CN115294136A - Artificial intelligence-based construction and detection method for textile fabric flaw detection model - Google Patents
Artificial intelligence-based construction and detection method for textile fabric flaw detection model Download PDFInfo
- Publication number
- CN115294136A CN115294136A CN202211228035.1A CN202211228035A CN115294136A CN 115294136 A CN115294136 A CN 115294136A CN 202211228035 A CN202211228035 A CN 202211228035A CN 115294136 A CN115294136 A CN 115294136A
- Authority
- CN
- China
- Prior art keywords
- fabric
- defect
- database
- detection model
- images
- 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
- 239000004744 fabric Substances 0.000 title claims abstract description 97
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 239000004753 textile Substances 0.000 title claims abstract description 27
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 17
- 238000010276 construction Methods 0.000 title claims abstract description 11
- 230000007547 defect Effects 0.000 claims abstract description 99
- 238000012360 testing method Methods 0.000 claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000009941 weaving Methods 0.000 claims abstract description 17
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000004519 manufacturing process Methods 0.000 claims description 10
- 230000002950 deficient Effects 0.000 claims description 9
- 238000013135 deep learning Methods 0.000 claims description 6
- 235000009161 Espostoa lanata Nutrition 0.000 claims description 4
- 240000001624 Espostoa lanata Species 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000003321 amplification Effects 0.000 claims description 3
- 238000005094 computer simulation Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 description 4
- 238000007639 printing Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000002203 pretreatment Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000282326 Felis catus Species 0.000 description 1
- 229920000297 Rayon Polymers 0.000 description 1
- 208000003464 asthenopia Diseases 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000005906 menstruation Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012372 quality testing Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration by the use of histogram techniques
-
- G06T5/90—
-
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses a construction and detection method of a textile fabric defect detection model based on artificial intelligence, which comprises the following steps of construction of a textile defect training set, pretreatment of textile defect images, construction of a textile defect database, construction of a test set, construction of a model through the database, training and use of the detection model, wherein the system method can reduce 50% of related textile defect detection personnel at least, improve labor productivity, reduce cost and reduce labor intensity of workers; meanwhile, weaving defects are reduced under the guidance of detection data, and the woven cloth is qualified.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a construction and detection method of a textile fabric flaw detection model based on artificial intelligence.
Background
At present, the main cloth inspecting mode in the textile industry is semi-automatic manual cloth inspecting, and a qualified inspector can meet the requirements of inspecting, correctly identifying and evaluating the defects of all textiles, but human visual detection cannot have deviation. In addition, since an inspector concentrates his or her efforts in an cloth inspection work environment (mainly asthenopia caused by strong light) for a maximum time of 20 to 30 minutes, fatigue occurs over the time period, and even if a person is fully attentive, only 200 events can be observed per hour, but if a special event is not observed within 20 seconds, the inspector's attention is reduced and only a response to a clear event is made. In fact, about 30% of defects are not detected, even though the reproducibility of defect scores from person to person in an examination environment rarely exceeds 50%, and as tests continue, a test deviation may or may not be deemed defect, even though weaving defects are generally rated defects, in individual cases are considered "harmless". But other deviations can only be rated as "harmless" or harmful defects with some degree of uncertainty. Thus, it is not possible for a "test" inspector to have an objective scale, often with many test mistakes.
(1) Differences in the textile must be sought and the entire textile must be covered. If the textile is moved at a speed of 0.08-0.33M/S, even in some cases up to 1M/S. In the distance range of lM, the human eye can only provide clear images of a circle with a diameter of 0.018M, and must therefore be moved continuously.
(2) The influence of the features of the defects, such as size and shape, contrast, weaving uniformity, etc., on the subsequent process steps must be evaluated in each case continuously: defects are considered to be classified into different categories, effects and possible remedies.
(3) The determination of whether a defect is being repaired or ignored must be made continuously throughout the move. The above factors illustrate the insurmountable deficiencies of manual textile testing, making such testing certainly unable to meet the requirements of testing with regard to high reliability and reproducibility, and the objectivity, justice and authority of quality testing is therefore greatly reduced.
Disclosure of Invention
In order to solve the main technical problem, the invention provides a construction detection method of a textile fabric flaw detection model based on artificial intelligence. The complete system can capture images of any object, and according to different parameters of quality and safety, once the defects appear, the defects are respectively classified, recorded and marked, and scored and assessed; if the alarm occurs continuously or repeatedly, the treatment is stopped in time.
The technical scheme of the invention is as follows:
the construction and detection method of the textile fabric flaw detection model based on artificial intelligence comprises the following steps,
step 6, adopting a pretr artificial intelligence ned + fine-tuning strategy, putting the constructed detection model network model on an ImageNet database for pre-training, and adjusting and optimizing a fabric database according to a training result to obtain a final defect detection model of the textile fabric;
step 7, storing the model obtained in the step 6 in a deep learning server, and installing the linear array camera on a cloth feeding system production line;
step 8, shooting the cloth on the production line by the linear array camera, transmitting the shot cloth into a deep learning server for identifying a detection model, preprocessing the picture containing the defects, and then transmitting the preprocessed picture to an online computer through a concentrator for displaying and storing;
step 9, once the defects appear, respectively classifying and marking out records, and grading and checking; if the treatment occurs continuously or repeatedly, an alarm must be given to stop the treatment in time.
Preferably, the fabric defect image database in the step 1 comprises normal cloth samples, 8 types of defect samples of broken warp, broken weft, broken silk, crosspieces, oil stains, broken holes, cotton balls and creases, which are common in the weaving process, and 256-level gray level images with 256 pixels by 256 pixels in size;
the database comprises 40000 defect sample pictures and 10000 normal cloth sample pictures;
the normal cloth images in the database are named as C0, and the 8 types of defect sample images are named as C1-C8 in sequence; the normal cloth sample image is the background image when no C1-C8 defects appear.
Preferably, the preprocessing method in step 2 is a histogram equalization method, an adaptive median filtering method and a gray level co-occurrence matrix.
Preferably, the system further comprises a length measuring instrument, the length measuring instrument is used for synchronously detecting the speed and the direction of the system and the cloth and is connected to the line camera through an I/O box, so that the line camera can locate the position of the defect.
Compared with the prior art, the invention adopts a precise meter length changer, a positioning method that the defects can not be manually penetrated by paper is detected, new methods such as viscose label printing, visible and invisible ink label printing, ultraviolet visible ink spot printing and the like are commonly used for selvedge printing, and computer yarn defect data can be stored for a long time, is subjected to drawing statistical analysis and is networked with a computer control center;
the system and the method can reduce 50% of related textile defect detection personnel at least, improve labor productivity, reduce cost and reduce labor intensity of workers; meanwhile, weaving defects are reduced under the guidance of detection data, and the woven cloth is qualified.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some examples of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive labor.
Figure 1 is a sample image of defects in accordance with the present invention.
Fig. 2 is a graph of the optimized depth residual error network CrossNet for fabric detection according to the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
Example 1
As shown in fig. 1 and fig. 2, the method for constructing and detecting the textile fabric flaw detection model based on artificial intelligence comprises the following steps,
The fabric defect image database comprises normal cloth samples, 8 types of defect samples of broken warp, broken weft, broken silk, crosspiece, oil stain, broken hole, cotton ball and crease which are common in the weaving process, and 256-level gray level images with the size of 256 × 256 pixels; the database comprises 40000 defect sample pictures and 10000 normal cloth sample pictures; the normal cloth images in the database are named as C0, and the 8 types of defect sample images are named as C1-C8 in sequence; and the normal cloth sample image is the background image when no C1-C8 defects appear. Fabric defect database sample distribution statistics are as follows:
| serial number | Sample name | Number of samples | Serial number | Sample name | Number of samples |
| C0 | Is normal | 1000 | C5 | Oil stain | 500 |
| C1 | Lack of menstruation | 500 | C6 | Hole breaking | 500 |
| C2 | Weft break | 500 | C7 | Cotton ball | 500 |
| C3 | Hook wire | 500 | C8 | Fold line | 500 |
| C4 | Weft stop | 500 |
When the same pretreatment method is selected for all weaving defect images, the quality of the weaving defect images cannot be enhanced; when the same weaving defect image is preprocessed by a plurality of methods, a plurality of important defect detail information is lost and the calculation amount is increased. Therefore, it is necessary to roughly determine the weaving defect image and then select an appropriate preprocessing method to achieve a good effect.
The images in the established weaving defect database are not all from the same cloth, so the quality of the images is different. Part of weaving defect images contain more noise points, and are caused by the thickness, irregular distribution and different textile production processes of textile raw materials; the contrast between the partially woven defect image object and the fabric texture is not high. Pretreatment is particularly required when the contrast between the defect target and the cloth texture background is low to highlight the detailed parts of the fabric defect. The contrast value of the weaving defect image can be scaled by the contrast value of the gray level co-occurrence matrix, and the defect details of the fabric defects with higher contrast are easier to clearly observe. Therefore, the contrast value in the gray level co-occurrence matrix can be adopted to identify the defect image, so as to achieve the purpose of adaptively selecting the corresponding preprocessing method.
The depth residual error network is mainly characterized in that a depth model with any number of layers can be constructed by using a residual error unit module. Generally, the more the network layer number is, the stronger the abstract feature extraction capability is, and the stronger the similarity feature classification capability is. In order to fully utilize the residual error units on the shallow layer of the front section part of the model to mainly learn the structural semantic features of the fabric texture image, such as color, texture structure, brightness and the like, the three residual error unit modules at the back can learn the fabric low-layer features obtained by the two residual error modules at the front, and the residual error network structure across layers is designed based on the idea of residual errors. Different from the traditional residual error network, the model enables a deep network to fully fuse the low-level characteristics of a front-end network, so that higher-level image semantic characteristics can be obtained through learning. For the condition that the scales of the feature maps output by the first two residual error units are relatively large, a global average pooling layer is adopted to reduce the dimensionality of the feature maps, and the problem of inconsistent input dimensionality in cross-layer transmission is solved.
step 6, adopting a pretr artificial intelligence ned + fine-tuning strategy, putting the constructed detection model network model on an ImageNet database for pre-training, and adjusting and optimizing a fabric database according to a training result to obtain a final defect detection model of the textile fabric;
the deep neural network needs a large amount of manual marking data for training due to the large number of parameters, however, the fabric database disclosed at present is rare, and the sample size of the database stored only is difficult to meet the training requirement of a large network. Therefore, the hyper-parameter training of the model adopts a pretr artificial intelligence ned + fine-tuning strategy, the built network model is put on an ImageNet database for pre-training, and then the pretr artificial intelligence ned model is finely tuned by using a fabric database built by the user. The strategy not only solves the problem that the fabric database small sample is easy to over-fit to the complex network, but also saves training time. The ImageNet image data set is the most applied image database with the strongest authority in the field of academic images at present, and comprises 1200 thousands of images covering 1000 different categories. The feature extraction capability obtained by training and learning in the ImageNet database can be transferred to fabric defect target detection, and the fabric defect detection performance of ResNet is improved.
The method comprises the following steps of randomly extracting samples from a fabric training set to carry out multiple comparison tests on the hyper-parameters of the migration network, and finally setting the hyper-parameters of the network according to the experimental effect as follows: the Learning Rate (Base Learning Rate) is set to 0.01, the Weight Decay factor (Weight Decay) is 0.0001, the Momentum factor (Momentum) is set to 0.9 and the Maximum number of Iterations (Maximum Iterations) is 60000 every 50 epochs. In the optimization of the objective function, a Stochastic Gradient Descent (SGD) algorithm is used. Random gradient descent is an unconstrained optimization problem, and solves two problems of gradient descent: slow convergence and falling into local optima. The method has the characteristics of high optimization speed and easiness in implementation, and is widely applied to training of a plurality of deep learning models. Compared to Batch Gradient Descent (BGD), SGD estimates the actual gradient by considering only a single training sample at a time. The idea of stochastic gradient descent is to minimize the loss function per sample, and the SGD processes only one training sample per iteration, each training sample consisting of uniformly sampling 50 positive samples and 80 negative samples. The initial values of the parameters of each layer of the network use Gaussian distribution with the standard deviation of 0.01.
Step 7, storing the model obtained in the step 6 in a deep learning server, and installing the linear array camera on a production line of a cloth feeding system;
step 8, shooting the cloth on the production line by the linear array camera, transmitting the shot cloth into a deep learning server for identifying a detection model, preprocessing the picture containing the defects, and then transmitting the preprocessed picture to an online computer through a concentrator for displaying and storing;
step 9, once the defects appear, respectively classifying and marking out records, and grading and checking; if the alarm occurs continuously or repeatedly, the treatment is stopped in time.
In addition, the invention also comprises a length measuring instrument which is used for synchronously detecting the speed and the direction of the system and the cloth and is connected to the linear array camera through an I/O box, so that the linear array camera can position the position of the defect.
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments or portions thereof without departing from the spirit and scope of the invention.
Claims (4)
1. The construction and detection method of the textile fabric flaw detection model based on artificial intelligence is characterized by comprising the following steps of,
step 1, constructing a fabric defect training set, collecting a sufficient number of images on an actual production line by adopting a linear array camera, segmenting the areas where the defects are located from the images, and then establishing a common type fabric defect point database;
step 2, preprocessing a fabric defect image, highlighting fabric texture image information and defect information, extracting defect characteristics, and establishing a fabric defect database;
step 3, establishing a test set, testing a detection model by adopting 150 defective fabric images in a TILDA fabric database, obtaining more test database samples by a data amplification method such as angle rotation and scaling, wherein icon data represent the number of corresponding defective type images and comprise defective types such as holes, spots, warp hanging, weft missing and warp missing, and the size of the images is 256 multiplied by 256;
step 4, building a detection model of the optimized depth residual error network CrossNet of the fabric detection by using the depth convolution neural network, resNet and the fabric defect database obtained in the step 2, and testing the detection model through the test set in the step 3;
step 5, marking data after the test is finished, marking a target by adopting a machine vision technology, and storing a marking result as data required by the training of the computer model;
step 6, adopting a pretr artificial intelligence ned + fine-tuning strategy, putting the constructed detection model network model on an ImageNet database for pre-training, and adjusting and optimizing a fabric database according to a training result to obtain a final defect detection model of the textile fabric;
step 7, storing the model obtained in the step 6 in a deep learning server, and installing the linear array camera on a production line of a cloth feeding system;
step 8, shooting the cloth on the production line by the linear array camera, transmitting the shot cloth into a deep learning server for identifying a detection model, preprocessing the picture containing the defects, and then transmitting the preprocessed picture to an online computer through a concentrator for displaying and storing;
step 9, once the defects appear, respectively classifying and marking out records, and grading and checking; if the treatment occurs continuously or repeatedly, an alarm must be given to stop the treatment in time.
2. The artificial intelligence based textile fabric defect detection model building and detecting method according to claim 1, wherein said step 1 fabric defect image database comprises normal cloth samples and 8 types of defect samples of warp broken, weft broken, silk hooked, crosspiece, oil stain, broken hole, cotton ball and crease during weaving, 256-level gray scale images with 256 pixels each;
the normal cloth images in the database are named as C0, and the 8 types of defect sample images are named as C1-C8 in sequence; the normal cloth sample image is the background image when no C1-C8 defects appear.
3. The artificial intelligence based textile fabric flaw detection model building and detecting method as claimed in claim 1, characterized in that the preprocessing method in step 2 is histogram equalization, adaptive median filtering and gray level co-occurrence matrix.
4. A method for building and detecting a flaw detection model of textile fabric based on artificial intelligence according to any one of claims 1-3, further comprising a length measuring instrument, wherein the length measuring instrument is used for synchronizing the speed and direction of the detection system and the fabric, and is connected to the line camera through an I/O box, so that the line camera can locate the position of the flaw.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211228035.1A CN115294136A (en) | 2022-10-09 | 2022-10-09 | Artificial intelligence-based construction and detection method for textile fabric flaw detection model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211228035.1A CN115294136A (en) | 2022-10-09 | 2022-10-09 | Artificial intelligence-based construction and detection method for textile fabric flaw detection model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115294136A true CN115294136A (en) | 2022-11-04 |
Family
ID=83833679
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211228035.1A Pending CN115294136A (en) | 2022-10-09 | 2022-10-09 | Artificial intelligence-based construction and detection method for textile fabric flaw detection model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115294136A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116309493A (en) * | 2023-03-24 | 2023-06-23 | 南通飞隼信息科技有限公司 | Method and system for detecting defects of textile products |
| CN116736783A (en) * | 2023-08-16 | 2023-09-12 | 江苏德顺纺织有限公司 | Intelligent remote control system and method for textile electrical equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20090094955A (en) * | 2008-03-04 | 2009-09-09 | 주찬 | A triangle teabag manufacture device |
| CN110490874A (en) * | 2019-09-04 | 2019-11-22 | 河海大学常州校区 | Weaving cloth surface flaw detecting method based on YOLO neural network |
-
2022
- 2022-10-09 CN CN202211228035.1A patent/CN115294136A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20090094955A (en) * | 2008-03-04 | 2009-09-09 | 주찬 | A triangle teabag manufacture device |
| CN110490874A (en) * | 2019-09-04 | 2019-11-22 | 河海大学常州校区 | Weaving cloth surface flaw detecting method based on YOLO neural network |
Non-Patent Citations (1)
| Title |
|---|
| 陈宇俊: "基于深度卷积神经网络的织物瑕疵检测方法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅰ辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116309493A (en) * | 2023-03-24 | 2023-06-23 | 南通飞隼信息科技有限公司 | Method and system for detecting defects of textile products |
| CN116736783A (en) * | 2023-08-16 | 2023-09-12 | 江苏德顺纺织有限公司 | Intelligent remote control system and method for textile electrical equipment |
| CN116736783B (en) * | 2023-08-16 | 2023-12-05 | 江苏德顺纺织有限公司 | Intelligent remote control system and method for textile electrical equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rasheed et al. | Fabric defect detection using computer vision techniques: a comprehensive review | |
| CN115294136A (en) | Artificial intelligence-based construction and detection method for textile fabric flaw detection model | |
| CN111325713B (en) | Neural network-based wood defect detection method, system and storage medium | |
| CN101739570B (en) | Cotton foreign fiber online classifying method and system | |
| CN100401043C (en) | Image inspecting method and device for detecting faults | |
| CN109724984B (en) | Defect detection and identification device and method based on deep learning algorithm | |
| CN110136101B (en) | Tire X-ray defect detection method based on twinning distance comparison | |
| CN108074231A (en) | A kind of magnetic sheet detection method of surface flaw based on convolutional neural networks | |
| CN106845556A (en) | A kind of fabric defect detection method based on convolutional neural networks | |
| CN108416774A (en) | A kind of fabric types recognition methods based on fine granularity neural network | |
| CN112102229A (en) | Intelligent industrial CT detection defect identification method based on deep learning | |
| CN108932712A (en) | A kind of rotor windings quality detecting system and method | |
| CN110610475A (en) | Visual defect detection method of deep convolutional neural network | |
| CN109509171A (en) | A kind of Fabric Defects Inspection detection method based on GMM and image pyramid | |
| CN108711148A (en) | A kind of wheel tyre defect intelligent detecting method based on deep learning | |
| CN113554080A (en) | Non-woven fabric defect detection and classification method and system based on machine vision | |
| CN109946304A (en) | Surface defects of parts on-line detecting system and detection method based on characteristic matching | |
| CN114972342B (en) | Method for detecting surface defects of gearbox gear | |
| CN105184777A (en) | Painted design fabric defect detection method based on image decomposition | |
| CN114841957A (en) | Steel plate surface defect detection method based on deep learning | |
| CN117218114B (en) | Mesh fabric defect rapid detection method based on image data processing | |
| Zhang et al. | Fabric defect detection based on visual saliency map and SVM | |
| CN206897873U (en) | A kind of image procossing and detecting system based on detection product performance | |
| CN117152119A (en) | Profile flaw visual detection method based on image processing | |
| CN116579984A (en) | Intelligent cloth inspection method and system for autonomous fault learning |
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: 20221104 |