to
Cross and the operation such as rotated, overturn, move to identified component, to cause the posture for being identified component to meet default standardization
Posture.The standardization posture is to refer to make identified component have clear readily identified detection site in subsequent operation
Posture.
The camera assembly at a slow speed can use common camera, and its capture rate is not less than 24fps, its role is to
Identified position of the component on conveyer of detection, when identified component is transferred into the shooting model of the high-speed camera component
When enclosing, for sending signal to start high-speed camera component.
The high-speed camera component uses high speed camera, its capture rate at least should be 100fps, preferably 200fps with
Upper, further preferred 240fps.The high speed camera is used to shoot identified component successively, particularly shoots one group
Each detection site in detection site, it is corresponding for calculating so as to obtain the time interval that corresponding detection site is shot successively
Distance between detection site.
The transfer assembly is used to identified component carrying out physical transport;For calculate foregoing corresponding detection site it
Between distance purpose, the transfer assembly should cause identified component through at the uniform velocity transmitting.In order to facilitate the shooting of camera assembly, respectively
There should be certain intervals between transfer assembly.
The computation module is microcomputer system, SCM system or other any executable required mathematical computations and signal
The system of input and output, its function of performing include:
I. the digital information of 3D structure component feedbacks is handled, and exports the 3D structural informations of identified object;
Ii. the 3D structural informations input posture of identified object is corrected into component, and controls posture to correct component regulation quilt
The position and attitude of identification object;
Iii. the photographing information of the analysis camera assembly at a slow speed, to judge to be identified whether component is taken the photograph positioned at the high speed
As the coverage of component;
Iv. the shooting for analyzing the high-speed camera component is newly played, and detection site quilt is corresponded to obtain in one group of detection site
The time interval of shooting;
V. according to the time interval, the length between the corresponding detection site is calculated.The present invention, which also provides one kind, to be made
The method being identified with the 3D visual identifying systems, the flow of methods described are:
(1) build component using the 3D structures identified component is identified, obtain the profile letter of identified component
Breath;
(2) component is corrected using the posture, its posture is judged according to the profile information of identified component, to the posture
Corrected, identified component is met default standardization posture, and determine one group of detecting position on identified component
The position of point;
(3) camera assembly is continuously shot to the identified component in standardization posture at a slow speed described in using, and is judged
Whether the identified component is transferred into the coverage of the high-speed camera component;
(4) when the identified component in standardization posture is transferred into the coverage of the high-speed camera component
When interior, picture shooting assembly is shot high described in startup, the identified component is continuously taken pictures, obtains one group of inspection of identified component
The time interval that detection site is taken is corresponded in location point;
(5) according to the time interval and the transfer rate of the transfer assembly, the length for corresponding to detection site is calculated
Degree.Calculation formula is:
L=(tB-tA) × v=Δs t × v
Wherein tA、tBFor the timestamp for the detection site being taken one after another, both differences are shooting time interval;V is
The rate travel of identified component;L is the distance of detection site A, B, as needs the size of quality inspection.
Further, 3D visual identifying systems of the present invention are applied to know workpiece in pipeline automation production
Not, especially for the identification and detection of workpiece size.
Beneficial effects of the present invention are:
1. device therefor structure of the present invention is simple, required component can use commercially available component, should not extra purchase or production
The component of specific use, equipment interoperability is strong, construction cost is low.
2. the automatization level of production line balance can be greatly improved in the present invention, by the way that 3D Visual identification technologies are applied to
In automatic production line, the identification and quality inspection of product/raw material workpiece can be realized automatically, particularly judge that workpiece size precision is
It is no to meet to require, manual intervention degree is significantly reduced, improves the yield rate of product.
, can be according to 3D Visual identification technologies and automatic 3. the present invention has stronger scalability and scalable type simultaneously
Change the future development for the treatment of technology, flexible expanded function or raising performance.
Brief description of the drawings
Clearly to show technical scheme, the required accompanying drawing used in embodiment will be made simply below
Introduce.
Fig. 1 is the method being identified using 3D visual identifying systems of the present invention.
Fig. 2 is workpiece used in embodiment 1, it is necessary to which the size of detection is the length L at the long end of workpiece, as corresponding detection site
Point A and point B composition one group of detection site.Dotted line represents workpiece rotation axes of symmetry in figure, and arrow represents workpiece motion s direction.
Embodiment
The specific function mode of the present invention is discussed in detail below by specific embodiment.
Workpiece as shown in Figure 2 is an annular workpieces.It has axially distributed point A and point B, distance between two points L are annular
The ring width of workpiece, as need the size of quality inspection.In order to realize the Automated inspection to size L, below scheme is taken.
The workpiece is transported on a moving belt, when reaching at the 3D structures structure component using dual camera, is taken the photograph by double
As the 3D topographical informations of head extraction workpiece, its profile information and attitude information are calculated.
Subsequently determine whether the posture of workpiece meets standardization posture, including its symmetry axis sensing, horizontality etc., and lead to
Cross manipulator crawl and do not conform to the workpiece of symbol, reapposed after adjusted posture on conveyer belt, finally making workpiece be in standardization
Posture, it is easy to follow-up to gather and handle.
Next judge whether workpiece enters the camera site of high-speed camera using slow-motion camera, specially transmitting
Take and a virtual start line is set, when workpiece passes through start line completely, start high-speed camera and carry out subsequent treatment.
When workpiece enters camera site, high-speed camera is taken pictures to the detection site of workpiece, i.e. point A and point B successively,
And obtain the interval time t that takes pictures.
So far, according to the speed v of conveyer belt conveying workpieces, L can be precisely calculated, formula is:
L=(tB-tA) × v=Δs t × v.
Further, the workpiece of specific length is unsatisfactory for for L, is judged as substandard product, can be then by from flowing water
Removed on line.
Described above is only the exemplary embodiment of the present invention, it is noted that for the common skill of the art
For art personnel, on the premise of its general principles are not departed from, some improvements and modifications can also be made, these improve and
Retouching also should be regarded as protection scope of the present invention.
Claims (9)
1. a kind of 3D visual identifying systems, it is characterised in that the system is included with lower component:
(1) 3D structures structure component, the 3D structural elements system are used to obtaining the 3D structural informations of identified object, and by institute
3D structural informations are stated to transmit in digitized form to computation module;
(2) posture corrects component, the 3D knots for the identified object that the posture correcting system provides according to 3D structural elements systems
Structure information, change the position and attitude of identified object, comply with default standardization posture;
(3) camera assembly at a slow speed, the camera assembly at a slow speed are used to detect identified object in the position of transit area, and by position
Confidence breath is sent to computation module in digitized form;
(4) high-speed camera component, one group of detection site of the identified object of high-speed camera component positioning, and obtain group inspection
The time interval that location point is taken;
(5) computation module, the computation module are used for the calculating of each component input signals in the 3D visual identifying systems, and will
Technical result inputs;
(6) identified object-order is delivered to each detection means by transfer assembly, the transfer assembly.
2. 3D visual identifying systems according to claim 1, it is characterised in that the 3D structures structure component includes double shootings
Head, the marginal information of the identified object of dual camera collection, and the marginal information is sent to the computation module, lead to
The identified object axis feature of calculating is crossed to obtain the attitude information of identified component.
3. 3D visual identifying systems according to claim 1, it is characterised in that the 3D structures structure component is thrown including grid
Image device and image collecting device, the optical grating projection device are adopted to identified subject surface projection gridding image by image
Acquisition means collect warp mesh image and are delivered to the computation module, and the attitude information of identified component is calculated.
4. 3D visual identifying systems according to claim 1, it is characterised in that the posture, which corrects component, includes a machinery
Hand, the manipulator are acted on the identified component, to cause the posture of identified component to meet default standardization
Posture.
5. 3D visual identifying systems according to claim 1, it is characterised in that the high-speed camera component shoots one group successively
Each detection site in detection site, and shooting signal is delivered to the computation module, so as to obtain in one group of detection site
The time interval that corresponding detection site is taken.
6. 3D visual identifying systems according to claim 1, it is characterised in that the function that the computation module performs includes:
I. the digital information of 3D structure component feedbacks is handled, and exports the 3D structural informations of identified object;
Ii. the 3D structural informations input posture of identified object is corrected into component, and it is identified to control posture to correct component regulation
The position and attitude of object;
Iii. the photographing information of the analysis camera assembly at a slow speed, to judge to be identified whether component is located at the high-speed camera group
The coverage of part;
Iv. the shooting for analyzing the high-speed camera component is newly played, and is taken with corresponding to detection site in one group of detection site of acquisition
Time interval;
V. according to the time interval, the length between the corresponding detection site is calculated.
7. 3D visual identifying systems according to claim 1, it is characterised in that the transfer assembly is used for identified component
It is sequentially transmitted, there is interval between the identified component, it is described to be transmitted as at the uniform velocity.
8. the application of 3D visual identifying systems according to claim 1, it is characterised in that the 3D visual identifying systems are used for
Workpiece is identified in pipeline automation production.
9. the method that 3D visual identifying systems are identified according to claim 1, it is characterised in that the flow of methods described
For:
(1) build component using the 3D structures identified component is identified, obtain the profile information of identified component;
(2) component is corrected using the posture, its posture is judged according to the profile information of identified component, the posture is carried out
Correct, identified component is met default standardization posture, and determine one group of detection site on identified component
Position;
(3) camera assembly is continuously shot to the identified component in standardization posture at a slow speed described in using, described in judgement
Whether identified component is transferred into the coverage of the high-speed camera component;
(4) when the identified component in standardization posture is transferred into the coverage of the high-speed camera component
When, picture shooting assembly is shot high described in startup, the identified component is continuously taken pictures, and obtains one group of detection of identified component
The time interval that detection site is taken is corresponded in site;
(5) according to the time interval and the transfer rate of the transfer assembly, the length for corresponding to detection site is calculated.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710662010.5A CN107609473A (en) | 2017-08-04 | 2017-08-04 | A kind of 3D visual identifying systems and its recognition methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710662010.5A CN107609473A (en) | 2017-08-04 | 2017-08-04 | A kind of 3D visual identifying systems and its recognition methods |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107609473A true CN107609473A (en) | 2018-01-19 |
Family
ID=61064580
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710662010.5A Pending CN107609473A (en) | 2017-08-04 | 2017-08-04 | A kind of 3D visual identifying systems and its recognition methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107609473A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108764105A (en) * | 2018-05-22 | 2018-11-06 | 广东美的智能机器人有限公司 | Assembly line workpiece identification method, server and terminal |
CN109719229A (en) * | 2019-01-01 | 2019-05-07 | 周义 | The car door limiter skeleton of view-based access control model detection reverses control system |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150036875A1 (en) * | 2013-07-31 | 2015-02-05 | Nvidia Corporation | Method and system for application execution based on object recognition for mobile devices |
CN106552772A (en) * | 2016-10-25 | 2017-04-05 | 北京京东尚科信息技术有限公司 | Visual identifying system and the classification sorting system using the visual identifying system |
CN106680286A (en) * | 2016-12-20 | 2017-05-17 | 深圳信息职业技术学院 | Machine vision-based laser processing system and machine vision-based laser processing method |
-
2017
- 2017-08-04 CN CN201710662010.5A patent/CN107609473A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150036875A1 (en) * | 2013-07-31 | 2015-02-05 | Nvidia Corporation | Method and system for application execution based on object recognition for mobile devices |
CN106552772A (en) * | 2016-10-25 | 2017-04-05 | 北京京东尚科信息技术有限公司 | Visual identifying system and the classification sorting system using the visual identifying system |
CN106680286A (en) * | 2016-12-20 | 2017-05-17 | 深圳信息职业技术学院 | Machine vision-based laser processing system and machine vision-based laser processing method |
Non-Patent Citations (1)
Title |
---|
WU526515597: ""手机外壳尺寸检测设备"", 《HTTPS://WENKU.BAIDU.COM/VIEW/F3CFFAB5BA0D4A7302763A67.HTML》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108764105A (en) * | 2018-05-22 | 2018-11-06 | 广东美的智能机器人有限公司 | Assembly line workpiece identification method, server and terminal |
CN108764105B (en) * | 2018-05-22 | 2022-03-22 | 广东美的智能机器人有限公司 | Assembly line workpiece identification method, server and terminal |
CN109719229A (en) * | 2019-01-01 | 2019-05-07 | 周义 | The car door limiter skeleton of view-based access control model detection reverses control system |
CN109719229B (en) * | 2019-01-01 | 2020-10-30 | 新昌县雷涛机械有限公司 | Car door limiter framework torsion control system based on visual detection |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11276194B2 (en) | Learning dataset creation method and device | |
CN104476552B (en) | A kind of method for carrying of robot based on machine vision section bar Handling device | |
CN106041937B (en) | A kind of control method of the manipulator crawl control system based on binocular stereo vision | |
US10442633B2 (en) | Robot system | |
CN104511436B (en) | A kind of express delivery method for sorting based on Robot Visual Servoing technology and system | |
CN107671008A (en) | A kind of part stream waterline automatic sorting boxing apparatus of view-based access control model | |
CN110580725A (en) | Box sorting method and system based on RGB-D camera | |
CN106824816B (en) | A kind of detection of PE bottle and method for sorting based on machine vision | |
CN106583268A (en) | PE bottle inspecting and sorting system based on machine vision | |
CN110302981B (en) | Solid waste sorting online grabbing method and system | |
CN104589357B (en) | The DELTA robot control system of view-based access control model tracking and method | |
CN105225225B (en) | A kind of leather system for automatic marker making method and apparatus based on machine vision | |
CN101872423A (en) | Method for tracking moving object on production line | |
CN103325106A (en) | Moving workpiece sorting method based on LabVIEW | |
CN106113067A (en) | A kind of Dual-Arm Mobile Robot system based on binocular vision | |
CN109623821A (en) | The visual guide method of mechanical hand crawl article | |
CN107609473A (en) | A kind of 3D visual identifying systems and its recognition methods | |
CN108355979A (en) | Target tracking sorting system on conveyer belt | |
CN105588440A (en) | Position detection method and device for sintering machine trolleys | |
CN106733686B (en) | A kind of assembly line object positioning method of view-based access control model and code-disc data fusion | |
CN103707300A (en) | Manipulator device | |
CN208067234U (en) | A kind of automatic riveting system based on machine vision | |
JP2019018339A (en) | Robot system | |
CN104615971B (en) | A kind of Quick Response Code Tracking Recognition method for chemical fertilizer quality tracing system | |
CN210233036U (en) | Binocular artificial intelligence arm teaching 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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180119 |
|
WD01 | Invention patent application deemed withdrawn after publication |