CN110517233A - A kind of defect classification learning system and its classification method based on artificial intelligence - Google Patents
A kind of defect classification learning system and its classification method based on artificial intelligence Download PDFInfo
- Publication number
- CN110517233A CN110517233A CN201910755707.6A CN201910755707A CN110517233A CN 110517233 A CN110517233 A CN 110517233A CN 201910755707 A CN201910755707 A CN 201910755707A CN 110517233 A CN110517233 A CN 110517233A
- Authority
- CN
- China
- Prior art keywords
- defect
- detection
- shooting area
- signal
- described image
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 80
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 9
- 238000001514 detection method Methods 0.000 claims abstract description 52
- 238000012545 processing Methods 0.000 claims abstract description 37
- 238000000605 extraction Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 9
- 230000005540 biological transmission Effects 0.000 claims description 8
- 230000002950 deficient Effects 0.000 claims description 7
- 230000003287 optical effect Effects 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000010408 sweeping Methods 0.000 claims 1
- 238000012360 testing method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000009659 non-destructive testing Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 208000003464 asthenopia Diseases 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
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
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
-
- 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
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a kind of defect classification learning system and its classification method based on artificial intelligence, including a camera apparatus are provided with image sensitive chip wherein there are two shooting areas for camera apparatus tool in each shooting area；One light supply apparatus；One detection device, the detection device includes a roll-in component and a conveyer belt, the roll-in component drives the conveyer belt to be recycled, and the top of the conveyer belt is arranged in the camera apparatus, and the lower section of the conveyer belt is arranged in correspond to shooting area in the light supply apparatus；An and image processing unit, described image processing unit communicates to connect the camera apparatus, with picture signal captured in detection image sensitive chip, and it controls the camera apparatus and shoots same coiled material surface in different shooting areas, and combine detection content completion twice for the defects detection of coiled material surface.The defect of coiled material surface can be recorded and analyzed and be stored.
Description
Technical field
The present invention relates to Surface testing field, in particular to a kind of defect classification learning system based on artificial intelligence and its
Classification method.
Background technique
Surface testing be premised on not damaging the service performance of detected object, with physics, chemistry, material science and
Based on engineering theory, various engineering material, components and product are effectively examined, so as to evaluating the complete of them
Property, continuity and security reliability.The feature on surface and distribution disclose property, shape, position and the quantity of defect, therefore
Surface testing focuses primarily upon the content of detection or more.It can be completed by comprehensive analysis to defect and evaluation for defect point
The study of class, and finally improve whole production efficiency.Therefore, correct identification and analyzing defect image account in coiled material surface detection
Considerable status.
The Surface testing of traditional coiled material product usually carries out the differentiation of product defects by manually.And the inspection of coiled material surface
It surveys, it is often necessary to which several inspectors complete the detection of thousands of coiled materials daily, and coiled material reaches certain amount in length
When, missing inspection, erroneous judgement problem caused by it inevitably will appear in actually detected because of inspector's asthenopia.Moreover, non-destructive testing
Technology forward direction Nondestructive Evaluation direction is developed, and automation and intelligence are trends of the times.Product defects image automatic identification technology is
The key point of intelligent non-destructive detecting device certainly will be used widely in non-destructive testing industry.
Summary of the invention
The defect classification learning system that the technical problem to be solved in the present invention is to provide a kind of based on artificial intelligence and its point
Class method can be recorded and analyzed and be stored for the defect of coiled material surface.
In order to solve the above-mentioned technical problem, the technical solution of the present invention is as follows:
A kind of defect classification learning system based on artificial intelligence, comprising:
One camera apparatus is provided with figure in each shooting area wherein there are two shooting areas for camera apparatus tool
As sensitive chip；
One light supply apparatus；
One detection device, the detection device include a roll-in component and a conveyer belt, described in the roll-in component drives
Conveyer belt is recycled, and the top of the conveyer belt is arranged in the camera apparatus, and the light supply apparatus is arranged described
The lower section of conveyer belt is to correspond to shooting area；And
One image processing unit, described image processing unit communicates to connect the camera apparatus, with the photosensitive core of detection image
Captured picture signal in piece, and control the camera apparatus and shoot same coiled material surface in different shooting areas, and
Detection content completes the defects detection for coiled material surface to joint twice.
Preferably, described image processing unit includes a scan module, a preprocessing module, a characteristic extracting module and one
Comparison module, the scan module communicate to connect the camera apparatus to obtain and scan described image signal, the pretreatment
Module communicates to connect the scan module to obtain the picture signal after scanning and be pre-processed, and the characteristic extracting module is logical
Letter connects the preprocessing module and the comparison module, to carry out defects detection to image module.
Preferably, described image processing unit further includes a case database, described in the case database communication connection
Comparison module, to complete the comparison of case data, and it is scarce in the case database when being not belonging in the comparison module
It falls into, is then stored in the case database to wait name.
Preferably, the light supply apparatus includes multiple LED light sources and an at least high frequency flashing light lamp, the LED light source difference
Position corresponding in different shooting areas is set, the high frequency flashing light lamp is then disposed therein in a shooting area,
To complete defects detection.
Preferably, the roll-in component further includes an encoder, and the encoder controls the transmission speed of the roll-in component
Degree.
The present invention also provides a kind of classification methods based on machine vision, comprising the following steps:
(a) start the defect classification learning system, when the coiled material product passes through first shooting area, setting exists
The image sensitive chip of first shooting area is shot, and picture signal is exported；
(b) scan and pre-process described image signal, in pretreatment stage, the next bat of waiting of doubtful defective problem
It takes the photograph corresponding described image signal in region and carries out defect diagonsis, there is no problem is then not necessarily to carry out the operation of next shooting area,
Export qualifying signal；
(c) when corresponding coiled material product passes through second shooting area, the image sense in the shooting area is set
Optical chip is shot, and picture signal is exported, and described image signal passes through the detection in defect diagonsis stage, belongs to known defect
When problem, then current flaw indication is exported, when belonging to unknown defect problem, is then stored in present case database.
Preferably, further comprising the steps of among step (b):
(b1) in pretreatment stage, described image signal is scanned；
(b2) picture signal after scanning, is split regional processing, described image signal is divided into multiple detection zones；
(b3) detection zone after segmentation is detected one by one, when a problem occurs, then exports detection signal to carry out
The operation of downstream；It does not go wrong, then exports qualifying signal without carrying out subsequent operation.
Preferably, further comprising the steps of among step (c):
(c1) in second shooting area, interference operation is carried out after removal interference to picture signal and defect occurs
Problem enters next step, and the then output qualifying signal not gone wrong need not carry out subsequent operation；
(c2) feature extraction is carried out to the picture signal for defect problem occur, and then the defect characteristic of extraction and the case
The defects of example database is compared, and compares the output met currently corresponding flaw indication, compares incongruent deposit and work as
In preceding case database.
Preferably, further comprising the steps of among step (c):
(c0) high frequency started when the light supply apparatus is changed to and is shot with the LED light source being always on from the LED light source being always on
Flash lamp.
By adopting the above technical scheme, secondary inspection is carried out to coiled material surface by image processing unit due to the detection system
It surveys, so that the beneficial effects of the present invention are:
The first, detection more efficiently, is divided into and detecting twice, only pre-processes in detection for the first time, does defect for the second time and examine
It is disconnected, while improving detection accuracy, improve detection speed.
The second, processing is split to same detection target in detection process, and when detecting for second, passes through light source
Depth detection is done in variation.
In third, detection process, there is unknown defect problem, is then stored in case database to wait name.
Detailed description of the invention
Fig. 1 is the device partial structure diagram of defect classification learning system of the present invention；
Fig. 2 is the image processing section flow diagram of defect classification learning system of the present invention；
Fig. 3 is the step schematic diagram of classification method of the present invention.
Specific embodiment
Specific embodiments of the present invention will be further explained with reference to the accompanying drawing.It should be noted that for
The explanation of these embodiments is used to help understand the present invention, but and does not constitute a limitation of the invention.In addition, disclosed below
The each embodiment of the present invention involved in technical characteristic can be combined with each other as long as they do not conflict with each other.
As shown in Figure 1, the present invention provides a kind of defect classification learning system based on artificial intelligence.The defect classification
Learning system includes a camera apparatus 10, a light supply apparatus 20, an image processing unit 30 and a detection device 40, wherein described
Detection device 40 includes a roll-in component 41 and a conveyer belt 42, wherein the conveyer belt 42 is arranged in the roll-in component 41
Both ends are provided with coiled material product on the conveyer belt 42, so that described to complete the cyclic transfer for the conveyer belt 42
Coiled material product can follow the conveyer belt 42 to be transmitted.
Further, the rolling device 41 is provided with an encoder 410, and the encoder 410 can control the roller
The velocity of rotation of component 41 is pressed, that is, controls the speed of the conveyer belt 42.The camera apparatus 10 is arranged in the transmission
With 42 top, the lower section of the conveyer belt 42 is arranged in the light supply apparatus 20.
Specifically, when the light supply apparatus 20 emits light source, so that the camera apparatus 10 is collecting the coiled material table
When face, there can be the picture being more clear.So that heretofore described image processing unit 30 is capable of handling more preferably.
Normally, the camera apparatus 10 includes image sensitive chip and digital interface group at so that the camera apparatus
10 can convert optical signals to orderly electric signal.The pith of the machine vision is the property of the camera apparatus 10
Can, the camera apparatus 10 is provided with a shooting area 100, and the conveyer belt 42 is arranged in wherein in the shooting area 100
A part, as the transmission of the conveyer belt 42 is mobile, the coiled material surface is taken by the shooting area 100.
In the shooting area 100 of the coiled material surface, the camera apparatus 10 takes the coiled material surface
Picture signal is transferred to described image processing unit 30.The picture signal of the camera apparatus 10 is transferred to described image processing
It is detected in unit 30, defect image defective is carried in picture signal when detecting, due to defect image and normogram
As there are grayscale differences, therefore 100 in the shooting area, the shooting speed of the camera apparatus 10 and the conveyer belt 42
Transmission speed has certain proportion, it is preferable that the shooting speed of the camera apparatus 10 and the transmission speed of the conveyer belt 42
It is identical, so that the camera apparatus 10 shoot once every a shooting area 100, of course, in order to guarantee to shoot matter
Amount and accuracy rate, the camera apparatus 10 are additionally provided with another shooting area 101, and the shooting area 101 is arranged in the bat
Take the photograph the rear in region 100, that is, the rear of 42 direction of transfer of the conveyer belt.
Specifically, in the previous shooting area 100, the described image signal of the camera apparatus 10 is transmitted to described
In image processing unit 30, when described image processing unit 30 is tested with some defect problems, described image processing unit 30
Another shooting area 100 is shot.That is, the camera apparatus 10 is respectively in 100 He of shooting area
It is provided with image sensitive chip in 101, therefore, in the shooting process, is tested with some defects in previous shooting area 100
After problem, the described image sensitive chip in latter shooting area 101 is just opened, so that institute in latter shooting area 101
Image sensitive chip is stated further to detect the progress of the coiled material surface in the shooting area 100.
More specifically, described image processing unit 30 locates the described image signal in previous shooting area 100 in advance
Reason, described image processing unit 30 are split processing to described image signal, to detect in current each cut zone
Defect problem, and to doubtful defective part in latter shooting area 100, it is further detected, to improve detection effect
Rate.
In above process, the lower section of the conveyer belt 42 is arranged in the light supply apparatus 20, so that the light supply apparatus
Camera apparatus 10 described in 20 light source face.Normally, the light supply apparatus 20 mainly uses LED light source, the LED light source
Stability and service life are higher.On the other hand, the light supply apparatus 20 is also configured with high frequency flashing light lamp, due to high frequency flashing light lamp
Light uniformity to be got well compared with LED light source.Therefore, the light supply apparatus 20 includes LED light source 21 and high frequency flashing light lamp 22, specifically
Ground, the LED light source 21 keep long bright state, and the high frequency flashing light lamp 22 is in described image processing unit 30 to the coiled material table
Face starts when detect for second, keeps standby mode, usually to save power supply.
As shown in Fig. 2, described image processing unit 30 includes a scan module 31, described to sweep according to aforesaid operations mode
Module 31 is retouched for receiving the picture signal of the camera apparatus 10, and is scanned into orderly electric signal.It further include a pretreatment
Module 32, the picture signal after 32 pairs of preprocessing module scannings pre-process, and include two steps in preprocessing process:
The first step divides described image signal, that is to say, that divide the coiled material surface, the area where the coiled material surface
Domain is divided into several pieces, it is preferable that can be divided into 9 parts.Certainly, cutting procedure is actually and completes in virtual process.
Second step carries out problem detection to the described image signal in each region after being divided into several regions,
That is being detected for whether there is problem in each region.Detect the problems in described image signal, in fact it could happen that accidentally
Sentence, but is not necessarily to the detection judged by accident in pretreatment stage.
Specifically, in pretreatment stage, the doubtful defective problem of the coiled material surface is detected in the cut zone
When, the coiled material in current shooting region 100 is transferred in next shooting area 101 by the conveyer belt 42, the figure
As processing unit 30 receives the described image signal in the shooting area 101.At this point, due in previous shooting area 101
Preprocessing process in, the doubtful problematic cut zone of described image signal is marked with, so that the latter shooting area
The image of the cut zone of the direct shot mark of the sensitive chip of camera apparatus 10 described in domain 101.
Therefore, in the latter shooting area 101, described image processing unit 30 is in the defect diagonsis stage, in defect
In diagnostic phases, it is divided into the following steps:
The first, interference operation is carried out, described image processing unit 30 directly carries out interference behaviour to described image signal
Make, specifically, is confirmed using gray level image.In this process, described image processing unit 30 is provided with one and goes to interfere
Module 33, it is described that interference module 33 is gone to carry out interference operation to the defects of described image signal feature, when judging by accident,
Erroneous judgement signal is then exported, without carrying out next step, directly exports current qualifying signal；It is on the contrary then continue next step and continue to sentence
It is disconnected.
In the whole process, described image processing unit 30 includes a case database 300, the case database 300
Include some basic typical defect characteristic picture signals, compares and measure for the later period.On the other hand, when not having
Detected defect in case database 300 described in typing is then stored in the case database 300 and waits name.In addition,
Described image processing unit 30 can also carry out the acquisition and comparison of case data by way of external connection internet, therefore at this
In kind technical solution, the case database is not needed.
The second, feature extraction is carried out, in the characteristic extraction procedure, described image processing unit 30 includes a feature extraction
Module 34 and a comparison module 35, the characteristic extracting module 34 proposes the characteristic signal in current image signal, described
The comparison that case database of the comparison module 35 to current signature signal and in relation to defect carries out then is exported when comparison meets
Corresponding defect characteristic signal is then stored in present case database when comparison is not met.
As shown in figure 3, therefore, according to above embodiment, the present invention provides classification process below:
L10, the starting detection device 40 and the camera apparatus 10 and light supply apparatus 20,41 band of roll-in component
Move the conveyer belt 42 and the coiled material product on the conveyer belt 42；
L11, when the coiled material product passes through first shooting area 100, the camera apparatus 10 is arranged at first
The image sensitive chip of shooting area 100 is shot, and picture signal is exported；
L12, described image processing unit 30 the scan module 31 picture signal is scanned, and then will scanning after
Picture signal be transferred to the preprocessing module 32 and carry out pretreatment stage；In pretreatment stage, doubtful defective problem
The next shooting area 101 of waiting in corresponding described image signal carry out defect diagonsis, it is then next without carrying out that there is no problem
The operation of shooting area 101.
L13, when corresponding coiled material product passes through second shooting area 101, the camera apparatus 10 is arranged described
Image sensitive chip in shooting area 101 is shot, export picture signal, described image signal by pretreatment stage it
Afterwards, described image signal passes through the defect diagonsis stage, when belonging to known defect problem, then exports current flaw indication,
When belonging to unknown defect problem, then present case database 300 is stored in.
Further, specifically further include following below scheme in L12:
L121, in pretreatment stage, described image signal is scanned；
Picture signal after L122, scanning, is split regional processing, and described image signal is divided into multiple detection zones
Domain；
L123, the detection zone after segmentation is detected one by one, when a problem occurs, then exports detection signal to carry out
The operation of downstream；It does not go wrong, then exports qualifying signal without carrying out subsequent operation.
Further, specifically further include following below scheme in L13:
L131, in second shooting area 101, the camera apparatus 10 is arranged in the shooting area 101
Image sensitive chip shoots and exports picture signal and carry out interference operation, after removal interference, the entrance of defect problem occurs
Next step, the then output qualifying signal not gone wrong need not carry out subsequent operation；
L132, to there is defect problem picture signal carry out feature extraction, and then the defect characteristic of extraction with it is described
The defects of case database 300 is compared, and compares the output met currently corresponding flaw indication, compares incongruent deposit
Enter in present case database 300.
In addition, the light supply apparatus 20 is changed to shoot with the LED light source being always on from the LED light source 21 being always in L13
The high frequency flashing light lamp of Shi Qidong.
According to aforesaid operations process, invention further provides a kind of classification methods based on machine vision, including
Following steps:
(a) start the defect classification learning system, when the coiled material product passes through first shooting area, setting exists
The image sensitive chip of first shooting area is shot, and picture signal is exported；
(b) scan and pre-process described image signal, in pretreatment stage, the next bat of waiting of doubtful defective problem
It takes the photograph corresponding described image signal in region and carries out defect diagonsis, there is no problem is then not necessarily to carry out the operation of next shooting area,
Export qualifying signal；
(c) when corresponding coiled material product passes through second shooting area, the image sense in the shooting area is set
Optical chip is shot, and picture signal is exported, and described image signal passes through the detection in defect diagonsis stage, belongs to known defect
When problem, then current flaw indication is exported, when belonging to unknown defect problem, is then stored in present case database.
Further, further comprising the steps of among step (b):
(b1) in pretreatment stage, described image signal is scanned；
(b2) picture signal after scanning, is split regional processing, described image signal is divided into multiple detection zones；
(b3) detection zone after segmentation is detected one by one, when a problem occurs, then exports detection signal to carry out
The operation of downstream；It does not go wrong, then exports qualifying signal without carrying out subsequent operation.
Further, further comprising the steps of among step (c):
(c1) in second shooting area, interference operation is carried out after removal interference to picture signal and defect occurs
Problem enters next step, and the then output qualifying signal not gone wrong need not carry out subsequent operation；
(c2) feature extraction is carried out to the picture signal for defect problem occur, and then the defect characteristic of extraction and the case
The defects of example database is compared, and compares the output met currently corresponding flaw indication, compares incongruent deposit and work as
In preceding case database.
It is further comprising the steps of among step (c):
(c0) high frequency started when the light supply apparatus is changed to and is shot with the LED light source being always on from the LED light source being always on
Flash lamp.
In conjunction with attached drawing, the embodiments of the present invention are described in detail above, but the present invention is not limited to described implementations
Mode.For a person skilled in the art, in the case where not departing from the principle of the invention and spirit, to these embodiments
A variety of change, modification, replacement and modification are carried out, are still fallen in protection scope of the present invention.
Claims (9)
1. a kind of defect classification learning system based on artificial intelligence characterized by comprising
One camera apparatus is provided with image sense wherein there are two shooting areas for camera apparatus tool in each shooting area
Optical chip；
One light supply apparatus；
One detection device, the detection device include a roll-in component and a conveyer belt, and the roll-in component drives the transmission
Band is recycled, and the top of the conveyer belt is arranged in the camera apparatus, and the light supply apparatus is arranged in the transmission
The lower section of band is to correspond to shooting area；And
One image processing unit, described image processing unit communicates to connect the camera apparatus, in detection image sensitive chip
Captured picture signal, and control the camera apparatus and shoot same coiled material surface in different shooting areas, and combine
Detection content completes the defects detection for coiled material surface twice.
2. defect classification learning system according to claim 1, which is characterized in that described image processing unit includes sweeping
Module, a preprocessing module, a characteristic extracting module and a comparison module are retouched, the scan module communicates to connect the camera dress
It sets to obtain and scan described image signal, the preprocessing module communicates to connect the scan module to obtain the figure after scanning
It as signal and is pre-processed, the characteristic extracting module communicates to connect the preprocessing module and the comparison module, with right
Image module carries out defects detection.
3. defect classification learning system according to claim 2, which is characterized in that described image processing unit further includes one
Case database, the case database communicate to connect the comparison module, to complete the comparison of case data, and when described
It is not belonging to the defects of described case database in comparison module, then is stored in the case database to wait name.
4. defect classification learning system according to claim 3, which is characterized in that the light supply apparatus includes multiple LED
Light source and at least a high frequency flashing light lamp, the LED light source is separately positioned on position corresponding in different shooting areas, described
High frequency flashing light lamp is then disposed therein in a shooting area, to complete defects detection.
5. defect classification learning system according to claim 4, which is characterized in that the roll-in component further includes a coding
Device, the encoder control the transmission speed of the roll-in component.
6. a kind of classification method based on artificial intelligence, which comprises the following steps:
(a) start the defect classification learning system, when the coiled material product passes through first shooting area, be arranged first
The image sensitive chip of a shooting area is shot, and picture signal is exported；
(b) scan and pre-process described image signal, in pretreatment stage, the next shooting area of the waiting of doubtful defective problem
Corresponding described image signal carries out defect diagonsis in domain, and there is no problem is then not necessarily to carry out the operation of next shooting area, output
Qualifying signal；
(c) when corresponding coiled material product passes through second shooting area, the image sensitive core in the shooting area is set
Piece is shot, and picture signal is exported, and described image signal passes through the detection in defect diagonsis stage, belongs to known defect problem
When, then current flaw indication is exported, when belonging to unknown defect problem, is then stored in present case database.
7. classification method according to claim 6, which is characterized in that further comprising the steps of among step (b):
(b1) in pretreatment stage, described image signal is scanned；
(b2) picture signal after scanning, is split regional processing, described image signal is divided into multiple detection zones；
(b3) detection zone after segmentation is detected one by one, when a problem occurs, then it is next to carry out exports detection signal
The operation of process；It does not go wrong, then exports qualifying signal without carrying out subsequent operation.
8. classification method according to claim 7, which is characterized in that further comprising the steps of among step (c):
(c1) in second shooting area, interference operation is carried out after removal interference to picture signal and defect problem occurs
Enter next step, the then output qualifying signal not gone wrong need not carry out subsequent operation；
(c2) feature extraction is carried out to the picture signal for defect problem occur, and then the defect characteristic of extraction and the case number of cases
It is compared according to the defects of library, compares the output met currently corresponding flaw indication, compare incongruent current case of deposit
In example database.
9. classification method according to claim 8, which is characterized in that further comprising the steps of among step (c):
(c0) high frequency flashing light started when the light supply apparatus is changed to and is shot with the LED light source being always on from the LED light source being always on
Lamp.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910755707.6A CN110517233A (en) | 2019-08-15 | 2019-08-15 | A kind of defect classification learning system and its classification method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910755707.6A CN110517233A (en) | 2019-08-15 | 2019-08-15 | A kind of defect classification learning system and its classification method based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110517233A true CN110517233A (en) | 2019-11-29 |
Family
ID=68626244
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910755707.6A Pending CN110517233A (en) | 2019-08-15 | 2019-08-15 | A kind of defect classification learning system and its classification method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110517233A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111707614A (en) * | 2020-04-13 | 2020-09-25 | 深圳市正宇兴电子有限公司 | Optical chip surface contamination defect detection method and linear laser vision detection system |
| CN111855576A (en) * | 2020-08-07 | 2020-10-30 | 济南海马机械设计有限公司 | Coiled material detection system and detection method |
| CN113077430A (en) * | 2021-03-30 | 2021-07-06 | 太原理工大学 | Laser chip defect detection and classification method and system based on SSD algorithm |
| CN114565609A (en) * | 2022-04-27 | 2022-05-31 | 河南银金达新材料股份有限公司 | On-line detection method for optical performance of photochromic film |
| CN114700290A (en) * | 2022-03-29 | 2022-07-05 | 中建材苏州防水研究院有限公司 | Device for cleaning and marking surface layer of high-molecular waterproof coiled material |
| CN115082445A (en) * | 2022-07-25 | 2022-09-20 | 山东鲁泰防水科技有限公司 | Method for detecting surface defects of building waterproof roll |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101819162A (en) * | 2010-05-13 | 2010-09-01 | 山东大学 | Empty bottle wall defect detection method and device |
| CN101996405A (en) * | 2010-08-30 | 2011-03-30 | 中国科学院计算技术研究所 | Method and device for rapidly detecting and classifying defects of glass image |
| CN202631443U (en) * | 2012-05-04 | 2012-12-26 | 南京文采科技有限责任公司 | Cloth inspection detecting device |
| CN103913460A (en) * | 2013-01-04 | 2014-07-09 | 北京兆维电子(集团)有限责任公司 | Online paper defect detecting system |
| CN104949990A (en) * | 2013-11-01 | 2015-09-30 | 南通大学 | Online detecting method suitable for defects of woven textiles |
| CN105044122A (en) * | 2015-08-25 | 2015-11-11 | 安徽工业大学 | Copper part surface defect visual inspection system and inspection method based on semi-supervised learning model |
| CN105158268A (en) * | 2015-09-21 | 2015-12-16 | 武汉理工大学 | Intelligent online detection method, system and device for defects of fine-blanked parts |
| CN105447512A (en) * | 2015-11-13 | 2016-03-30 | 中国科学院自动化研究所 | Coarse-fine optical surface defect detection method and coarse-fine optical surface defect detection device |
| CN106780473A (en) * | 2016-12-23 | 2017-05-31 | 西安交通大学 | A kind of magnet ring defect multi-vision visual detection method and system |
| CN108037138A (en) * | 2017-12-23 | 2018-05-15 | 陕西科技大学 | A kind of web inspection system and detection method for being used to detect the two-sided defect of paper |
| CN108320282A (en) * | 2018-01-25 | 2018-07-24 | 陕西科技大学 | The ceramic wall and floor bricks surface defect detection apparatus and its method of multicharacteristic information fusion |
| CN108341230A (en) * | 2018-01-12 | 2018-07-31 | 中国林业科学研究院木材工业研究所 | A kind of wood-based plate surface defect on-line measuring device based on synchronization principles |
| CN109030495A (en) * | 2018-06-26 | 2018-12-18 | 大连鉴影光学科技有限公司 | A kind of optical element defect inspection method based on machine vision technique |
| CN109916910A (en) * | 2019-03-27 | 2019-06-21 | 中建材凯盛机器人（上海）有限公司 | Photovoltaic glass edge fault detection system and corresponding method |
-
2019
- 2019-08-15 CN CN201910755707.6A patent/CN110517233A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101819162A (en) * | 2010-05-13 | 2010-09-01 | 山东大学 | Empty bottle wall defect detection method and device |
| CN101996405A (en) * | 2010-08-30 | 2011-03-30 | 中国科学院计算技术研究所 | Method and device for rapidly detecting and classifying defects of glass image |
| CN202631443U (en) * | 2012-05-04 | 2012-12-26 | 南京文采科技有限责任公司 | Cloth inspection detecting device |
| CN103913460A (en) * | 2013-01-04 | 2014-07-09 | 北京兆维电子(集团)有限责任公司 | Online paper defect detecting system |
| CN104949990A (en) * | 2013-11-01 | 2015-09-30 | 南通大学 | Online detecting method suitable for defects of woven textiles |
| CN105044122A (en) * | 2015-08-25 | 2015-11-11 | 安徽工业大学 | Copper part surface defect visual inspection system and inspection method based on semi-supervised learning model |
| CN105158268A (en) * | 2015-09-21 | 2015-12-16 | 武汉理工大学 | Intelligent online detection method, system and device for defects of fine-blanked parts |
| CN105447512A (en) * | 2015-11-13 | 2016-03-30 | 中国科学院自动化研究所 | Coarse-fine optical surface defect detection method and coarse-fine optical surface defect detection device |
| CN106780473A (en) * | 2016-12-23 | 2017-05-31 | 西安交通大学 | A kind of magnet ring defect multi-vision visual detection method and system |
| CN108037138A (en) * | 2017-12-23 | 2018-05-15 | 陕西科技大学 | A kind of web inspection system and detection method for being used to detect the two-sided defect of paper |
| CN108341230A (en) * | 2018-01-12 | 2018-07-31 | 中国林业科学研究院木材工业研究所 | A kind of wood-based plate surface defect on-line measuring device based on synchronization principles |
| CN108320282A (en) * | 2018-01-25 | 2018-07-24 | 陕西科技大学 | The ceramic wall and floor bricks surface defect detection apparatus and its method of multicharacteristic information fusion |
| CN109030495A (en) * | 2018-06-26 | 2018-12-18 | 大连鉴影光学科技有限公司 | A kind of optical element defect inspection method based on machine vision technique |
| CN109916910A (en) * | 2019-03-27 | 2019-06-21 | 中建材凯盛机器人（上海）有限公司 | Photovoltaic glass edge fault detection system and corresponding method |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111707614A (en) * | 2020-04-13 | 2020-09-25 | 深圳市正宇兴电子有限公司 | Optical chip surface contamination defect detection method and linear laser vision detection system |
| CN111707614B (en) * | 2020-04-13 | 2023-08-18 | 深圳市正宇兴电子有限公司 | Optical chip surface contamination defect detection method and line laser vision detection system |
| CN111855576A (en) * | 2020-08-07 | 2020-10-30 | 济南海马机械设计有限公司 | Coiled material detection system and detection method |
| CN111855576B (en) * | 2020-08-07 | 2023-08-18 | 深圳市玻尔智造科技有限公司 | Coiled material detection system and detection method |
| CN113077430A (en) * | 2021-03-30 | 2021-07-06 | 太原理工大学 | Laser chip defect detection and classification method and system based on SSD algorithm |
| CN114700290A (en) * | 2022-03-29 | 2022-07-05 | 中建材苏州防水研究院有限公司 | Device for cleaning and marking surface layer of high-molecular waterproof coiled material |
| CN114700290B (en) * | 2022-03-29 | 2024-03-12 | 中建材苏州防水研究院有限公司 | Device for cleaning and marking surface layer of high-molecular waterproof coiled material |
| CN114565609A (en) * | 2022-04-27 | 2022-05-31 | 河南银金达新材料股份有限公司 | On-line detection method for optical performance of photochromic film |
| CN114565609B (en) * | 2022-04-27 | 2022-07-15 | 河南银金达新材料股份有限公司 | On-line detection method for optical performance of photochromic film |
| CN115082445A (en) * | 2022-07-25 | 2022-09-20 | 山东鲁泰防水科技有限公司 | Method for detecting surface defects of building waterproof roll |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110517233A (en) | A kind of defect classification learning system and its classification method based on artificial intelligence | |
| CN109724984B (en) | Defect detection and identification device and method based on deep learning algorithm | |
| CN111862064B (en) | Silver wire surface flaw identification method based on deep learning | |
| CN110346377A (en) | Nonwoven surface detection system and its detection method based on machine vision | |
| CN106802215A (en) | A kind of device for detecting water leakage of water pipe and detection method | |
| CN102175692A (en) | System and method for detecting defects of fabric gray cloth quickly | |
| CN205538710U (en) | Inductance quality automatic check out system based on machine vision | |
| CN105388162A (en) | Raw material silicon wafer surface scratch detection method based on machine vision | |
| CN1485616A (en) | Fowl eggs quality non-destruction automatic detection grading apparatus and process | |
| CN106290392A (en) | A kind of little micro-bearing surface pitting defects online test method and system thereof | |
| CN108711148A (en) | A kind of wheel tyre defect intelligent detecting method based on deep learning | |
| CN111369516A (en) | Transformer bushing heating defect detection method based on infrared image recognition | |
| CN104483320A (en) | Digitized defect detection device and detection method of industrial denitration catalyst | |
| CN110756457A (en) | Automatic wall and floor tile identifying and classifying system and method | |
| CN210071686U (en) | Fruit grading plant based on orthogonal binocular machine vision | |
| CN106706656A (en) | Machine vision-based zipper detection device and method | |
| CN106248680A (en) | A kind of engine commutator quality detecting system based on machine vision and detection method | |
| CN103785622A (en) | Machine-vision-based part sorting device and method | |
| CN113894055A (en) | Hardware surface defect detection and classification system and method based on machine vision | |
| CN112268910A (en) | Visual detection system and method for cigarette appearance defects | |
| CN115060742A (en) | Printed circuit board defect detection system and method based on visual edge calculation | |
| CN113916127A (en) | Visual inspection system and method for appearance of valve guide pipe finished product | |
| CN114494103A (en) | Defect detection method and device | |
| CN115170501A (en) | Defect detection method, system, electronic device and storage medium | |
| CN103364413A (en) | Identifier detection and product grading control system and method |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191129 |