CN106157303A - A kind of method based on machine vision to Surface testing - Google Patents
A kind of method based on machine vision to Surface testing Download PDFInfo
- Publication number
- CN106157303A CN106157303A CN201610487518.1A CN201610487518A CN106157303A CN 106157303 A CN106157303 A CN 106157303A CN 201610487518 A CN201610487518 A CN 201610487518A CN 106157303 A CN106157303 A CN 106157303A
- Authority
- CN
- China
- Prior art keywords
- region
- threshold segmentation
- image
- noise
- machine vision
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of method based on machine vision to Surface testing, comprise the following steps: 1) use LED circular lamp direct dark field frontlighting mode to gather image；2) dynamic threshold segmentation method is used by the region of interesting extraction of scratch detection in region, surface out；3) again by using provincial characteristics, regional morphology to use erosion operation to remove miscellaneous point and little outthrust in cut zone in region, surface, it is ensured that there are enough precision in calculating；4) radiation conversion, image smoothing, connected region is finally used to extract scheduling algorithm and detect cut and show its result, this method detection surface defect, preferably reduce in image acquisition process, it is limited to the interference of ambient light, has isolated metal defect with becoming more meticulous and be all the edge being highlighted.
Description
Technical field
The present invention relates to field, particularly to a kind of method based on machine vision to Surface testing.
Background technology
In traditional process of producing product, generally the surface defects detection to product is to use manual detection
Method.Along with the development of science and technology, particularly the development of computer technology, occurs in that Computer Vision Detection Technique.
The system utilizing this new technique to design is not affected by adverse circumstances and subjective factors, can detect product quickly and accurately
The quality of product, completes the Detection task that manually cannot complete.Machine Vision Detection combines Computer Image Processing and pattern is known
Not theoretical, it combines computer technology, data structure, image procossing, the phase of the different field such as pattern recognition and soft project
Close knowledge.
During detecting metal surface, being limited to the restriction of shooting environmental, the metal surface cut shot exists
Black background region is shown as highlighted, and metal edge shows it is also highlighted, be difficult to careful accurately pass through with regard to current technology
Machine vision technique is accomplished strictly to distinguish.
Summary of the invention
The technical problem to be solved is to provide a kind of method based on machine vision to Surface testing, this method
Detection surface defect, preferably reduces in image acquisition process, is limited to the interference of ambient light；Can separate with becoming more meticulous most
Go out metal defect and be all the edge being highlighted, to solve the above-mentioned multinomial defect caused in prior art.
For achieving the above object, the present invention provides following technical scheme: a kind of based on machine vision to Surface testing
Method, comprises the following steps:
1) LED circular lamp direct dark field frontlighting mode is used to gather image；
2) dynamic threshold segmentation method is used by the region of interesting extraction of scratch detection in region, surface out；
3) again by using provincial characteristics, regional morphology to use erosion operation to remove in cut zone miscellaneous in region, surface
Point and little outthrust, it is ensured that in calculating, have enough precision；
4) radiation conversion, image smoothing, connected region is finally used to extract scheduling algorithm and detect cut and show its result.
Preferably, described step 2) in, the purpose of image threshold is according to gray level, collection of pixels carries out one and draws
Point, each subset obtained forms a region corresponding with real-world scene, has consistent attribute inside regional, and
Adjacent area layout has this consistent attribute, Threshold segmentation operation to be defined as
S={ (r, c) ∈ Rgmin≤fr, c≤gmax}；
Therefore, in gray value in image ROIR is in a certain appointment intensity value ranges by Threshold segmentation, all point chooses output
In the S of region, make gmin=0 or gmax=2b-1, if illumination can keep constant, threshold value gmin and gmax can be when system be arranged
Determined choosing and never with being adjusted. Threshold segmentation is divided into fixed threshold segmentation and dynamic threshold segmentation；
The operation that image and its local background are compared by dynamic threshold segmentation is referred to as dynamic threshold segmentation and processes, and uses
fr, c represents input picture, uses gr, c represents the image after smoothing, then processes as follows to the dynamic threshold segmentation of bright object
S={ (r, c) ∈ Rfr, c-gr, c >=gdiff}；
And the dynamic threshold segmentation process to dark object is S={ (r, c) ∈ Rfr, c-gr, c≤-gdiff}。
Preferably, described step 3) in, the algorithm of employing is a=R=∑ (r, c) ∈ R1=∑ ni-1cei-csi+1；
From above formula, the area a in region is exactly the R that counts in region.
Preferably, described step 4) in, it is to carry territory one of this point of the value of any in digital picture or Serial No.
In the Mesophyticum of each point value replace, allow the pixel value of surrounding close to actual value, thus eliminate isolated noise spot；
Method is the two-dimentional sleiding form of certain structure, pixel in plate is ranked up according to the size of pixel value, raw
Become monotone increasing (or decline) for 2-D data sequence, containing noise in the result of Threshold segmentation, this is not end product,
The process of noise, by using image smoothing to suppress.
Preferably, described step 4) in, testing result during removing noise, all connections being less than 4 pixels
Region is counted as noise and is removed, in order to distinguish noise and defect, it is assumed that noise is equally distributed, and belongs to a cut together
Defect be close to each other, therefore, it can medium and small for defect area gap be closed by expanding. in order to the company of calculating
Logical region, it is necessary to define suitable two pixels and should be considered to communicate with each other.
Use above technical scheme to provide the benefit that: the method detection surface defect of the present invention, preferably reduce figure
As in gatherer process, being limited to the interference of ambient light, isolate metal defect with becoming more meticulous and be highlighted with being all
Edge.Moreover, this method is simple to operation, and feasibility is the strongest.
Accompanying drawing explanation
Fig. 1 is the control block diagram of the present invention.
Detailed description of the invention
Describe the preferred embodiment of the present invention below in conjunction with the accompanying drawings in detail.
Fig. 1 shows the detailed description of the invention of the present invention: a kind of method based on machine vision to Surface testing, including following
Step:
1) LED circular lamp direct dark field frontlighting mode is used to gather image；
2) dynamic threshold segmentation method is used by the region of interesting extraction of scratch detection in region, surface out；
3) again by using provincial characteristics, regional morphology to use erosion operation to remove in cut zone miscellaneous in region, surface
Point and little outthrust, it is ensured that in calculating, have enough precision；
4) radiation conversion, image smoothing, connected region is finally used to extract scheduling algorithm and detect cut and show its result.
Described step 2) in, the purpose of image threshold is according to gray level, collection of pixels carries out a division, obtains
Each subset form a region corresponding with real-world scene, there is consistent attribute, and adjacent region inside regional
Territory layout has this consistent attribute, Threshold segmentation operation to be defined as
S={ (r, c) ∈ Rgmin≤fr, c≤gmax}；
Therefore, in gray value in image ROIR is in a certain appointment intensity value ranges by Threshold segmentation, all point chooses output
In the S of region, make gmin=0 or gmax=2b-1, if illumination can keep constant, threshold value gmin and gmax can be when system be arranged
Determined choosing and never with being adjusted；Threshold segmentation is divided into fixed threshold segmentation and dynamic threshold segmentation；
The operation that image and its local background are compared by dynamic threshold segmentation is referred to as dynamic threshold segmentation and processes, and uses
fr, c represents input picture, uses gr, c represents the image after smoothing, then processes as follows to the dynamic threshold segmentation of bright object
S={ (r, c) ∈ Rfr, c-gr, c >=gdiff}；
And the dynamic threshold segmentation process to dark object is S={ (r, c) ∈ Rfr, c-gr, c≤-gdiff}。
Described step 3) in, the algorithm of employing is a=R=∑ (r, c) ∈ R1=∑ ni-1cei-csi+1；
From above formula, the area a in region is exactly the R that counts in region.
Described step 4) in, it is to carry each point value in territory one of this point of the value of any in digital picture or Serial No.
Mesophyticum replace, allow the pixel value of surrounding close to actual value, thus eliminate isolated noise spot；
Method is the two-dimentional sleiding form of certain structure, pixel in plate is ranked up according to the size of pixel value, raw
Become monotone increasing (or decline) for 2-D data sequence, containing noise in the result of Threshold segmentation, this is not end product,
The process of noise, by using image smoothing to suppress.
Described step 4) in, testing result is during removing noise, and all connected regions being less than 4 pixels are seen
Make noise and be removed, in order to distinguish noise and defect, it is assumed that noise is equally distributed, and the defect belonging to a cut together is
Close to each other, therefore, it can by expanding gap Guan Bi medium and small for defect area. in order to calculate connected region,
Suitable two pixels must be defined should be considered to communicate with each other.
The method detection surface defect of the present invention, preferably reduces in image acquisition process, is limited to ambient light
Interference.Most become more meticulous to have isolated metal defect and be all the edge being highlighted.Moreover, this method is the most easily grasped
Making, feasibility is the strongest.
By using LED circular lamp direct dark field frontlighting mode to gather image, then use dynamic threshold segmentation
Method by the region of interesting extraction of scratch detection in region, surface out, then by use provincial characteristics, regional morphology to table
Region, face uses erosion operation to remove miscellaneous point and little outthrust in cut zone, it is ensured that have enough precision in calculating, finally
Radiation conversion, image smoothing, connected region is used to extract scheduling algorithm and detect cut and show its result.
The directivity of illumination generally has two kinds: diffusion and direct irradiation.
During diffusion, light is almost equally in the intensity of all directions.During direct irradiation, the light that light source sends concentrates on non-
In the narrowest spatial dimension.Detecting liking surface scratch herein, owing to this type of defects detection area is little, cut is inconspicuous waits bar
Part, under bright field illumination mode, it is difficult to obtain preferable cut image.Therefore this detection uses the direct dark field of LED circular lamp
Lighting system, ring light and body surface, in the least angle, so can highlight breach and the projection of measured object, so drawing
Trace, texture or engraving character etc. are enhanced, and see to become apparent from.
The image collected is not provided that in image the information comprising object.In order to obtain the object information in image, must
Must carry out image segmentation, image is divided into some regions by image segmentation exactly, and in the same area, the feature of image is close；And
In different regions, characteristics of image difference is bigger.Characteristics of image can be the feature of image itself, such as gray scale, the edge of pixel
Profile and texture etc..Imagethresholding is a kind of the most frequently used, is also simplest image partition method simultaneously.Image threshold
The purpose changed is according to gray level, and collection of pixels carries out a division, and each subset obtained forms one and real-world scene
Corresponding region, has consistent attribute, and adjacent area layout has this consistent attribute inside regional.Threshold segmentation
Operation is defined as
S={ (r, c) ∈ Rgmin≤fr, c≤gmax}
Therefore, in gray value in image ROIR is in a certain appointment intensity value ranges by Threshold segmentation, all point chooses output
In the S of region.Make gmin=0 or gmax=2b-1.If illumination can keep constant, threshold value gmin and gmax can be when system be arranged
Determined choosing and never with being adjusted.Threshold segmentation is divided into fixed threshold segmentation and dynamic threshold segmentation.Dynamic threshold segmentation will
The operation that image and its local background compare is referred to as dynamic threshold segmentation and processes, and represents input picture with fr, c, with gr,
C represents the image after smoothing, then process as follows to the dynamic threshold segmentation of bright object
S={ (r, c) ∈ Rfr, c-gr, c >=gdif}
And the dynamic threshold segmentation process to dark object is
S={ (r, c) ∈ Rfr, c-gr, c≤-gdiff}
In dynamic threshold segmentation processes, the size of smoothing filter determines the size of the divided object out of energy.
If filter size is the least, then the local background estimated at the center of object is by undesirable.
Through process above, region or the sub-pixel precision profile extracted from image can be obtained.But they are only
Contain the original description to segmentation result.Some region or profile must also be selected below, as segmentation from segmentation result
In result, undesired part is removed.Up to the present, simplest provincial characteristics is the area in region:
A=R=∑ (r, c) ∈ R1=∑ ni-1cei-csi+1
From above formula, the area a in region is exactly the R that counts in region.If region represents with a width bianry image, that
Area with first summation equation zoning in formula；If region run-length encoding represents, then by public affairs
The area of second summation equation zoning in formula 4.One region can be considered a union of its all strokes, and
The area of each stroke is as easy as rolling off a log calculating.Note few a lot of than first cumulative formula of the item of second cumulative formula.So,
The stroke representation in region can make the calculating speed of region area soon a lot, and this feature is to almost all of provincial characteristics all
It is suitable for.
After the most a series of process, area-of-interest can be carried out defects detection, need to reuse dynamic
The operation of state Threshold segmentation detects defect, can estimate background with median filter.
The ultimate principle of medium filtering is that one of this point of the value of any in digital picture or Serial No. is carried in territory
The Mesophyticum of each point value replaces, and allows the pixel value of surrounding close to actual value, thus eliminates isolated noise spot.Method is certain knot
The two-dimentional sleiding form of structure, is ranked up pixel in plate according to the size of pixel value, and generate monotone increasing (or decline) is
2-D data sequence.Containing noise in the result of Threshold segmentation, this is not end product.The process of noise, is schemed by use
Suppress as smoothing.
By aforesaid operations, surface scratch detection terminates substantially, due to during removing noise, all is less than 4
The connected region of pixel is counted as noise and is removed.In order to distinguish noise and defect, it is assumed that noise is equally distributed, and with
The defect belonging to a cut is close to each other, therefore, it can by expanding gap Guan Bi medium and small for defect area.In order to enable
Enough calculate connected region, it is necessary to define suitable two pixels and should be considered to communicate with each other.It is more than this detection process, logical
Cross aforesaid operations, just can obtain wanted testing result.
From assigned catalogue, read in the template of surface scratch image continuously, and image size is configured, use LED ring
Shape light direct dark field illumination gained surface scratch figure.
Cut is shown as highlighted in black background region, but 4 in the edge on surface and surface plane part
The edge of inner square is also highlighted, in order to distinguish the edge of cut and surface, is first partitioned into bright marginal area.So
After from the region on surface, deduct the region being partitioned into, thus the area-of-interest of scratch detection is narrowed down to the district after subtracting each other
Territory.
Being processed by above, next step determines the plane needing detection, therefore to extract area-of-interest.Need
The bright border on surface and middle 4 little foursquare bright borders are removed from segmentation result.It is first necessary to know that surface exists
Direction in image and size, for obtaining direction and the size on surface, reuse regional morphology and be just partitioned into 4 of inside
Square.The little cavity on the inner square edge being above partitioned into, inner square border is filled first by 2 closed operations
On have gap.
So far, cut is appointed in the bright borderline region being partitioned into.In order to detect cut, need cut from dividing
Cut in result and separate.Shape due to the foursquare borderline region of known internal, it is possible to use suitably structural element is opened
Cut is removed in computing.Generate a structural element for this, the rectangle parallel by two axles forms, and represents two of inner square
Opposite side.
When generating rectangle in suitable direction, structural element can not rotate.However it is necessary that according to direction transformation square
Shape center.
Opening operation can serve as template matching, all points matched with structural element in returning input area.As
Being expected, result contains inner square border.But result is appointed containing surface portion external boundary, this is because interior square
To the distance of surface-boundary as interior foursquare length of side size.In order to remove as boundary member, take away computing result and
Occur simultaneously in region, surface after corrosion.
So obtain containing only the region RegionSquares on 4 inner square borders.Surface to be checked is just
It it is the difference in region, surface and interior square border.
Before calculating difference, circular configuration element is used to corrode to remove border to region, surface.The radius of circle
For Border-Width with BorderTolerance's and, the two value is all predefined.Radius adds
Remove when BorderTolerance is to detect with border very close to pixel, these pixel grey scales can be by border
Impact, may be wrongly judged defect.In like manner, in representing, square borderline region also to expand.Obtain containing table
The area-of-interest Re-gionSurface of face detection plane.Notice that superficial white border and interior square white border do not wrap
Containing in the zone.
Through process above, it now is possible to area-of-interest is carried out defects detection: reuse dynamic threshold and divide
Cut operation to detect defect, now can estimate background with median filter.Based on known maximum scratch width
ScratchWidthMax, utilizes Scratch-WidthMax to remove all cuts as median filter radius.Dark owing to using
Visual field frontlighting, cut is bright region in the picture, can readily use predefined
ScratchGrayDiffMin is as Threshold segmentation.
In this case, all connected regions being less than 4 pixels are seen as noise and are removed.Not all make an uproar
Sound is removed the most completely, improves threshold value further and may remove the discontinuous defect area of part simultaneously.Make an uproar to distinguish
Sound and defect, it is assumed that noise is equally distributed, and the defect belonging to a cut together is close to each other, therefore, it can pass through
Expand gap Guan Bi medium and small for defect area.The defect originally disconnected connects together after expanding, to the district after expanding
Connected region is recalculated in territory.In order to obtain the original-shape of defect, take the friendship of unexpanded front original area and connected region
Collection.
Notice that intersection operation does not affect the connectedness of each composition, then, be increase only the profile of connected region by expansion.
Finally select all regions bigger than predetermined minimum cut.
Above-described is only the preferred embodiment of the present invention, it is noted that for those of ordinary skill in the art
For, without departing from the concept of the premise of the invention, it is also possible to make some deformation and improvement, these broadly fall into the present invention
Protection domain.
Claims (5)
1. one kind based on the machine vision method to Surface testing, it is characterised in that comprise the following steps:
1) LED circular lamp direct dark field frontlighting mode is used to gather image；
2) dynamic threshold segmentation method is used by the region of interesting extraction of scratch detection in region, surface out；
3) again by use provincial characteristics, regional morphology region, surface use erosion operation is removed in cut zone miscellaneous point and
Little outthrust, it is ensured that have enough precision in calculating；
4) radiation conversion, image smoothing, connected region is finally used to extract scheduling algorithm and detect cut and show its result.
Method based on machine vision to Surface testing the most according to claim 1, it is characterised in that described step 2)
In, the purpose of image threshold is according to gray level, and collection of pixels carries out a division, and each subset obtained forms one
The region corresponding with real-world scene, has consistent attribute, and adjacent area layout has this consistent genus inside regional
Property, Threshold segmentation operation is defined as
S={ (r, c) ∈ Rgmin≤fr, c≤gmax}；
Therefore, in gray value in image ROIR is in a certain appointment intensity value ranges by Threshold segmentation, all point chooses output area
In S, make gmin=0 or gmax=2b-1, if illumination can keep constant, threshold value gmin and gmax can be determined when system is arranged
Select and never with being adjusted；Threshold segmentation is divided into fixed threshold segmentation and dynamic threshold segmentation；
The operation that image and its local background are compared by dynamic threshold segmentation is referred to as dynamic threshold segmentation and processes, and uses fr, c
Represent input picture, use gr, c represents the image after smoothing, then processes as follows to the dynamic threshold segmentation of bright object
S={ (r, c) ∈ Rfr, c-gr, c >=gdiff}；
And the dynamic threshold segmentation process to dark object is S={ (r, c) ∈ Rfr, c-gr, c≤-gdiff}。
Method based on machine vision to Surface testing the most according to claim 1, it is characterised in that described step 3)
In, the algorithm of employing is a=R=∑ (r, c) ∈ R1=∑ ni-1cei-csi+1；
From above formula, the area a in region is exactly the R that counts in region.
Method based on machine vision to Surface testing the most according to claim 1, it is characterised in that described step 4)
In, it is to carry the Mesophyticum of each point value in territory one of this point of the value of any in digital picture or Serial No. to replace, allows around
Pixel value close to actual value, thus eliminate isolated noise spot；
Method is the two-dimentional sleiding form of certain structure, pixel in plate is ranked up according to the size of pixel value, generates single
Adjust rise (or decline) for 2-D data sequence, containing noise in the result of Threshold segmentation, this is not end product, noise
Process, by use image smoothing suppress.
Method based on machine vision to Surface testing the most according to claim 1, it is characterised in that described step 4)
In, testing result is during removing noise, and all connected regions being less than 4 pixels are counted as noise and are removed, for
Differentiation noise and defect, it is assumed that noise is equally distributed, and the defect belonging to a cut together is close to each other, therefore,
By expanding, gap medium and small for defect area can be closed；In order to calculate connected region, it is necessary to define suitable two
Pixel should be considered to communicate with each other.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610487518.1A CN106157303A (en) | 2016-06-24 | 2016-06-24 | A kind of method based on machine vision to Surface testing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610487518.1A CN106157303A (en) | 2016-06-24 | 2016-06-24 | A kind of method based on machine vision to Surface testing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106157303A true CN106157303A (en) | 2016-11-23 |
Family
ID=57349525
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610487518.1A Pending CN106157303A (en) | 2016-06-24 | 2016-06-24 | A kind of method based on machine vision to Surface testing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106157303A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107255641A (en) * | 2017-06-06 | 2017-10-17 | 西安理工大学 | A kind of method that Machine Vision Detection is carried out for GRIN Lens surface defect |
| CN107358598A (en) * | 2017-05-24 | 2017-11-17 | 上海视马艾智能科技有限公司 | A kind of scratch detection method and apparatus |
| CN107358597A (en) * | 2017-05-24 | 2017-11-17 | 上海视马艾智能科技有限公司 | A kind of glue point method of inspection and device |
| CN107977960A (en) * | 2017-11-24 | 2018-05-01 | 南京航空航天大学 | A kind of car surface scratch detection algorithm based on improved SUSAN operators |
| CN108459030A (en) * | 2018-02-08 | 2018-08-28 | 东华大学 | One kind being applied to non-planar plastic smooth surface flaw on-line measuring device and method |
| CN109374638A (en) * | 2018-12-18 | 2019-02-22 | 王章飞 | A kind of wood floor surface detection device and its detection method based on machine vision |
| CN109829904A (en) * | 2019-01-29 | 2019-05-31 | 京东方科技集团股份有限公司 | Detect the method, apparatus of dust, electronic equipment, readable storage medium storing program for executing on screen |
| CN110021012A (en) * | 2019-03-27 | 2019-07-16 | 安徽皓视光电科技有限公司 | Mobile lens sight glass defect inspection method based on machine vision technique |
| CN110514665A (en) * | 2019-09-03 | 2019-11-29 | 博科视(苏州)技术有限公司 | A kind of detection method of electronic product plastic shell scratch defects |
| CN111060526A (en) * | 2018-10-16 | 2020-04-24 | 苏州微微视视觉科技有限公司 | Method for detecting surface defects of metal target based on picture processing |
| CN111223080A (en) * | 2020-01-02 | 2020-06-02 | 长江存储科技有限责任公司 | Wafer detection method and device, electronic equipment and storage medium |
| CN112508937A (en) * | 2020-12-22 | 2021-03-16 | 北京百度网讯科技有限公司 | Method and device for generating scratch data, electronic equipment and storage medium |
| WO2021168733A1 (en) * | 2020-02-27 | 2021-09-02 | 京东方科技集团股份有限公司 | Defect detection method and apparatus for defect image, and computer-readable storage medium |
| CN115359044A (en) * | 2022-10-18 | 2022-11-18 | 南通群富金属科技有限公司 | Metal part surface scratch detection method based on image enhancement |
| CN115575321A (en) * | 2022-09-26 | 2023-01-06 | 广州惠中木业有限公司 | Wood board surface defect detection system and method based on mechanical vision |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102759528A (en) * | 2012-07-09 | 2012-10-31 | 陕西科技大学 | Method for detecting diseases of crop leaves |
| CN103700116A (en) * | 2012-09-27 | 2014-04-02 | 中国航天科工集团第二研究院二O七所 | Background modeling method for movement target detection |
| CN105261037A (en) * | 2015-10-08 | 2016-01-20 | 重庆理工大学 | Moving object detection method capable of automatically adapting to complex scenes |
| CN105388162A (en) * | 2015-10-28 | 2016-03-09 | 镇江苏仪德科技有限公司 | Raw material silicon wafer surface scratch detection method based on machine vision |
-
2016
- 2016-06-24 CN CN201610487518.1A patent/CN106157303A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102759528A (en) * | 2012-07-09 | 2012-10-31 | 陕西科技大学 | Method for detecting diseases of crop leaves |
| CN103700116A (en) * | 2012-09-27 | 2014-04-02 | 中国航天科工集团第二研究院二O七所 | Background modeling method for movement target detection |
| CN105261037A (en) * | 2015-10-08 | 2016-01-20 | 重庆理工大学 | Moving object detection method capable of automatically adapting to complex scenes |
| CN105388162A (en) * | 2015-10-28 | 2016-03-09 | 镇江苏仪德科技有限公司 | Raw material silicon wafer surface scratch detection method based on machine vision |
Non-Patent Citations (2)
| Title |
|---|
| 何小利等: "基于机器视觉技术的物体表面缺陷检测", 《洛阳师范学院学报》 * |
| 刘泽等: "钢轨表面缺陷检测机器视觉系统的设计", 《电子测量与仪器学报》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107358598A (en) * | 2017-05-24 | 2017-11-17 | 上海视马艾智能科技有限公司 | A kind of scratch detection method and apparatus |
| CN107358597A (en) * | 2017-05-24 | 2017-11-17 | 上海视马艾智能科技有限公司 | A kind of glue point method of inspection and device |
| CN107255641B (en) * | 2017-06-06 | 2019-11-22 | 西安理工大学 | A method of Machine Vision Detection is carried out for self-focusing lens surface defect |
| CN107255641A (en) * | 2017-06-06 | 2017-10-17 | 西安理工大学 | A kind of method that Machine Vision Detection is carried out for GRIN Lens surface defect |
| CN107977960A (en) * | 2017-11-24 | 2018-05-01 | 南京航空航天大学 | A kind of car surface scratch detection algorithm based on improved SUSAN operators |
| CN107977960B (en) * | 2017-11-24 | 2021-08-06 | 南京航空航天大学 | Saloon car surface scratch detection algorithm based on improved SUSAN operator |
| CN108459030A (en) * | 2018-02-08 | 2018-08-28 | 东华大学 | One kind being applied to non-planar plastic smooth surface flaw on-line measuring device and method |
| CN111060526A (en) * | 2018-10-16 | 2020-04-24 | 苏州微微视视觉科技有限公司 | Method for detecting surface defects of metal target based on picture processing |
| CN109374638A (en) * | 2018-12-18 | 2019-02-22 | 王章飞 | A kind of wood floor surface detection device and its detection method based on machine vision |
| CN109374638B (en) * | 2018-12-18 | 2022-01-18 | 深圳市鼎源检测技术有限公司 | Wood floor surface detection device based on machine vision and detection method thereof |
| CN109829904A (en) * | 2019-01-29 | 2019-05-31 | 京东方科技集团股份有限公司 | Detect the method, apparatus of dust, electronic equipment, readable storage medium storing program for executing on screen |
| CN110021012A (en) * | 2019-03-27 | 2019-07-16 | 安徽皓视光电科技有限公司 | Mobile lens sight glass defect inspection method based on machine vision technique |
| CN110021012B (en) * | 2019-03-27 | 2023-09-26 | 安徽皓视光电科技有限公司 | Mobile phone lens window glass defect detection method based on machine vision technology |
| CN110514665A (en) * | 2019-09-03 | 2019-11-29 | 博科视(苏州)技术有限公司 | A kind of detection method of electronic product plastic shell scratch defects |
| CN111223080A (en) * | 2020-01-02 | 2020-06-02 | 长江存储科技有限责任公司 | Wafer detection method and device, electronic equipment and storage medium |
| WO2021168733A1 (en) * | 2020-02-27 | 2021-09-02 | 京东方科技集团股份有限公司 | Defect detection method and apparatus for defect image, and computer-readable storage medium |
| CN112508937A (en) * | 2020-12-22 | 2021-03-16 | 北京百度网讯科技有限公司 | Method and device for generating scratch data, electronic equipment and storage medium |
| CN115575321A (en) * | 2022-09-26 | 2023-01-06 | 广州惠中木业有限公司 | Wood board surface defect detection system and method based on mechanical vision |
| CN115575321B (en) * | 2022-09-26 | 2023-07-21 | 广州惠中木业有限公司 | Wood board surface defect detection system and method based on machine vision |
| CN115359044A (en) * | 2022-10-18 | 2022-11-18 | 南通群富金属科技有限公司 | Metal part surface scratch detection method based on image enhancement |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106157303A (en) | A kind of method based on machine vision to Surface testing | |
| CN109961049B (en) | Cigarette brand identification method under complex scene | |
| CN107025652B (en) | A kind of flame detecting method based on kinetic characteristic and color space time information | |
| CN103366167B (en) | System and method for processing image for identifying alphanumeric characters present in a series | |
| CN101299268B (en) | Semantic object dividing method suitable for low depth image | |
| CN108416766B (en) | Double-side light-entering type light guide plate defect visual detection method | |
| CN104794491A (en) | Fuzzy clustering steel plate surface defect detection method based on pre classification | |
| CN105445277A (en) | Visual and intelligent detection method for surface quality of FPC (Flexible Printed Circuit) | |
| CN104103082A (en) | Image saliency detection method based on region description and priori knowledge | |
| CN106296744A (en) | A kind of combining adaptive model and the moving target detecting method of many shading attributes | |
| CN106780486A (en) | A kind of Surface Defects in Steel Plate image extraction method | |
| CN108062508B (en) | The extracting method of equipment in substation's complex background infrared image | |
| CN103295013A (en) | Pared area based single-image shadow detection method | |
| CN106096491B (en) | Automatic identification method for microaneurysms in fundus color photographic image | |
| CN111754538B (en) | Threshold segmentation method for USB surface defect detection | |
| CN109685760A (en) | A kind of SLM powder bed powder laying image convex closure depression defect detection method based on MATLAB | |
| Lian et al. | A novel method on moving-objects detection based on background subtraction and three frames differencing | |
| Mustafa | Feature selection using sequential backward method in melanoma recognition | |
| CN109949294A (en) | A kind of fracture apperance figure crack defect extracting method based on OpenCV | |
| JP2013105245A (en) | Image processing method, device and program | |
| CN106683062A (en) | Method of checking the moving target on the basis of ViBe under a stationary camera | |
| CN104268845A (en) | Self-adaptive double local reinforcement method of extreme-value temperature difference short wave infrared image | |
| Selamet et al. | Automatic detection and classification of defective areas on metal parts by using adaptive fusion of faster R-CNN and shape from shading | |
| CN103065145A (en) | Vehicle movement shadow eliminating method | |
| CN110110665B (en) | Detection method for hand area in driving environment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161123 |