CN106570511A - Keyboard character defect online detection system based on computer vision and keyboard character defect online detection method thereof - Google Patents
Keyboard character defect online detection system based on computer vision and keyboard character defect online detection method thereof Download PDFInfo
- Publication number
- CN106570511A CN106570511A CN201610943910.2A CN201610943910A CN106570511A CN 106570511 A CN106570511 A CN 106570511A CN 201610943910 A CN201610943910 A CN 201610943910A CN 106570511 A CN106570511 A CN 106570511A
- Authority
- CN
- China
- Prior art keywords
- character
- keyboard
- image
- computer vision
- defect
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/243—Aligning, centring, orientation detection or correction of the image by compensating for image skew or non-uniform image deformations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/06—Recognition of objects for industrial automation
Abstract
The invention relates to a keyboard character online detection system based on computer vision and a keyboard character online detection method thereof. The system comprises an industrial camera, a computer, a single-chip microcomputer system, a frequency converter, a motor and a conveying belt. The method comprises the operation steps that (1) the image quality is enhanced by using an image gray processing and median filtering method; (2) image matching is performed by using surf feature points; (3) the character positions are determined; (4) sequence numbering is performed on the characters; and (5) defect character separation is performed. An SVM classification method is used by the detection algorithm, the normal characters are marked as positive samples, the wrong characters are marked as negative samples, and the size, the center and the moment features of the samples are extracted to be trained. The system is easy to install and high in detection efficiency and plays an important role in the keyboard production process.
Description
Technical field
The present invention relates to a kind of method based on support vector machine, a kind of keyboard word based on computer vision of specific design
Symbol defect on-line detecting system and method, detect to keyboard defect, by being automatically positioned keypad character, extract the big of them
Little, center of gravity, girth, the feature such as square is trained, and realizes the detection to character defect, construct complete set by computer,
Industrial camera, light source, the keypad character on-line detecting system of the composition of single-chip microcomputer and method.
Background technology
In keyboard production process, need that button on keyboard is printed and installed, but in printing and installation process
In, it is easy to occur printing dirty, the phenomenon such as setup error.For the quality of the keyboard production that is rigid in checking up, monitor in real time is needed
The quality of leaflet on keyboard production line, in time removes underproof product.
Past, printing enterprise was all completing the quality testing of product with control, a veteran technology by workman
Workman is exactly a detecting system, but manual detection does not ensure that steady quality, usually occurs that flase drop, missing inspection show
As, and manual detection speed is too slow, cost is too high, examination criteria differs.Under the pressure of the shortcoming and defect of traditional detection, seek one
Plant the trend of the times that objective efficient keyboard printing defects online measuring technique is printing industry development.
At present, computer vision quickly grows, and with visual processes technology defects detection is done, and highlights in industrial quarters
Big advantage, reduces production cost, improves production efficiency.Therefore computer vision technique is applied in the inspection of keyboard defect
Big advantage will be played in survey.
The content of the invention
Present invention aims to the defect that prior art is present, there is provided a kind of keyboard based on computer vision lacks
Sunken on-line detecting system and method, by it in combination with automation control system, the keyboard defect for having built complete set is online
Detecting system and method, improve the efficiency in production process, it is ensured that the quality of production.
To reach above-mentioned purpose, idea of the invention is that:
1. the detecting system with computer system as core is set up, the letter of transmission tape running speed will be controlled by computer
Single-chip microcomputer number is sent to, single-chip microcomputer is converted into analogue signal and gives frequency changer through D/A, and then controls the operation speed of conveyer belt
Degree, simultaneous computer controls the exposure time interval of CCD camera.
2. image capturing system, by internetwork connection mode the image that industry camera is collected is obtained.
3. Image semantic classification, to the picture that shoots, we need to carry out gray processing, binaryzation and noise reduction, remove unrelated letter
Breath, it is ensured that substrate processing time.
4. keyboard is extracted, and the picture for photographing finds position of the keyboard in picture by template matching method, is easy to the later stage
Keyboard is processed.
5. keycap is extracted, and to keyboard by binary conversion treatment, the character being partitioned on keyboard, facilitates later stage character to contrast.
6. slant correction, keyboard does not ensure that on a production line its each position, all in horizontality, may be located
In heeling condition, by carrying out Feature Points Matching with template picture, obtain its transformation matrix, so can by keyboard transform to
Template picture position identical state, improves the accuracy of later stage contrast.
7. feature extraction, by extracting character girth on keyboard, area, center of gravity, moment characteristics, as the spy of character on keyboard
Levy.
8. Defect Comparison, the feature of said extracted is input in support vector machine classifier and is trained, and obtains one
The grader for training, when detection when, by it is to be checked measure picture be input to grader just can be by defect keyboard detection out.
Conceived according to foregoing invention, the present invention adopts following technical proposals, a kind of keypad character based on computer vision
Defect on-line detecting system, including industry camera (1), computer (2), Single Chip Microcomputer (SCM) system (3), converter (4), motor (5)
With conveyer belt (6), it is characterised in that:Keyboard (7) on the industry camera (1) alignment conveyer belt (6) shoots, and it is exported
It is connected to computer (2), computer (2) connection Single Chip Microcomputer (SCM) system (3), output Jing converters (4) the connection motor of single-chip microcomputer (3)
(5), motor (5) drives conveyer belt (6) to carry out;Keypad character image is obtained using industry camera (1), using computer vision
Svm classifier method detected image deflects, judges the quality of production of keyboard in process of production.Comprise the following steps that:
1. using computer (2) and single-chip microcomputer msp430 serial communications, the speed of conveyer belt (6) is controlled;
2. image gray processing, the method for medium filtering is used to improve the quality of image;
3. images match is carried out using surf characteristic points, inclined image rotation is same into being in template picture
One level;
4. contours extract is carried out to character using eight adjacent methods, by obtaining its profile, calculates these character outlines
Minimum enclosed rectangle, determine the position of character;
5. the coordinate of boundary rectangle is calculated, order label is carried out to character;
6. the size of character is extracted, center, moment characteristics are trained using SVM classifier, by normal key
Disk character and defect character separation.
Conceived according to foregoing invention, the present invention adopts following technical proposals:
A kind of keypad character defect on-line detecting system based on computer vision, including industry camera, computer, list
Piece machine system, converter, motor and conveyer belt, it is characterised in that:Keyboard on the industry camera alignment conveyer belt shoots,
And its output is connected to computer, computer connection Single Chip Microcomputer (SCM) system, the output Jing converters connection motor of single-chip microcomputer, motor belt motor
Dynamic conveyer belt operation;Keypad character image is obtained using industry camera, is lacked image using computer vision svm classifier method
It is trapped into going detection, judges the quality of production of keyboard in process of production;Using computer and single-chip microcomputer msp430 serial communications, control
The speed of conveyer belt processed.
A kind of keypad character defect online detection method based on computer vision, is operated using said system, its
It is characterised by that concrete operation step is as follows:
1. image gray processing, the method for medium filtering is used to improve the quality of image;
2. images match is carried out using surf characteristic points, inclined image rotation is in into same water into template picture
It is flat;
3. contours extract is carried out to character using eight adjacent methods, by obtaining its profile, calculates these character outlines most
Little boundary rectangle, determines the position of character;
4. the coordinate of boundary rectangle is calculated, order label is carried out to character;
5. extract the size of character, center, moment characteristics are trained using SVM classifier, by normal keyboard character with lack
Sunken character separation.
The step (2) carries out images match using surf characteristic points, and its purpose is in order to by the inclination picture on conveyer belt
The angle of template picture is rotated into, concrete grammar is as follows:
1. respectively to template image and conveyer belt photographs to keyboard picture extract its surf feature, the spy for obtaining respectively
It is a little one 64 vector tieed up to levy.
2. the surf characteristic points on two width images are traveled through respectively, their Euclidean distance is calculated, and find the minimum point of distance,
It is exactly the good point of two width images match, using the point for matching, obtains single mapping matrix H.
3. using single mapping matrix H, the keyboard picture on conveyer belt is carried out into affine change, obtains the image for having corrected.
The step (5) is trained using SVM classifier, and it is positive sample that 2000 normal picture indicias are taken respectively,
2000 defective picture indicias are negative sample, and their size, center of gravity, girth, moment characteristics, as SVM's are extracted respectively
Input, is trained, and using the SVM classifier for obtaining defects detection is carried out.
The present invention compared with prior art, enters with following obvious prominent substantive distinguishing features and notable technology
Step:
1. when image rectification is carried out, characteristic matching is carried out using surf features, obtain picture to be measured with template picture
Mapping relations, using surf features ensure it is stable in the case of, real-time also can be guaranteed, can be quick, it is accurately complete
Into the correction of testing image.
2. in order to obtain the position of testing image and template image each character, herein by Mathematical Morphology method and eight neighborhood
Contour detecting method, obtains the character zone of primary segmentation.And using the sort algorithm for designing, obtain region of interest
Domain, meeting has the index of a determination on a button, it is ensured that detection every time can be by the character of accurately contrast same position.
3. when character is contrasted, by the size for extracting character, center, moment characteristics can effectively represent the spy of keypad character
Levy, and be trained by using SVM classifier, it is not enough to successfully manage sample, most normal keyboard character and defect at last
Character separation.
Description of the drawings
Fig. 1 is overall structure diagram of the present invention.
Fig. 2 is global procedures block diagram of the present invention.
Fig. 3 is image correction program block diagram.
Fig. 4 is character outline tracing program block diagram.
Fig. 5 is the flow chart of software in the embodiment of the present invention three.
Fig. 6 is to build SVM classifier flow chart.
Fig. 7 is present invention correction matching schematic diagram.
Fig. 8 is the schematic diagram after correction.
Fig. 9 is the schematic diagram of Morphological scale-space.
Figure 10 is character locating schematic diagram.
Figure 11 is character sequence schematic diagram.
Specific embodiment
The preferred embodiments of the present invention combine accompanying drawing it is shallow state it is as follows:
Embodiment one:
Referring to Fig. 1, based on the keypad character defect on-line detecting system of computer vision, including industry camera (1), meter
Calculation machine (2), Single Chip Microcomputer (SCM) system (3), converter (4), motor (5) and conveyer belt (6), it is characterised in that:The industry camera
(1) keyboard (7) being aligned on conveyer belt (6) shoots, and its output is connected to computer (2), computer (2) connection single-chip microcomputer system
System (3), output Jing converters (4) connection motor (5) of single-chip microcomputer (3), motor (5) drives conveyer belt (6) operation;Using industry
Photographic head (1) obtains keypad character image, is detected image deflects using computer vision svm classifier method, judges
The quality of production of keyboard in production process;Using computer (2) and single-chip microcomputer msp430 serial communications, conveyer belt (6) is controlled
Speed.
Embodiment two:
Referring to Fig. 2, Figure 11, based on the keypad character defect on-line checking detection method of computer vision, using according to power
Profit requires being operated based on the keypad character defect on-line detecting system of computer vision described in 1, it is characterised in that concrete
Operating procedure is as follows:
6. image gray processing, the method for medium filtering is used to improve the quality of image;
7. images match is carried out using surf characteristic points, inclined image rotation is in into same water into template picture
It is flat;
8. contours extract is carried out to character using eight adjacent methods, by obtaining its profile, calculates these character outlines most
Little boundary rectangle, determines the position of character;
9. the coordinate of boundary rectangle is calculated, order label is carried out to character;
The size of character is extracted, center, moment characteristics are trained using SVM classifier, by normal keyboard character and defect
Character separation.
The described keypad character defect online detection method based on computer vision, it is characterised in that the step (2)
Carry out images match using surf characteristic points, its purpose in order to will on conveyer belt incline picture rotation into template picture angle
Degree, concrete grammar is as follows:
4. respectively to template image and conveyer belt photographs to keyboard picture extract its surf feature, the spy for obtaining respectively
It is a little one 64 vector tieed up to levy.
5. the surf characteristic points on two width images are traveled through respectively, their Euclidean distance is calculated, and find the minimum point of distance,
It is exactly the good point of two width images match, using the point for matching, obtains single mapping matrix H.
6. using single mapping matrix H, the keyboard picture on conveyer belt is carried out into affine change, obtains the image for having corrected.
The step (5) is trained using SVM classifier, and it is positive sample that 2000 normal picture indicias are taken respectively,
2000 defective picture indicias are negative sample, and their size, center of gravity, girth, moment characteristics, as SVM's are extracted respectively
Input, is trained, and using the SVM classifier for obtaining defects detection is carried out.
Embodiment three:
Not only Detection accuracy is high but also detection speed is fast for the method for keyboard defect on-line checking based on computer vision.
The present embodiment can be divided into two stages of hardware designs and software design, and detailed step is as follows:
1. hardware designs (referring to Fig. 1)
Fig. 1 for the system a hardware frame, detecting system mainly by a computer, msp430 single-chip microcomputers, one
The industry camera composition of 5000000 pixels.Because CCD camera and conveyer belt are all to work independently, it must be determined that CCD camera
Time of exposure so that every time camera can photograph on keyboard, so time of exposure must be calculated.
If the distance between two neighboring keyboard is Δ s, transmission tape running speed is v, and time of exposure is Δ t, by Δ s=
Δ t × v can show that the time of exposure of camera, so only needing computer and msp430 microcomputer series interface communications, sends control life
Order, it is ensured that keyboard can be shot complete by camera every time.
In order to control to transmit tape running speed v, single-chip microcomputer is converted into analogue signal and send a frequency changer, enters by D/A modules
And control speed of service v of conveyer belt.
2. software design (referring to Fig. 5)
2.1 Image semantic classification
(1) gray proces
Mainly cromogram is transformed into into gray-scale maps, coloured picture is made up of each RGB channel, and gray proces are mainly by the every of coloured picture
Individual pixel rgb value is calculated as gray value.Formula is as follows:
Gray=0.11R+0.59G+0.30B (2.1)
(2) Filtering Processing
The purpose of image filtering is that, in order to reduce the noise of keyboard image, these noises are from many-sided reason.In keyboard
During production, due to reflective stronger, these local images easily whiten, so as to be formed for keyboard recess and corner
It is miscellaneous, the normal cutting procedure of interferogram picture, it is necessary to filtered.We are carried out respectively using mean filter and medium filtering
Experiment, wave filter size adopts 3*3,5*5,7*7.Can effectively be filtered using the medium filtering mode of 7*7 through Experimental comparison
Miscellaneous point in except image.
(3) binaryzation
Binary image can in the background project object pixel so that the contrast of image strengthens, moreover it is possible to reduce follow-up
The complexity of process and time.Because illumination condition is almost fixation during keyboard detection, so in conversion process,
What a threshold value we only need to arrange, and the pixel of the point is 1 more than this threshold value, is 0 less than this threshold value.Binaryzation formula is such as
Under:
2.2 image slant corrections
When producing on keyboard production line, do not ensure that their position is horizontally oriented forever, can cause to incline, such as
Fruit angle of inclination is larger, then can have a strong impact on the identification to defect character, so we will be according to the practical situation of production to figure
As being corrected.We are corrected using the matching algorithm based on surf, and matching is as shown in Fig. 2 the figure such as Fig. 3 institutes after correction
Show.
If p=is (x1,y1) coordinate of certain point, q=(x in image to be corrected2,y2) be correction after image midpoint coordinate,
Then there is affine transformation relationship:
Then need to obtain affine matrix, make the affine matrix be:
In order to obtain above-mentioned affine matrix, need to obtain the p in image, the position of q, extract respectively image to be rotated and
The surf features of template image, the matching that the method matched with arest neighbors obtains two width images is right, in conjunction with RANSAC and minimum
Square law, obtains the affine matrix between image.
2.3 keycaps are extracted
In order to contrast whether keypad character mistake occurs, it would be desirable to which keypad character is extracted from image.Because
A very big part is unrelated background in keyboard image, and carrying out defect character recognition to us does not have any help.
(1) Morphological scale-space
A region interested is only existed in order to allow on a button, it would be desirable at morphologic method
Reason.Its basic thought is to go to measure and extract the correspondingly-shaped in figure using the structural element with certain form, to reach
Graphical analyses and the purpose of identification.Its elementary operation has 4:They are carried out group by expansion, burn into opening operation and closed operation
Conjunction can reach and a region interested is only existed on a button.We pass through the such a processing stream of experimental design
Journey:Expansion, closure, opening operation, expansion.As a result as shown in figure 4, thus obtaining area-of-interest, while ensure that same
Only one of which area-of-interest on button.
(2) contours extract
By morphologic method, the size of character can be primarily determined that, but the position of character zone can't be determined.
By the profile for extracting region, their minimum enclosed rectangle is asked to be assured that the position of character, minimum enclosed rectangle is surrounded
Region be exactly area-of-interest.
Here we take eight adjacent methods to carry out the extraction of profile.Algorithm is as follows:For any one pixel P, its
8 pixels of surrounding are respectively P1, P2, P3, P4, P5, P6, P7, P8.On bianry image, from P points, search for around it
8 pixels whether be 1, if all 1, P points are not profiles, not all 1, then P points be profile, now from P point up times
Pin sets out, and finds the pixel for 1, determines whether profile, and always recurrence is gone down, until entire image it is searched complete, so
Just can be by the contours extract of target area out.
(3) character sequence
By the method for contours extract, the position in picture of each character of keyboard is obtained, as shown in Figure 5.But
What the order of these positions was not to determine, such as character ' A ' not can determine which position this character is, so we
The character position for extracting must be ranked up, determine the position of each character.Sort algorithm step is described as follows:
A. travel through all of character outline set, choose first character in set, calculate it minimum external square
The vertical coordinate of shape, continues to travel through remaining character outline, and whether judge the vertical coordinate of their minimum enclosed rectangles has intersecting.If sat
Mark scope has intersecting, then these characters are in same a line on keyboard, thus the rower of character can be determined.
B. to each character row, the center of calculating character is ranked up according to the abscissa of its center, completes
Sequence from left to right.Each character is set to have a label to be determined, as a result as shown in Figure 6.
2.4 character features are extracted
Image characteristics extraction is that the result figure after the completion of image procossing is extracted into alphabetical feature, the first foundation of feature
There is certain principle, be characterised by can be good at description letter.For keypad character only can not be effective using a feature
The difference for finding out character, so multiple features must be extracted to achieve the goal.
(1) character boundary
In character defects detection, when defect occurs in character, such as character is long or shorter, then can pass through character
Length and height are reflected.The profile of character is obtained by contours extract algorithm, the minimum enclosed rectangle of profile is obtained, then it is minimum
The length and width of boundary rectangle are exactly the size of character.
(2) character center of gravity
Alphabetical center of gravity be in binary picture own ' 1 ' a statistical value, each character correspondence position gray scale
Value regards its quality in the position as, it is possible to their oeverall quality and position of centre of gravity is calculated, if character is defective
Words, can pass through the reacting condition of its position of centre of gravity out.Assume piece image M*N binary image, f (x, y)=0 or
F (x, y)=1, if pixel number is k in region, then institute position a little and average as focus point, formula in statistics set P
It is as follows:
(3) character moment characteristics
In view of keyboard production process can cause testing image and template image in the region of interest misalignment of pixels or
The problem that person's scale size differs, in order to obtain character translation invariant shape, rotational invariance and constant rate, using character
Central moment can be with this problem of effectively solving.For width digital picture I (x, y), the central moment of the p+q of region R is:
Although Central Moment Feature will not change after image occurs translation and change of scale, after rotation still
Can change, so must be completed by combination center square.These central moment combinations are as follows:
H1=μ20+μ02 (2.7)
H2=(μ20-μ02)2+4μ11 2 (2.8)
H3=(μ30-3μ12)2+(3μ21-μ03)2 (2.9)
H4=(μ30+μ12)2+(μ21+μ03)2 (2.10)
H6=(μ20-μ02)·[(μ30+μ12)2-(μ21+μ03)2]+4μ11·(μ30+μ12)·(μ21+μ03) (2.12)
When character occurs defect, bending moment will numerically not produce difference in Hu for testing image and template image, lead to
Experiment is crossed, defect character H is found1And H2Differ greatly with normal character, therefore we only use H1And H2As feature.
2.5 support vector machine classifiers are designed
Support vector machine are built upon on the VC of Statistical Learning Theory dimensions theory and Structural risk minization basis, it
It is, to find a kind of efficient machine learning method of the optimal classification surface as target, to solve small sample, non-linear and higher-dimension mould
Formula identification convexity shows many advantages.Therefore SVM is selected to carry out training sample.
Normal keypad character is labeled as into positive sample, defective keypad character is labeled as negative sample, and it is extracted respectively
Size, center of gravity, girth, the feature such as square puts it in SVM classifier and is trained.Finally obtain one to train
Model, then use it for keyboard defects detection.
Claims (6)
1. a kind of keypad character defect on-line detecting system based on computer vision, including industry camera(1), computer
(2), Single Chip Microcomputer (SCM) system(3), converter(4), motor(5)And conveyer belt(6), it is characterised in that:The industry camera(1)It is right
Quasi- conveyer belt(6)On keyboard(7)Shoot, and its output is connected to computer(2), computer(2)Connection Single Chip Microcomputer (SCM) system
(3), single-chip microcomputer(3)Output Jing converters(4)Connection motor(5), motor(5)Drive conveyer belt(6)Operation;Taken the photograph using industry
As head(1)Keypad character image is obtained, is detected image deflects using computer vision svm classifier method, judge giving birth to
The quality of production of keyboard during product;Using computer(2)With single-chip microcomputer msp430 serial communications, conveyer belt is controlled(6)Speed
Degree.
2. a kind of keypad character defect on-line checking detection method based on computer vision, adopts according to claim 1
Operated based on the keypad character defect on-line detecting system of computer vision, it is characterised in that concrete operation step is such as
Under:
1. image gray processing, the method for medium filtering is used to improve the quality of image;
2. images match is carried out using surf characteristic points, inclined image rotation is in into same level into template picture;
3. contours extract is carried out to character using eight adjacent methods, by obtaining its profile, the minimum for calculating these character outlines is outer
Rectangle is connect, the position of character is determined;
4. the coordinate of boundary rectangle is calculated, order label is carried out to character;
5. the size of character is extracted, center, moment characteristics are trained using SVM classifier, by normal keyboard character and defect word
Symbol is separated.
3. the keypad character defect online detection method based on computer vision according to claim 2, it is characterised in that
The step(2)Images match is carried out using surf characteristic points, its purpose is in order to by the inclination picture rotation Cheng Mo on conveyer belt
The angle of plate picture, concrete grammar is as follows:
1. respectively to template image and conveyer belt photographs to keyboard picture extract its surf feature, the characteristic point for obtaining respectively
It is the vector of one 64 dimension;
2. the surf characteristic points on two width images are traveled through respectively, their Euclidean distance is calculated, the minimum point of distance is found, and are exactly
The good point of two width images match, using the point for matching, obtains single mapping matrix H;
3. using single mapping matrix H, the keyboard picture on conveyer belt is carried out into affine change, obtains the image for having corrected.
4. the keypad character defect online detection method based on computer vision according to claim 2, it is characterised in that
The step(3)Contours extract is carried out to character using eight adjacent methods, method is as follows:A foreground pixel is hit every time, is scanned
Position returns to the upper background pixel for finding the pixel, and town of then taking advantage of a situation checks 8 pixels around foreground pixel, until finding
Next foreground pixel, algorithm terminates when starting point pixel is found for the second time, then the foreground pixel passed through in tracking
Just constitute profile.
5. the keypad character defect online detection method based on computer vision according to claim 2, it is characterised in that
The step(4)The coordinate of boundary rectangle is calculated, order label is carried out to character, it is main by all of character outline collection of traversal
Close, choose first character in set, calculate it minimum enclosed rectangle vertical coordinate, continue to travel through remaining character wheel
Exterior feature, whether have intersecting, if coordinate range has intersecting, these characters are in keyboard if judging the vertical coordinate of their minimum enclosed rectangles
It is upper to be in same a line, thus the rower of character can be determined.
6. the keypad character defect online detection method based on computer vision according to claim 2, it is characterised in that
The step(5)It is trained using SVM classifier, it is positive sample that 2000 normal picture indicias are taken respectively, and 2000 have
The picture indicia of defect is negative sample, and their size is extracted respectively, and center of gravity, girth, moment characteristics, as the input of SVM, are carried out
Training, using the SVM classifier for obtaining defects detection is carried out.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610943910.2A CN106570511B (en) | 2016-11-02 | 2016-11-02 | Keyboard character defect online detection system and method based on computer vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610943910.2A CN106570511B (en) | 2016-11-02 | 2016-11-02 | Keyboard character defect online detection system and method based on computer vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106570511A true CN106570511A (en) | 2017-04-19 |
CN106570511B CN106570511B (en) | 2021-04-09 |
Family
ID=58536409
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610943910.2A Active CN106570511B (en) | 2016-11-02 | 2016-11-02 | Keyboard character defect online detection system and method based on computer vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106570511B (en) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107328793A (en) * | 2017-06-30 | 2017-11-07 | 航天新长征大道科技有限公司 | A kind of ornaments surface word print flaw detection method and device based on machine vision |
CN107367514A (en) * | 2017-08-28 | 2017-11-21 | 广东工业大学 | A kind of POS defect inspection method based on machine vision |
CN108416765A (en) * | 2018-01-30 | 2018-08-17 | 华南理工大学 | A kind of character defect automatic testing method and system |
CN108548820A (en) * | 2018-03-28 | 2018-09-18 | 浙江理工大学 | Cosmetics paper labels defect inspection method |
CN108647706A (en) * | 2018-04-24 | 2018-10-12 | 广州大学 | Article identification classification based on machine vision and flaw detection method |
CN109087286A (en) * | 2018-07-17 | 2018-12-25 | 江西财经大学 | A kind of detection method and application based on Computer Image Processing and pattern-recognition |
CN109211136A (en) * | 2018-08-31 | 2019-01-15 | 广州大学 | A kind of Watch glass cover board profile tolerance detection method |
CN109738450A (en) * | 2019-01-09 | 2019-05-10 | 合肥联宝信息技术有限公司 | The detection method and device of keyboard of notebook computer |
CN109871895A (en) * | 2019-02-22 | 2019-06-11 | 北京百度网讯科技有限公司 | The defect inspection method and device of circuit board |
CN110363250A (en) * | 2019-07-23 | 2019-10-22 | 北京隆普智能科技有限公司 | A kind of method and its system of 3-D image intelligent Matching |
CN110596121A (en) * | 2019-09-12 | 2019-12-20 | 南京旷云科技有限公司 | Keyboard appearance detection method and device and electronic system |
CN110658194A (en) * | 2019-09-30 | 2020-01-07 | 联想(北京)有限公司 | Keyboard detection method and keyboard detection equipment |
CN110674876A (en) * | 2019-09-25 | 2020-01-10 | 北京猎户星空科技有限公司 | Character detection method and device, electronic equipment and computer readable medium |
CN110706195A (en) * | 2018-06-22 | 2020-01-17 | 技嘉科技股份有限公司 | Keyboard detection method and system |
CN111553357A (en) * | 2019-12-12 | 2020-08-18 | 电子科技大学中山学院 | Product function durability test self-learning method containing sliding pattern identification |
CN112613337A (en) * | 2020-11-07 | 2021-04-06 | 泰州无印广告传媒有限公司 | Keyboard label missing detection platform and method |
WO2021109011A1 (en) * | 2019-12-04 | 2021-06-10 | 电子科技大学 | Intelligent capacitor internal defect detection method based on ultrasound image |
CN114972825A (en) * | 2022-07-19 | 2022-08-30 | 天津七一二移动通信有限公司 | Method for realizing automatic detection of MMI key icon matrix |
TWI779836B (en) * | 2021-09-15 | 2022-10-01 | 英業達股份有限公司 | Keyboard file verification method based on image processing |
CN116008294A (en) * | 2022-12-13 | 2023-04-25 | 无锡微准科技有限公司 | Key cap surface particle defect detection method based on machine vision |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103134803A (en) * | 2011-11-25 | 2013-06-05 | 西安中科麦特电子技术设备有限公司 | PCB optical detection system control circuit |
CN103323457A (en) * | 2013-05-20 | 2013-09-25 | 中国农业大学 | Fruit appearance defect real-time on-line detection system and detection method |
CN103592310A (en) * | 2013-11-22 | 2014-02-19 | 昆山视杰维光电科技有限公司 | System for detecting keyboard |
CN104568986A (en) * | 2015-01-26 | 2015-04-29 | 中国科学院半导体研究所 | Method for automatically detecting printing defects of remote controller panel based on SURF (Speed-Up Robust Feature) algorithm |
US20150206601A1 (en) * | 2014-01-21 | 2015-07-23 | SK Hynix Inc. | Semiconductor memory device, test control system, and method of operating test control system |
-
2016
- 2016-11-02 CN CN201610943910.2A patent/CN106570511B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103134803A (en) * | 2011-11-25 | 2013-06-05 | 西安中科麦特电子技术设备有限公司 | PCB optical detection system control circuit |
CN103323457A (en) * | 2013-05-20 | 2013-09-25 | 中国农业大学 | Fruit appearance defect real-time on-line detection system and detection method |
CN103592310A (en) * | 2013-11-22 | 2014-02-19 | 昆山视杰维光电科技有限公司 | System for detecting keyboard |
US20150206601A1 (en) * | 2014-01-21 | 2015-07-23 | SK Hynix Inc. | Semiconductor memory device, test control system, and method of operating test control system |
CN104568986A (en) * | 2015-01-26 | 2015-04-29 | 中国科学院半导体研究所 | Method for automatically detecting printing defects of remote controller panel based on SURF (Speed-Up Robust Feature) algorithm |
Non-Patent Citations (1)
Title |
---|
管兴坚: "自动光学检测中的键盘缺陷字符关键检测技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107328793A (en) * | 2017-06-30 | 2017-11-07 | 航天新长征大道科技有限公司 | A kind of ornaments surface word print flaw detection method and device based on machine vision |
CN107367514A (en) * | 2017-08-28 | 2017-11-21 | 广东工业大学 | A kind of POS defect inspection method based on machine vision |
CN108416765A (en) * | 2018-01-30 | 2018-08-17 | 华南理工大学 | A kind of character defect automatic testing method and system |
CN108548820A (en) * | 2018-03-28 | 2018-09-18 | 浙江理工大学 | Cosmetics paper labels defect inspection method |
CN108548820B (en) * | 2018-03-28 | 2023-08-15 | 浙江理工大学 | Cosmetic paper label defect detection method |
CN108647706A (en) * | 2018-04-24 | 2018-10-12 | 广州大学 | Article identification classification based on machine vision and flaw detection method |
CN108647706B (en) * | 2018-04-24 | 2022-03-22 | 广州大学 | Article identification classification and flaw detection method based on machine vision |
CN110706195A (en) * | 2018-06-22 | 2020-01-17 | 技嘉科技股份有限公司 | Keyboard detection method and system |
CN109087286A (en) * | 2018-07-17 | 2018-12-25 | 江西财经大学 | A kind of detection method and application based on Computer Image Processing and pattern-recognition |
CN109211136A (en) * | 2018-08-31 | 2019-01-15 | 广州大学 | A kind of Watch glass cover board profile tolerance detection method |
CN109738450B (en) * | 2019-01-09 | 2021-06-29 | 合肥联宝信息技术有限公司 | Method and device for detecting notebook keyboard |
CN109738450A (en) * | 2019-01-09 | 2019-05-10 | 合肥联宝信息技术有限公司 | The detection method and device of keyboard of notebook computer |
CN109871895A (en) * | 2019-02-22 | 2019-06-11 | 北京百度网讯科技有限公司 | The defect inspection method and device of circuit board |
CN110363250A (en) * | 2019-07-23 | 2019-10-22 | 北京隆普智能科技有限公司 | A kind of method and its system of 3-D image intelligent Matching |
CN110596121A (en) * | 2019-09-12 | 2019-12-20 | 南京旷云科技有限公司 | Keyboard appearance detection method and device and electronic system |
CN110674876A (en) * | 2019-09-25 | 2020-01-10 | 北京猎户星空科技有限公司 | Character detection method and device, electronic equipment and computer readable medium |
CN110658194A (en) * | 2019-09-30 | 2020-01-07 | 联想(北京)有限公司 | Keyboard detection method and keyboard detection equipment |
WO2021109011A1 (en) * | 2019-12-04 | 2021-06-10 | 电子科技大学 | Intelligent capacitor internal defect detection method based on ultrasound image |
CN111553357A (en) * | 2019-12-12 | 2020-08-18 | 电子科技大学中山学院 | Product function durability test self-learning method containing sliding pattern identification |
CN112613337A (en) * | 2020-11-07 | 2021-04-06 | 泰州无印广告传媒有限公司 | Keyboard label missing detection platform and method |
TWI779836B (en) * | 2021-09-15 | 2022-10-01 | 英業達股份有限公司 | Keyboard file verification method based on image processing |
CN114972825A (en) * | 2022-07-19 | 2022-08-30 | 天津七一二移动通信有限公司 | Method for realizing automatic detection of MMI key icon matrix |
CN114972825B (en) * | 2022-07-19 | 2022-12-13 | 天津七一二移动通信有限公司 | Method for realizing automatic detection of MMI key icon matrix |
CN116008294A (en) * | 2022-12-13 | 2023-04-25 | 无锡微准科技有限公司 | Key cap surface particle defect detection method based on machine vision |
CN116008294B (en) * | 2022-12-13 | 2024-03-08 | 无锡微准科技有限公司 | Key cap surface particle defect detection method based on machine vision |
Also Published As
Publication number | Publication date |
---|---|
CN106570511B (en) | 2021-04-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106570511A (en) | Keyboard character defect online detection system based on computer vision and keyboard character defect online detection method thereof | |
CN108982508B (en) | Plastic package IC chip defect detection method based on feature template matching and deep learning | |
Cao et al. | A pixel-level segmentation convolutional neural network based on deep feature fusion for surface defect detection | |
CN108764229B (en) | Water gauge image automatic identification method based on computer vision technology | |
CN102426649B (en) | Simple steel seal digital automatic identification method with high accuracy rate | |
CN111402209B (en) | U-Net-based high-speed railway steel rail damage detection method | |
CN104700092B (en) | A kind of small characters digit recognition method being combined based on template and characteristic matching | |
CN109190623B (en) | Method for identifying brand and model of projector | |
CN107545239A (en) | A kind of deck detection method matched based on Car license recognition with vehicle characteristics | |
CN107067002A (en) | Road licence plate recognition method in a kind of dynamic video | |
CN111650220A (en) | Vision-based image-text defect detection method | |
CN106446952A (en) | Method and apparatus for recognizing score image | |
CN106529532A (en) | License plate identification system based on integral feature channels and gray projection | |
CN112233090A (en) | Film flaw detection method based on improved attention mechanism | |
CN105184225B (en) | A kind of multinational banknote image recognition methods and device | |
CN103761743A (en) | Solid wood floor surface defect detecting method based on image fusion and division | |
US11790517B2 (en) | Subtle defect detection method based on coarse-to-fine strategy | |
WO2015131468A1 (en) | Method and system for estimating fingerprint pose | |
CN106599890A (en) | Transformer substation patrol robot digital type instrument identification algorithm | |
CN112508857B (en) | Aluminum product surface defect detection method based on improved Cascade R-CNN | |
CN107886539A (en) | High class gear visible detection method under a kind of industrial scene | |
CN113870202A (en) | Far-end chip defect detection system based on deep learning technology | |
CN115184380A (en) | Printed circuit board welding spot abnormity detection method based on machine vision | |
WO2021109011A1 (en) | Intelligent capacitor internal defect detection method based on ultrasound image | |
CN110363196B (en) | Method for accurately recognizing characters of inclined text |
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 |