CN105784712B - A kind of automatic cloth inspection method using contact-type image sensor - Google Patents
A kind of automatic cloth inspection method using contact-type image sensor Download PDFInfo
- Publication number
- CN105784712B CN105784712B CN201610112469.3A CN201610112469A CN105784712B CN 105784712 B CN105784712 B CN 105784712B CN 201610112469 A CN201610112469 A CN 201610112469A CN 105784712 B CN105784712 B CN 105784712B
- Authority
- CN
- China
- Prior art keywords
- standard deviation
- denoted
- child window
- fabric
- pixel value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- 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
- 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
Abstract
The present invention relates to a kind of automatic cloth inspection methods using contact-type image sensor, fabric image is acquired using contact-type image sensor, then industrial control computer is transmitted data to by CameraLink interfaces, the industrial control computer is detected fabric image, by the average value for calculating the pixel value of the child window of fabric image and standard picture, the every standard deviation of row pixel value and the sum of the absolute value of difference of standard deviation of each column pixel value, there are whether fault for judgement fabric.The present invention carries out fabric Image Acquisition using CIS sensors, and at low cost, simple in structure, detection method speed used is fast, and precision is high, strong antijamming capability.
Description
Technical field
The invention belongs to textile image processing technology fields, are related to a kind of automatic cloth inspection side using contact-type image sensor
Method.
Background technology
As sensor technology and textile process equip the fast development of intelligence degree, conventionally employed artificial fabric defect
Point detection mode has been difficult to adapt to the needs of current textile industry upgrading.For this purpose, in conjunction with artificial intelligence technology, using machine generation
The fast automatic detecting that collisional transfer is carried out for artificial vision is the new direction and new demand of current development.Application number
CN201010189591.3 discloses a kind of automatic cloth inspection, which carries out fabric image using 4 intelligent CCD cameras
Information collection, then analyze and learn fabric characteristics of image, so that it is determined that in image fault position, then will be seen that fault letter
Breath passes to computer.Though however, the detection efficiency and speed of the fabric that the invention improves to a certain extent, used intelligence
Energy CCD camera is professional precision instrument, expensive, and needs high quality camera lens and special light source to assist fabric image to adopt
Collection.Further, since CCD camera is worked using line scan mode, due to optical energy loss when imaging, during the image collected will appear
Between the dark situation in bright both sides, the defect detection precision seriously affected.Application number 201410201132.0 discloses one kind and is based on connecing
The patent of invention of the collisional transfer detecting system of touch imaging sensor, the patent of invention use contact-type image sensor
(CIS) then array acquisition fabric image sends images to and carries out defects identification and classification in digital signal processor, to
Realize the automatic detection of collisional transfer.Although the invention reduces the cost of image capturing system on certain procedures, do not carry
And specific detection algorithm and precision.
Invention content
Technical problem to be solved by the invention is to provide a kind of automatic cloth inspection methods using contact-type image sensor.
A kind of automatic cloth inspection method using contact-type image sensor of the present invention, fabric image use contact type image
Sensor acquires, and then transmits data to industrial control computer by CameraLink interfaces, the industrial control computer is to fabric
Image is detected, by the average value of calculating fabric image and the pixel value of the child window of standard picture, per row pixel value
The sum of the absolute value of difference of standard deviation of standard deviation and each column pixel value, there are whether fault for judgement fabric.
As preferred technical solution:
A kind of automatic cloth inspection method using contact-type image sensor as described above, the calculating and the judgement
The specific steps are:
Step 1) standard picture is that the parameter of the fabric image A without fault calculates:
1.1) gaussian filtering is carried out to the fabric image A of no fault, is denoted as A1.Gaussian filtering is using the filtering for having weight
Device, and Parameter adjustable can effectively remove the noise in grey cloth image;
1.2) apply 4 filter M1~M4 successively to A1It is filtered, and is added after filter result is taken absolute value, remembered
For A2;Since the direction randomness of fault is strong, the filter of 4 different directions used herein, being respectively used to detection angles is
0 °, 45 °, 90 °, 135 ° of fault.Most of faults can be also detected only with two filters of M1 and M3, but are for angle
45 ° and 135 ° of fault effect is poor, and if the filter using more perspective is detected, detection result improves unknown
It is aobvious, and calculate complicated higher;
1.3) by A2The continuous child window for being divided into p pixels × p pixels without overlapping, value 16≤p≤64;
1.4) A is calculated2In j-th of child window pixel value average value, be denoted as S1(j);
1.5) A is calculated2In the standard deviation of j-th child window per row pixel value, then calculate the standard deviation of often row standard deviation,
It is denoted as S2(j).Preferably, A can first be calculated2In the average value of j-th child window per row pixel value, then calculate often row average value
Standard deviation, the purpose is to the horizontal faults of protrusion.
1.6) A is calculated2In j-th of child window each column pixel value standard deviation, then calculate the standard deviation of each column standard deviation,
It is denoted as S3(j).Preferably, A can first be calculated2In j-th of child window each column pixel value average value, then calculate each column average value
Standard deviation, the purpose is to the vertical faults of protrusion.
1.7) A is calculated2In all child window S1(j)、S2(j) and S3(j) average value, is denoted as S1、S2And S3;
The parameter of step 2) fabric image calculates:
2.1) gaussian filtering is carried out to fabric image B, is denoted as B1, it is denoted as B1.Gaussian filter parameter used herein with
Step 1.1) is equally.
2.2) apply 4 filter M1~M4 to B1It is handled, and is added after handling result is taken absolute value, be denoted as B2;
2.3) by B2The continuous child window for being divided into p pixels × p pixels without overlapping, value 16≤p≤64;
2.4) B is calculated2In j-th of child window pixel value average value, be denoted as T1(j);
2.5) B is calculated2In the standard deviation of j-th child window per row pixel value, then calculate the standard deviation of often row standard deviation,
It is denoted as T2(j);
2.6) B is calculated2In j-th of child window each column pixel value standard deviation, then calculate the standard deviation of each column standard deviation,
It is denoted as T3(j);
Step 3) mathematic interpolation and judgement:
If current j-th of child window meets | T1(j)-S1|+|T2(j)-S2|+|T3(j)-S3| > δ, then it will current sub- window
Mouth is determined with fault;δ values 0.05≤δ≤0.2.
A kind of automatic cloth inspection method using contact-type image sensor as described above, described 4 filter M1~
M4 is:
A kind of automatic cloth inspection method using contact-type image sensor as described above finds fault then by fault information
It is preserved, while fault is marked in fabric by labelling apparatus.
A kind of automatic cloth inspection method using contact-type image sensor as described above, contact-type image sensor length
It it is 2 meters, i.e., maximum detection fabric breadth is 2 meters, 150~300 pixel/inch of optical resolution.
Advantageous effect
1, the cloth inspecting machine that automatic cloth inspection method of the invention uses is at low cost, small;
2, detection method speed is fast, and precision is high;
3, software system development is flexible, easy care.
Description of the drawings
Fig. 1 is the structural schematic diagram for the automatic cloth inspecting machine for realizing the automatic cloth inspection method of the present invention
Fig. 2 is the fabric image that the present invention has fault
Fig. 3 is the fabric image that the present invention has fault
Fig. 4 is detection result figure of the present invention to Fig. 2 fabric images
Fig. 5 is detection result figure of the present invention to Fig. 3 fabric images
Specific implementation mode
The invention will be further elucidated with reference to specific embodiments.It should be understood that these embodiments are merely to illustrate this hair
It is bright rather than limit the scope of the invention.In addition, it should also be understood that, after reading the content taught by the present invention, art technology
Personnel can make various changes or modifications the present invention, and such equivalent forms equally fall within the application the appended claims and limited
Fixed range.
The structural schematic diagram of the automatic cloth inspecting machine of the automatic cloth inspection methods of Fig. 1 to realize the present invention, including fabric yardage roll 1,
CIS sensors 2 (contact-type image sensor), defect marking device 3, industrial control computer 4.CIS sensors 2 are for acquiring fabric
Image, and image is transferred to industrial control computer 4 by CameraLink and is preserved, if detection method finds fault, send out
The number of delivering letters carries out defect position label to defect marking device 3.Below in conjunction in Fig. 2 and Fig. 3 to detection provided by the present invention
Method specific implementation step is illustrated.
Embodiment 1
Step 1
(1) acquisition carries out Gauss filter to image A, is denoted as A corresponding to Fig. 2 without fault fabric image A1, Gauss used
Filter size is 5 × 5, standard deviation 3;
(2) apply 4 filter M1~M4 to A1It is handled, and is added after handling result is taken absolute value, be denoted as A2;
(3) by A2The continuous child window for being divided into the pixel of 40 pixels × 40 without overlapping;
(4) A is calculated2In the 1st child window pixel point value average value, S1(1)=0.097;
(5) A is calculated2In the every row pixel point value of the 1st child window standard deviation, then calculate the standard of often row standard deviation
Difference, S2(1)=0.010;
(6) A is calculated2In the 1st child window each column pixel point value standard deviation, then calculate the standard of each column standard deviation
Difference, S3(1)=0.012;
(7) A is calculated2In all child window S1(j), S2(j) and S3(j) average value, obtains S1=0.114, S2=0.012
And S3=0.014
Step 2
(1) Fig. 4 application Gauss filters are handled, is denoted as B1, filtering size used is 5 × 5, standard deviation 3;
(2) apply 4 filter M1~M4 to B1It is handled, and is added after handling result is taken absolute value, be denoted as B2;
(3) by B2The continuous child window for being divided into the pixel of 40 pixels × 40 without overlapping;
(4) B is calculated2In the 1st child window pixel point value average value, T1(1)=0.097;
(5) B is calculated2In the every row pixel point value of the 1st child window standard deviation, then calculate the standard of often row standard deviation
Difference, T2(1)=0.011;
(6) B is calculated2In the 1st child window each column pixel point value standard deviation, then calculate the standard of each column standard deviation
Difference, T3(1)=0.012;
(7) δ=0.04 is taken, is calculated | T1(1)-S1|+|T2(1)-S2|+|T3(1)-S3|=0.021 < 0.04, therefore current son
Window does not have fault.
Embodiment 2
Step 1
(1) acquisition carries out Gauss filter to image A, is denoted as A corresponding to Fig. 3 without fault fabric image A1, Gauss used
Filter size is 5 × 5, standard deviation 3;
(2) apply 4 filter M1~M4 to A1It is handled, and is added after handling result is taken absolute value, be denoted as A2;
(3) by A2The continuous child window for being divided into the pixel of 32 pixels × 32 without overlapping;
(4) A is calculated2In the 10th child window pixel point value average value, S1(10)=0.148;
(5) A is calculated2In the every row pixel point value of the 10th child window standard deviation, then calculate the standard of often row standard deviation
Difference, S2(10)=0.018;
(6) A is calculated2In the 10th child window each column pixel point value standard deviation, then calculate the standard of each column standard deviation
Difference, S3(10)=0.026;
(7) A is calculated2In all child window S1(j), S2(j) and S3(j) average value, obtains S1=0.119, S2=0.014
And S3=0.016
Step 2
(1) Fig. 4 application Gauss filters are handled, is denoted as B1, filtering size used is 5 × 5, standard deviation 3;
(2) apply 4 filter M1~M4 to B1It is handled, and is added after handling result is taken absolute value, be denoted as B2;
(3) by B2The continuous child window for being divided into the pixel of 32 pixels × 32 without overlapping;
(4) B is calculated2In the 80th child window pixel point value average value, T1(80)=0.196;
(5) B is calculated2In the every row pixel point value of the 80th child window standard deviation, then calculate the standard of often row standard deviation
Difference, T2(80)=0.055;
(6) B is calculated2In the 80th child window each column pixel point value standard deviation, then calculate the standard of each column standard deviation
Difference, T3(80)=0.065;
(7) δ=0.06 is taken, is calculated | T1(80)-S1|+|T2(80)-S2|+|T3(80)-S3|=0.167>0.06, thus it is current
Sub- window has fault.
The detection result to Fig. 2 and Fig. 3 is set forth in Fig. 4 and Fig. 5.
Claims (5)
1. a kind of automatic cloth inspection method using contact-type image sensor, fabric image is adopted using contact-type image sensor
Collection, then transmits data to industrial control computer by CameraLink interfaces, it is characterized in that:The industrial control computer is to fabric
Image is detected, by the average value of calculating fabric image and the pixel value of the child window of standard picture, per row pixel value
The sum of the absolute value of difference of standard deviation of standard deviation and each column pixel value, there are whether fault for judgement fabric.
2. a kind of automatic cloth inspection method using contact-type image sensor according to claim 1, which is characterized in that institute
State calculate with the judgement the specific steps are:
Step 1) standard picture is that the parameter of the fabric image A without fault calculates:
1.1) gaussian filtering is carried out to the fabric image A of no fault, is denoted as A1;
1.2) apply 4 filter M1~M4 successively to A1It is filtered, and is added after filter result is taken absolute value, be denoted as A2;
1.3) by A2The continuous child window for being divided into p pixels × p pixels without overlapping, value 16≤p≤64;
1.4) A is calculated2In j-th of child window pixel value average value, be denoted as S1(j);
1.5) A is calculated2In the standard deviation of j-th child window per row pixel value, then calculate the standard deviation of often row standard deviation, be denoted as
S2(j);
1.6) A is calculated2In j-th of child window each column pixel value standard deviation, then calculate the standard deviation of each column standard deviation, be denoted as
S3(j);
1.7) A is calculated2In all child window S1(j)、S2(j) and S3(j) average value, is denoted as S1、S2And S3;
The parameter of step 2) fabric image calculates:
2.1) gaussian filtering is carried out to fabric image B, is denoted as B1;
2.2) apply 4 filter M1~M4 to B1It is handled, and is added after handling result is taken absolute value, be denoted as B2;
2.3) by B2The continuous child window for being divided into p pixels × p pixels without overlapping, value 16≤p≤64;
2.4) B is calculated2In j-th of child window pixel value average value, be denoted as T1(j);
2.5) B is calculated2In the standard deviation of j-th child window per row pixel value, then calculate the standard deviation of often row standard deviation, be denoted as
T2(j);
2.6) B is calculated2In j-th of child window each column pixel value standard deviation, then calculate the standard deviation of each column standard deviation, be denoted as
T3(j);
Step 3) mathematic interpolation and judgement:
If current j-th of child window meets | T1(j)-S1|+|T2(j)-S2|+|T3(j)-S3| > δ then sentence current sub-window
Surely there is fault;δ values 0.05≤δ≤0.2.
3. a kind of automatic cloth inspection method using contact-type image sensor according to claim 2, which is characterized in that institute
The 4 filter M1~M4 stated are:
4. a kind of automatic cloth inspection method using contact-type image sensor according to claim 1, which is characterized in that hair
Existing fault then preserves fault information, while marking fault in fabric by labelling apparatus.
5. a kind of automatic cloth inspection method using contact-type image sensor according to claim 1, which is characterized in that connect
Touch imaging sensor length is 2 meters, i.e., maximum detection fabric breadth is 2 meters, 150~300 pixel/inch of optical resolution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610112469.3A CN105784712B (en) | 2016-02-29 | 2016-02-29 | A kind of automatic cloth inspection method using contact-type image sensor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610112469.3A CN105784712B (en) | 2016-02-29 | 2016-02-29 | A kind of automatic cloth inspection method using contact-type image sensor |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105784712A CN105784712A (en) | 2016-07-20 |
CN105784712B true CN105784712B (en) | 2018-07-17 |
Family
ID=56386516
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610112469.3A Active CN105784712B (en) | 2016-02-29 | 2016-02-29 | A kind of automatic cloth inspection method using contact-type image sensor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105784712B (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103954632A (en) * | 2014-03-14 | 2014-07-30 | 东华大学 | Contact image sensor based gray fabric defect detection system |
CN104048966A (en) * | 2014-03-14 | 2014-09-17 | 东华大学 | Big-law-based cloth cover defect detection and classification method |
CN104181170A (en) * | 2014-09-05 | 2014-12-03 | 熊菊莲 | Fruit appearance detection method based on spectrum image analysis |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS63163150A (en) * | 1986-12-25 | 1988-07-06 | Matsushita Electric Ind Co Ltd | Printed circuit board inspecting device |
JPH0333262A (en) * | 1989-06-28 | 1991-02-13 | Toyota Autom Loom Works Ltd | Fabric inspection device |
-
2016
- 2016-02-29 CN CN201610112469.3A patent/CN105784712B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103954632A (en) * | 2014-03-14 | 2014-07-30 | 东华大学 | Contact image sensor based gray fabric defect detection system |
CN104048966A (en) * | 2014-03-14 | 2014-09-17 | 东华大学 | Big-law-based cloth cover defect detection and classification method |
CN104181170A (en) * | 2014-09-05 | 2014-12-03 | 熊菊莲 | Fruit appearance detection method based on spectrum image analysis |
Non-Patent Citations (4)
Title |
---|
A real-time computer vision-based platform for fabric inspection part 1:algorithm;Jian Zhou and Jun Wang;《The Journal of The Textile Institute》;20151231;第106卷(第12期);1282-1292 * |
Real-Time Vision-Based System for Textile Fabric Inspection;Radovan Stojanovic et al.;《Real-Time Imaging》;20011231;第7卷;507-518 * |
基于机器视觉的坯布疵点实时自动检测平台;李冠志等;《东华大学学报(自然科学版)》;20140228;第40卷(第1期);11-16 * |
基于机器视觉的坯布自动检测技术;李勇等;《纺织学报》;20070831;第28卷(第8期);124-128 * |
Also Published As
Publication number | Publication date |
---|---|
CN105784712A (en) | 2016-07-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107203990B (en) | Label breakage detection method based on template matching and image quality evaluation | |
CN109454006B (en) | Detection and classification method based on device for online detection and classification of chemical fiber spindle tripping defects | |
CN107064160B (en) | Textile surface flaw detection method and system based on significance detection | |
CN106952257B (en) | A kind of curved surface label open defect detection method based on template matching and similarity calculation | |
CN107941808B (en) | 3D printing forming quality detection system and method based on machine vision | |
CN104361314B (en) | Based on infrared and transformer localization method and device of visual image fusion | |
CN109507192A (en) | A kind of magnetic core detection method of surface flaw based on machine vision | |
WO2014139273A1 (en) | Weld seam defect detection method | |
CN108267455B (en) | Device and method for detecting defects of printed characters of plastic film | |
CN110991360B (en) | Robot inspection point position intelligent configuration method based on visual algorithm | |
CN103051872A (en) | Method for detecting conveyor belt deviation based on image edge extraction | |
CN109409289A (en) | A kind of electric operating safety supervision robot security job identifying method and system | |
CN112858321A (en) | Steel plate surface defect detection system and method based on linear array CCD | |
CN105699386B (en) | A kind of automatic cloth inspection labeling method using contact-type image sensor | |
CN110880184A (en) | Method and device for carrying out automatic camera inspection based on optical flow field | |
CN206223683U (en) | A kind of tabular workpiece with hole surface defect detection apparatus | |
CN110503623A (en) | A method of Bird's Nest defect on the identification transmission line of electricity based on convolutional neural networks | |
CN105738376B (en) | A kind of automatic cloth inspecting machine using contact-type image sensor | |
CN116797977A (en) | Method and device for identifying dynamic target of inspection robot and measuring temperature and storage medium | |
CN105784712B (en) | A kind of automatic cloth inspection method using contact-type image sensor | |
CN211292638U (en) | Quick automatic checkout device of pen tube printing | |
CN107121063A (en) | The method for detecting workpiece | |
CN109406539B (en) | Transparent medicine bottle bottom accumulated material defect detection system and method | |
CN104112271A (en) | Detection method and system for housing side defect of mobile terminal | |
CN113077414B (en) | Steel plate surface defect detection method and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |