CN109825944A - A kind of online defect detection method of fabric knitting based on line laser - Google Patents
A kind of online defect detection method of fabric knitting based on line laser Download PDFInfo
- Publication number
- CN109825944A CN109825944A CN201910034418.7A CN201910034418A CN109825944A CN 109825944 A CN109825944 A CN 109825944A CN 201910034418 A CN201910034418 A CN 201910034418A CN 109825944 A CN109825944 A CN 109825944A
- Authority
- CN
- China
- Prior art keywords
- speck
- yarn
- laser
- motion
- knitting needle
- 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
Abstract
The present invention relates to a kind of online defect detection methods of fabric knitting based on line laser, including the following steps: laser rays or face is irradiated on yarn and knitting needle, yarn and knitting needle and laser are formed by light curtain and interweave with forming speck;Video camera acquisition is formed by speck;To each frame image in the regulation sampling time, yarn and knitting needle are obtained using Threshold segmentation and laser plane interweaves the speck position to be formed, extracts the centre coordinate of speck;Same speck position in different images is arranged in order, sequence is constituted to obtain motion profile of the single speck within the sampling time and obtains speck motion in one dimension signal graph;By Wavelet transformation, speck motion in one dimension signal is converted into two-dimentional spectrogram;Deep learning is carried out by CNN neural network, two-dimentional spectrogram corresponding to the corresponding two-dimentional spectrogram of the normal yarn characteristics of motion and abnormal conditions is marked respectively and is trained study.
Description
Technical field
The invention belongs to quality of textile products detection fields, are related to a kind of online defect detection of the fabric knitting based on line laser
Method.
Background technique
China is maximum fabric production in the world and goes abroad that in actual production, defect detection is mostly important.Currently,
The control of fabric defects fabric quality and detection are critically important, and fabric is usually to be completed by worker's range estimation, have detection speed low, work
The disadvantages of subjective factor of people influences greatly testing result, and there are also false detection rate and high omission factors.It is with export processing
Main, the lower high level production equipment for lacking autonomous production of textile added value is with the states such as import Germany and Japan machine
It is main.
Interconnection plane and automation control transformation are carried out to warp knit equipment, intelligence manufacture demonstration is realized to warp knit enterprise, be through
Compile important development direction of the industry under intelligence manufacture.In weaving process, due to mechanical breakdown or operating mistake etc.,
Cloth can have some flaws.Textile enterprise traditionally uses manual type to carry out fabric defects detection, in detection accuracy, speed
It is all undesirable in terms of degree and recall rate.Therefore, it is necessary to carry out the online vision detection system research of tricot machine Fabric Defect, base
The real-time detection of Fabric Defect during realizing of Robot Vision warp knit woven fabric.Currently, warp knit weave faults is quick, smart
The problem of quasi- identification, yet there are no relatively successfully general warp knit weaving vision detection system also in the exploratory stage, from it is practical
There are also comparable distances for line measurement.
Summary of the invention
The purpose of the present invention is to propose to a kind of online defect detection methods of fabric knitting based on line laser, can be in weaving
Broken yarn and the typical problems such as needle motion rule is abnormal be measured in real time and shut down prompting, fundamentally fabric is avoided to knit
Make the generation of middle fault.The advantages that adaptable good, the fault judgement of the present invention is in time, system cost is low, reliable and stable is full
Sufficient fabric defects weaves detection demand online.Technical solution is as follows:
A kind of online defect detection method of fabric knitting based on line laser, including the following steps:
(1) laser rays or face are irradiated on yarn and knitting needle, yarn and knitting needle and laser are formed by light curtain intertexture shape
At speck, so that yarn and knitting needle are lighted by laser light source;
(2) video camera acquisition is formed by speck, acquires the instantaneous picture of yarn and knitting needle in weaving process;
(3) to each frame image in the regulation sampling time, yarn and knitting needle is obtained using Threshold segmentation and laser plane is handed over
The speck position to be formed is knitted, the centre coordinate of speck is extracted by the method for center of gravity calculation;
(4) the same speck position in different images is arranged in order, sequence is constituted, to obtain single speck at this
Motion profile in sampling time obtains speck motion in one dimension signal graph;By Wavelet transformation, by speck motion in one dimension signal
Be converted to two-dimentional spectrogram;
(5) deep learning is carried out by CNN neural network, by the corresponding two-dimentional spectrogram of the normal yarn characteristics of motion and exception
Two-dimentional spectrogram corresponding to situation marks respectively and is trained study, so that judging whether to occur includes yarn broken string, yarn
Abnormal conditions including loosening.
The invention proposes a kind of online defect detection method of fabric knitting, by means of line laser by yarn and knitting needle point
It is bright, it acquires yarn and knitting needle image in weaving process in real time using high-speed camera, was weaved by the study of deep learning algorithm
The characteristics of motion of yarn and knitting needle in journey judges that broken yarn and needle motion are advised in weaving by judging that the characteristics of motion is destroyed
The typical problems such as rule exception, are measured in real time and shut down prompting, fundamentally avoid the generation of fault in fabric knitting.This
The advantages that adaptable good, fault judgement is timely, system cost is low, reliable and stable is invented, meets fabric defects and weaves online
Detection demand.
Detailed description of the invention
The online defect detection system structure diagram of Fig. 1 fabric.1 video camera;2 laser line generators;3 knitting needles;4 hot spots;5 yarns
Line;6 knitting lathes;7 woven fabrics.
The online defect detection flow chart figure of Fig. 2 fabric of the present invention.
Fig. 3 detection algorithm flow chart.
Fig. 4 subarea processing schematic diagram.
Specific embodiment
Of the invention is described in detail with reference to the accompanying drawings and examples.
System structure of the invention is as shown in Fig. 2, be installed on knitting lathe front end for linear structural laser device, and laser rays is from knitting
Needle lathe front end oblique upper direct irradiation yarn, another laser transmission laser line irradiate knitting needle, and video camera is also mounted on
Lathe front end and laser deviate certain distance, synchronous acquisition yarn and knitting needle speck image.
Overhaul flow chart of the invention is as shown in Figure 1.6 on knitting lathe, laser rays or face are irradiated to yarn 5 and knitted
On needle 3, yarn and knitting needle and laser are formed by light curtain and interweave with forming speck 4, so that yarn and knitting needle are lighted by laser light source;
The acquisition of video camera 1 is formed by speck, can acquire the instantaneous picture of yarn and knitting needle in weaving process.High speed acquisition weaving
The realtime graphic of yarn and knitting needle in the process, the characteristics of motion that both can capture in field range, the characteristics of motion can be with
Learnt by deep learning algorithm.The characteristics of motion is acquired and judges in real time, when the characteristics of motion is destroyed, it can be determined that yarn
There is the problems such as broken yarn or abnormal needle motion in the corresponding position of line or knitting needle, also may determine that the tendency occurred for fault,
To which parking avoids the waste of weaving process in time.
Specific detection algorithm all uses Threshold segmentation as shown in figure 3, first to each frame picture in the regulation sampling time
It obtains yarn and knitting needle and laser plane interweaves the speck position to be formed, the method for passing through center of gravity calculation extracts the center of speck
Coordinate.Then, the same speck position in different pictures is arranged in order, sequence is constituted, to obtain single speck at this
Motion profile in sampling time.The buffeting of lathe causes yarn vibration direction relatively simple during yarn weaving, therefore bright
Spot is regarded as moving reciprocatingly in one-dimensional direction, can state speck motion conditions by one-dimensional signal, obtains speck fortune
Dynamic signal graph.Again by Wavelet transformation, speck motion in one dimension signal is converted into two-dimentional spectrogram.Finally, passing through CNN nerve net
Network carries out deep learning.By two-dimentional spectrogram point corresponding to the corresponding two-dimentional spectrogram of the normal yarn characteristics of motion and abnormal conditions
It does not mark and is trained study, to judge whether it is the abnormal conditions such as yarn broken string, yarn loosening.
To reduce system cost, need for knitting lathe using few as far as possible video camera and laser, but if one
Visual field captured by set measuring device is excessive, and the position judgement when there is broken yarn or needle motion exception there is difficulty, because
This, above judgment method can only judge to exist but specific location cannot be navigated to.For this purpose, being adopted in journey processed above
The mode for being learnt with subregion and being judged, that is to say, that it is some zonules, these zonules that video camera, which is acquired image segmentation,
Study and real-time judge are carried out respectively, can be which or which zonule occurs with accurate judgement once being abnormal,
Abnormal position is just navigated to quickly, it is specific as Fig. 3 illustrates.
The above laser can use random color visible light source or near-infrared light source, be using the benefit of near-infrared light source
Human eye is invisible, to avoid the visual field of interference execute-in-place worker, but what is matched therewith just needs using with good close
The video camera of infrared spectrum response characteristic.
The present invention is as follows for the specific implementation process of fabric defect detection method:
(1) build the online defect detection system of fabric, by linear structural laser device be installed on knitting lathe front end, laser rays from
Knitting needle lathe front end oblique upper direct irradiation yarn, another laser transmission laser line irradiate knitting needle, and video camera is also mounted
Deviate certain distance in lathe front end and laser;
(2) video camera acquisition is formed by speck, can acquire the instantaneous picture of yarn and knitting needle in weaving process;
(3) knitting lathe starts weaving process, and the reality acquired in a period of time yarn and knitting needle weaving process in real time adopts figure,
Region division appropriate is carried out to real figure of adopting, respectively and deep learning is carried out according to region, obtains the yarn and knitting needle in each region
The characteristics of motion;
(4) circulation acquires the moving image of yarn and knitting needle in practical weaving process in real time, utilizes deep learning result point
Analyse whether the characteristics of motion in each subregion belongs to normally;
(5) it finds that the characteristics of motion in any subregion is abnormal, determines that yarn or needle motion are extremely local, carry out shutdown behaviour
Make, carries out troubleshooting.
Claims (1)
1. a kind of online defect detection method of fabric knitting based on line laser, including the following steps:
(1) laser rays or face are irradiated on yarn and knitting needle, yarn and knitting needle and laser be formed by light curtain interweave to be formed it is bright
Spot, so that yarn and knitting needle are lighted by laser light source.
(2) video camera acquisition is formed by speck, acquires the instantaneous picture of yarn and knitting needle in weaving process;
(3) to each frame image in the regulation sampling time, yarn and knitting needle and laser plane intertexture shape are obtained using Threshold segmentation
At speck position, the centre coordinate of speck is extracted by the method for center of gravity calculation;
(4) the same speck position in different images is arranged in order, sequence is constituted, to obtain single speck in the sampling
Motion profile in time obtains speck motion in one dimension signal graph;By Wavelet transformation, speck motion in one dimension signal is converted to
Two-dimentional spectrogram;
(5) deep learning is carried out by CNN neural network, by the corresponding two-dimentional spectrogram of the normal yarn characteristics of motion and abnormal conditions
Corresponding two-dimentional spectrogram marks respectively and is trained study, so that judging whether to occur includes yarn broken string, yarn loosening
Abnormal conditions inside.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910034418.7A CN109825944B (en) | 2019-01-15 | 2019-01-15 | Online fabric weaving defect detection method based on line laser |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910034418.7A CN109825944B (en) | 2019-01-15 | 2019-01-15 | Online fabric weaving defect detection method based on line laser |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109825944A true CN109825944A (en) | 2019-05-31 |
CN109825944B CN109825944B (en) | 2020-04-28 |
Family
ID=66861098
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910034418.7A Active CN109825944B (en) | 2019-01-15 | 2019-01-15 | Online fabric weaving defect detection method based on line laser |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109825944B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111429422A (en) * | 2020-03-19 | 2020-07-17 | 中国工程物理研究院激光聚变研究中心 | Laser near-field state analysis method and device based on deep learning |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3760184A (en) * | 1972-03-10 | 1973-09-18 | Sick Erwin Fa | Photoelectric monitoring device for pluralities of threads |
US6522402B1 (en) * | 1998-04-30 | 2003-02-18 | California Institute Of Technology | Apparatus and method for analyzing microscopic samples based on optical parametric oscillation |
CN102661724A (en) * | 2012-04-10 | 2012-09-12 | 天津工业大学 | RGBPSP (red green blue phase shift profilometry) three-dimensional color reconstruction method applied to online detection for fabric defects |
CN103116351A (en) * | 2012-12-28 | 2013-05-22 | 福州科迪电子技术有限公司 | Spinning defective cloth detection camera and detection system thereof |
CN104849275A (en) * | 2015-04-22 | 2015-08-19 | 上海工程技术大学 | High speed digital imaging device and method for yarn appearance |
CN105155126A (en) * | 2015-08-25 | 2015-12-16 | 哈尔滨展达机器人自动化有限责任公司 | Broken line detection device based on laser vision for warp knitting |
-
2019
- 2019-01-15 CN CN201910034418.7A patent/CN109825944B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3760184A (en) * | 1972-03-10 | 1973-09-18 | Sick Erwin Fa | Photoelectric monitoring device for pluralities of threads |
US6522402B1 (en) * | 1998-04-30 | 2003-02-18 | California Institute Of Technology | Apparatus and method for analyzing microscopic samples based on optical parametric oscillation |
CN102661724A (en) * | 2012-04-10 | 2012-09-12 | 天津工业大学 | RGBPSP (red green blue phase shift profilometry) three-dimensional color reconstruction method applied to online detection for fabric defects |
CN103116351A (en) * | 2012-12-28 | 2013-05-22 | 福州科迪电子技术有限公司 | Spinning defective cloth detection camera and detection system thereof |
CN104849275A (en) * | 2015-04-22 | 2015-08-19 | 上海工程技术大学 | High speed digital imaging device and method for yarn appearance |
CN105155126A (en) * | 2015-08-25 | 2015-12-16 | 哈尔滨展达机器人自动化有限责任公司 | Broken line detection device based on laser vision for warp knitting |
Non-Patent Citations (2)
Title |
---|
刘小敏等: "基于图像处理的织物疵点检测算法综述", 《微处理机》 * |
朱昊: "图像工程和模式识别在针织中的应用", 《四川丝绸》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111429422A (en) * | 2020-03-19 | 2020-07-17 | 中国工程物理研究院激光聚变研究中心 | Laser near-field state analysis method and device based on deep learning |
Also Published As
Publication number | Publication date |
---|---|
CN109825944B (en) | 2020-04-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103529051B (en) | A kind of Woven textiles flaw automatic on-line detection method | |
CN103604809B (en) | A kind of online visible detection method of pattern cloth flaw | |
CN101799434B (en) | Printing image defect detection method | |
CN202330297U (en) | Device for automatically detecting texture defects | |
CN108931531A (en) | A kind of Automatic Detection of Fabric Defects method, system and computer readable storage medium | |
WO2022021774A1 (en) | Online detection method for circular weft knitting horizontal stripe defects based on grayscale gradient method | |
CN110097538A (en) | A kind of online cloth examination device of loom and defects identification method | |
CN202770781U (en) | Fabric defect online detection device based on machine vision | |
CN205317684U (en) | Flaw detecting system dyes cloth based on image | |
US20220170189A1 (en) | A device and a method for real-time identification of defects in fabrics, during weaving | |
CN109385876A (en) | A kind of the intelligent quality management system and its method of woven fabric | |
CN109632817A (en) | A kind of online defect detection method of fabric knitting based on collimated laser beam | |
CN110400306A (en) | Non-woven fabrics defect detection method based on morphologic filtering and convolutional neural networks | |
CN109825944A (en) | A kind of online defect detection method of fabric knitting based on line laser | |
CN107144570A (en) | A kind of tufting machine row's yarn error-detecting method based on machine vision | |
CN115508282B (en) | Online intelligent cloth inspection detection system | |
CN106757649B (en) | A kind of system and method that jacquard fabrics line version replicates automatically | |
CN106841235B (en) | Image extraction device and image imaging method based on machine vision fabric weaving flower-shaped pattern | |
CN108664588A (en) | Automatic method for online detection and control of cloth skewing | |
CN112391731B (en) | Online detection method for yarn breakage during weaving of warp knitting machine | |
CN202416024U (en) | Automatic fabric inspection machine using hybrid light source | |
CN108221339A (en) | A kind of perching machine detecting device of clothes processing | |
CN105717133B (en) | Automatic cloth inspecting machine based on linear interpolation method correcting image | |
KR20230137584A (en) | Vision inspection system for defect detection of fabric | |
Neumann et al. | In-process fault detection for textile fabric production: onloom imaging |
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 |