CN107271437B - One kind being based on twin-channel retainer front and back sides image discriminating method - Google Patents
One kind being based on twin-channel retainer front and back sides image discriminating method Download PDFInfo
- Publication number
- CN107271437B CN107271437B CN201710260334.6A CN201710260334A CN107271437B CN 107271437 B CN107271437 B CN 107271437B CN 201710260334 A CN201710260334 A CN 201710260334A CN 107271437 B CN107271437 B CN 107271437B
- Authority
- CN
- China
- Prior art keywords
- picture
- channel
- led light
- retainer
- group
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
Abstract
The invention discloses one kind to be based on twin-channel retainer front and back sides image discriminating method, illumination is provided using two groups of LED light, image taking is carried out using two groups of channels of camera, when shooting, it is separately turned on two groups of LED light and shoots two groups of pictures, the picture of acquisition is merged and is compared, is compared by public domain and standard value, judges the front and back sides of the end face on retainer.The interference that the present invention effectively prevent image speck caused by workpiece and tooling surface reflection to identify feature, while requirement of the image recognition to light source is also effectively reduced, and reduce influence of the scattering of light to picture quality, to effectively prevent from judging by accident.
Description
Technical field
Machine components detection device technical field of the present invention, more specifically, more particularly to a kind of based on twin-channel guarantor
Hold frame front and back sides image discriminating method.
Background technique
Retainer partly wraps up all or part of rolling element, and moves therewith, to separation roller, guides rolling element
And it holds it in bearing.There is the retainer in the equipment of higher rotation speed for automotive trannsmission system etc., for size essence
The requirement of degree is very high.In process of production if dimensional accuracy is not up to standard, the face of being installed on vehicle will lead to keep retainer
The inordinate wear of frame, the problems such as influencing cage service life or traffic safety.Retainer mentioned by the present invention refer to by
The surfaces externally and internally formed after turning is spherical surface, and both ends are the metal component of plane, one in existing retainer processing technology
As be by manually differentiating front and back sides by mechanical measurement to retainer, then guarantee to carry out down one manufacturing procedure when
It waits, these retainers can enter in process equipment according to the same direction.It is produced in practice large batch of, on the one hand, by
In heavy workload, labor intensity is very high, and it is positive and negative to cut retainer it is possible that the case where failing to judge or differentiating fault for artificial cognition
Face difference is not that cannot accurately determine the front and back sides of retainer it is obvious that if the precision of measuring tool is not high enough,
The accurate differentiation rate of each the retainer front and back sides differentiated is not can guarantee exactly.
Number of patent application is that the Chinese invention of CN2015102506173 discloses a kind of retainer front and back sides discriminating gear,
The width that retainer center hole chamfering is shot by camera, when camera central axis is horizontal by 45 degree of angles,
The fillet surface of retainer center hole is just vertical with camera, and the chamfering width of camera shooting at this time is developed width,
Can determine whether the front and back sides of retainer compared with benchmark chamfering width according to the chamfering width taken at this time, identification effect is high,
Reduce the problem that artificial cognition error is big, identification effect is low.But in actual pilot production, it has been found that the program is deposited
In following problem:
1) due to the size for being inner chamfer of the positive and negative region feature of retainer, it is necessary to realize that chamfer image is clapped by 45 degree of light sources
Take the photograph realization image procossing;But chamfering is shot to the more demanding of light source, the retainer smaller for some chamferings, shooting
Image annulus is very narrow, it is easy to the case where judging by accident;
2) retainer has many greasy dirt stations on the surface of retainer during production process and automatic charging
And at the chamfering of retainer, there is unsharp situation so as to cause the image shot, it is serious to result even in annulus
Image fracture, so as to cause the situation of image recognition mistake;
3) require have lower cost when the production link due to equipment application, retainer is made up of cylinder and pneumatic clamper
Gripper crawl be put into camera in the following, a manipulator can adapt to the holding of different-diameter as much as possible again simultaneously
Frame, therefore it is easy to cause the positioning accuracy of retainer low, retainer, which is likely to occur angular deflection after being crawled, causes light source to reflect
The unstable situation of intensity, to influence the quality of image;
4) retainer may be such that the center of circle of outer circle and inner circle is not exclusively overlapped due to processing itself, only take in annulus
One section carry out know method for distinguishing cannot use, it is necessary to whole image is identified;In view of the cost factor of camera, in reality
The very high camera of resolution ratio can not be taken in the production process of border, therefore also brings difficulty to image recognition.
Simultaneously as whole system is very high to the reliability requirement of image recognition, absolutely not go out in permission production process
The case where now judging by accident, it is necessary to reduce the quantity of suspicious product, therefore the equipment can not well adapt to actual production process.
Summary of the invention
It is an object of the invention to solve the problem above-mentioned, proposing one kind can be avoided erroneous judgement, and it is positive and negative to improve retainer
Face discrimination precision based on twin-channel retainer front and back sides image discriminating method.
The present invention is through the following technical solutions to achieve the above objectives: a kind of be based on twin-channel retainer front and back sides image
Method of discrimination provides illumination using two groups of LED light, carries out image taking using camera, two groups of LED light are respectively end face identification
LED light and chamfering identify that LED light, two groups of LED light are annularly arranged, end face identification LED light is arranged in chamfering identification LED light
Surface, the surface of end face identification LED light is arranged in camera, and the axial line of camera and end face identify LED light and fall
The axial line of angle identification LED light is located on the same line;End face identification LED light includes the first lampshade of annular, setting the
First point light source of the ring shape on one lampshade, the non-reflective curtain of black that the first lampshade bottom is arranged in and setting are the
The filter one point light source outside and be fixed on the first lampshade;The chamfering identification LED light includes the second lampshade of annular, sets
The second point light source of the ring shape on the second lampshade is set, is arranged in inside lampshade and by the second point inside the second lampshade
The light that light source issues reflects the reflective side on 45 degree of angle vertical irradiations to retainer chamfering;
When shooting, end face identification LED light is first turned on, chamfering is closed and identifies LED light, control camera shoots first group
Picture, and first group of picture is handled using computer, it only extracts one of RGB channel channel and generates the second group picture
Piece;End face identification LED light is turned off, chamfering is opened and identifies LED light, control camera shoots third group picture, and utilizes meter
Calculation machine handles third width picture, and another channel extracted in RGB channel generates the 4th group of picture;In a computer will
Second group of different picture of channel and the 4th group of picture make lamination process, using in the picture after overlapping have twin-channel part
It makes comparisons with the normal pictures of the standard component shot in advance, the end face for judging that retainer is now placed in top is positive or anti-
Face.
Further, shoot first group of picture, if computer to first group of picture extract be the channel R image, count
Calculation machine can extract the image of the channel G or channel B to third group picture;If what computer extracted first group of picture is the channel G
Image, then computer can extract the image of the channel R or channel B to third group picture;What first group of picture extracted is channel B
Image, then computer can extract the image in the channel G or the channel R to third group picture.
Further, the standard picture of the standard component shot in advance had both included the standard shot when retainer is face-up
Image include again retainer reverse side upward when the standard picture shot, the picture after overlapping carries out with two width standard pictures simultaneously
Comparison, if the picture after overlapping has the outer profile of any one pair in the outer profile and two width standard pictures of binary channels part simultaneously
Error is in ± n pixel, then the standard picture according to error in ± n pixel judges the front and back sides of retainer;
If it is more than ± n that the picture after overlapping, which has the outer profile error in the outer profile and two width standard pictures of binary channels part simultaneously,
A pixel, without judging result, and then carry out retainer itself whether He Ge detection.
The beneficial effects of the present invention are:
1, the present invention shoots two groups of pictures of end face and chamfering using one group of camera respectively, and takes two groups of pictures respectively
Middle different channel is carrying out lamination process, is made comparisons using treated picture with standard picture, not only when shooting mutually not
It can have an impact, and due to comparing two factors simultaneously, can largely improve the accuracy of front and back sides judgement.
2, the present invention uses the picture after overlapping to have in the outer profile and two width standard pictures of binary channels part simultaneously and appoints
Whether the secondary outer profile of meaning one is used as judgment method in ± n pixel, which relies on single image with conventional
The mode that feature is identified is significantly improved, and single picture can be effectively prevent to make due to light source or environmental disturbances factor
At identification mistake so as to cause seriously endangering the case where.
3, the present invention reduces hair in the first lampshade bottom setting non-reflective curtain of black of end face identification LED light as far as possible
Penetrate light, prevent chamfering identification LED light light shine directly into end face identification LED light lampshade bottom reflected, prevent with
The reflective accuracy for leading to chamfering picture shooting in retainer end face caused by this.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of the used device of the present invention.
Fig. 2 is the structural schematic diagram of identification LED light in end face of the present invention.
Fig. 3 is the structural schematic diagram of chamfering identification LED light of the present invention.
In figure, 1- retainer, 2- chamfering identification LED light, the end face 3- identify LED light, 4- camera, the first lampshade of 5-, 6-
First point light source, the non-reflective curtain of 7- black, 8- filter, the second lampshade of 9-, the second point light source of 10-, the reflective side 11-.
Specific embodiment
The present invention will be further explained below with reference to the attached drawings:
As shown in Fig. 1 ~ 3, one kind being based on twin-channel retainer front and back sides image discriminating method, is mentioned using two groups of LED light
For illumination, image taking is carried out using camera 4, two groups of LED light are respectively that end face identification LED light 3 and chamfering identify LED light 2,
Two groups of LED light annularly arrange that end face identification LED light 3 is arranged in the surface of chamfering identification LED light 2, and camera 4 is arranged
In the surface of end face identification LED light 3, and the axial line of camera 4 and end face identification LED light 3 and chamfering identify LED light 2
Axial line is located on the same line;End face identification LED light 3 includes the first lampshade 5 of annular, is arranged on the first lampshade 5
First point light source 6 of one ring shape, the non-reflective curtain 7 of black that 5 bottom of the first lampshade is arranged in and setting are in the first point light
The filter 8 source 6 outside and be fixed on the first lampshade 5;The chamfering identification LED light 2 includes the second lampshade 9 of annular, sets
The second point light source 10 of the ring shape on the second lampshade 9 is set, is arranged in inside lampshade and by inside the second lampshade 9
The reflective side 11 on light reflection 45 degree of angle vertical irradiations to 1 chamfering of retainer that two point light sources 10 issue.
The light that end face identification LED light 3 issues is light downwards annular in shape, by the anti-of 1 upper surface of retainer
It is shone directly on camera after penetrating, when arranging end face identification LED light 3, the setting of its 5 lower edge of the first lampshade will guarantee
After the light that identification LED light 3 issues in end face projects vertically, it will not shine directly into chamfering identification LED light 2.
The light that chamfering identifies that LED light 2 issues is oblique light ray annular in shape;Since the bottom of end face identification LED light 3 is
The non-reflective curtain 7 of black, even if so the light of chamfering identification LED light 2 is irradiated to black after the reflection of the chamfering of retainer
Non-reflective curtain 7 also can directly be absorbed by the non-reflective curtain 7 of black, without being reflected on the end face of retainer 1, to avoid
Influence to judging result.
When shooting, end face identification LED light 3 is first turned on, chamfering is closed and identifies LED light 2, control camera 4 shoots the
One group of picture, and first group of picture is handled using computer, it only extracts one of RGB channel channel and generates second group
Picture;End face identification LED light 3 is turned off, chamfering is opened and identifies LED light 2, control camera 4 shoots third group picture, and benefit
Third width picture is handled with computer, another channel extracted in RGB channel generates the 4th group of picture;In computer
Middle second group of picture that channel is different and the 4th group of picture make lamination process, twin-channel using having in the picture after overlapping
Part is made comparisons with the normal pictures of the standard component shot in advance, and the end face for judging that retainer is now placed in top is that front is gone back
It is reverse side.
Shoot first group of picture, if computer to first group of picture extract be the channel R image, computer is to third
Group picture can extract the image of the channel G or channel B;If computer to first group of picture extract be the channel G image, count
Calculation machine can extract the image of the channel R or channel B to third group picture;What first group of picture extracted is the image of channel B, then counts
Calculation machine can extract the image in the channel G or the channel R to third group picture.
The standard picture of the standard component shot in advance had not only included the standard picture shot when retainer is face-up but also had wrapped
The standard picture shot when including retainer reverse side upward, the picture after overlapping are compared with two width standard pictures simultaneously, if
There is picture after overlapping the outer profile error of any one pair in the outer profile and two width standard pictures of binary channels part to exist simultaneously
In ± n pixel, then the standard picture according to error in ± n pixel judges the front and back sides of retainer;If overlapping
It is more than ± n pixel that picture afterwards, which has the outer profile error in the outer profile and two width standard pictures of binary channels part simultaneously,
Point, without judging result, and then carry out retainer itself whether He Ge detection.
Above-described embodiment is presently preferred embodiments of the present invention, is not a limitation on the technical scheme of the present invention, as long as
Without the technical solution that creative work can be realized on the basis of the above embodiments, it is regarded as falling into the invention patent
Rights protection scope in.
Claims (3)
1. one kind is based on twin-channel retainer front and back sides image discriminating method, it is characterised in that: provide photograph using two groups of LED light
It is bright, image taking is carried out using camera (4), two groups of LED light are respectively end face identification LED light (3) and chamfering identification LED light
(2), two groups of LED light are annularly arranged, end face identification LED light (3) is arranged in the surface of chamfering identification LED light (2), camera shooting
Head (4) setting is in the surface of end face identification LED light (3), and the axial line of camera (4) and end face identify LED light (3) and fall
The axial line of angle identification LED light (2) is located on the same line;End face identification LED light (3) include annular the first lampshade (5),
The first point light source (6) for the ring shape being arranged on the first lampshade (5), the black being arranged in the first lampshade (5) bottom is not anti-
Light curtain cloth (7) and setting are in the filter (8) the first point light source (6) outside and be fixed on the first lampshade (5);The chamfering
Identification LED light (2) includes the second point light source of the second lampshade (9) of annular, the ring shape being arranged on the second lampshade (9)
(10), it is arranged inside lampshade and the light that internal the second point light source (10) of the second lampshade (9) issues is reflected into 45 degree of angles
Reflective side (11) on vertical irradiation to retainer (1) chamfering;
When shooting, end face identification LED light (3) is first turned on, closes chamfering identification LED light (2), control camera (4) is shot
First group of picture, and first group of picture is handled using computer, it only extracts one of RGB channel channel and generates second
Group picture;End face identification LED light (3) is turned off, chamfering identification LED light (2) is opened, third group is shot in control camera (4)
Picture, and third width picture is handled using computer, another channel extracted in RGB channel generates the 4th group picture
Piece;Second group of different picture of channel and the 4th group of picture are made into lamination process in a computer, using in the picture after overlapping
It makes comparisons with twin-channel part with the normal pictures of the standard component shot in advance, judges that retainer is now placed in top
End face is front or reverse side.
2. according to claim 1 a kind of based on twin-channel retainer front and back sides image discriminating method, it is characterised in that:
Shoot first group of picture, if computer to first group of picture extract be the channel R image, computer can to third group picture
To extract the image in the channel G or channel B;If computer to first group of picture extract be the channel G image, computer is to
Three groups of pictures can extract the image of the channel R or channel B;What first group of picture extracted is the image of channel B, then computer is to the
Three groups of pictures can extract the image in the channel G or the channel R.
3. according to claim 1 a kind of based on twin-channel retainer front and back sides image discriminating method, it is characterised in that:
The standard picture of the standard component shot in advance had not only included the standard picture shot when retainer is face-up but also had included retainer
Reverse side upward when the standard picture shot, the picture after overlapping is compared with two width standard pictures simultaneously, if after overlapping
Picture has the outer profile error of any one pair in the outer profile and two width standard pictures of binary channels part in ± n pixel simultaneously
In point, then the standard picture according to error in ± n pixel judges the front and back sides of retainer;If the picture after overlapping is same
When outer profile and two width standard pictures with binary channels part in outer profile error be more than ± n pixel, without
Judging result, so carry out retainer itself whether He Ge detection.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710260334.6A CN107271437B (en) | 2017-04-20 | 2017-04-20 | One kind being based on twin-channel retainer front and back sides image discriminating method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710260334.6A CN107271437B (en) | 2017-04-20 | 2017-04-20 | One kind being based on twin-channel retainer front and back sides image discriminating method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107271437A CN107271437A (en) | 2017-10-20 |
CN107271437B true CN107271437B (en) | 2019-07-09 |
Family
ID=60073968
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710260334.6A Active CN107271437B (en) | 2017-04-20 | 2017-04-20 | One kind being based on twin-channel retainer front and back sides image discriminating method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107271437B (en) |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5253986B2 (en) * | 2008-12-04 | 2013-07-31 | Juki株式会社 | Button recognition device and button recognition method |
DE102009058215B4 (en) * | 2009-12-15 | 2014-12-18 | Aktiebolaget Skf | Method and device for surface inspection of a bearing component |
CN203287330U (en) * | 2013-06-02 | 2013-11-13 | 杭州欧谱洛博自动化技术有限公司 | Full-angle vision comprehensive detection system for appearance of bearing roller |
CN203824524U (en) * | 2014-04-16 | 2014-09-10 | 杭州欧谱洛博自动化技术有限公司 | Detection device detecting roller chamfer and end face simultaneously |
CN104165653B (en) * | 2014-08-27 | 2016-06-08 | 洛阳久德轴承模具技术有限公司 | For gathering device and the acquisition method thereof of bearing roller end face and chamfering image |
CN104897672B (en) * | 2015-05-15 | 2017-08-25 | 浙江工业大学 | A kind of retainer positive and negative discriminating gear |
CN105092577B (en) * | 2015-05-15 | 2017-08-01 | 浙江工业大学 | A kind of retainer positive and negative discriminating gear |
-
2017
- 2017-04-20 CN CN201710260334.6A patent/CN107271437B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN107271437A (en) | 2017-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104764750B (en) | Elevator balanced compensated chain automatic detection device for quality and method based on machine vision | |
CA2151344C (en) | Lens inspection system and method | |
CN105092577B (en) | A kind of retainer positive and negative discriminating gear | |
CN107533012B (en) | Melt the surface defect inspection apparatus and surface defect inspection method of coated steel sheet | |
CN104897672B (en) | A kind of retainer positive and negative discriminating gear | |
CN102636490A (en) | Method for detecting surface defects of dustproof cover of bearing based on machine vision | |
CN106931881A (en) | A kind of polyphaser visible detection method | |
CN107255452B (en) | Double LED light formula retainer front and back sides image-recognizing method | |
CN105139384B (en) | The method and apparatus of defect capsule detection | |
WO2020103310A1 (en) | Mask defect detection device and method | |
CN112129783A (en) | Transparent medicine bottle bottom defect detection device and detection method based on deep learning | |
CN107247054A (en) | A kind of retainer positive and negative image discriminating method based on bi-chromatic illumination | |
CN107271437B (en) | One kind being based on twin-channel retainer front and back sides image discriminating method | |
CN107202677B (en) | Retainer front and back sides recognition methods based on the illumination of double LED light | |
CN107084986B (en) | Utilize the method for carrying out retainer front and back sides image recognition with operation | |
CN207248761U (en) | Utilize the device that retainer positive and negative image recognition is carried out with computing | |
CN107247053B (en) | A kind of image-recognizing method identifying retainer front and back sides | |
CN206725456U (en) | A kind of retainer positive and negative pattern recognition device | |
CN103718025A (en) | Container mouth portion inspection method and device | |
CN107290341B (en) | A kind of retainer front and back sides image-recognizing method | |
CN103499586B (en) | A kind of cillin bottle head detection device for lamp inspection machine | |
CN104362111B (en) | Silicon wafer edge breakage automatic detection method | |
CN207397296U (en) | A kind of object identification device based on figure viewed from behind imaging | |
CN102495077A (en) | Automatic detecting method of machine vision system for detecting flaws of infusion bottle | |
CN110470670A (en) | A kind of magnetic core defect detecting system and method based on image procossing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |