CN111652883A - Glass surface defect detection method based on deep learning - Google Patents
Glass surface defect detection method based on deep learning Download PDFInfo
- Publication number
- CN111652883A CN111652883A CN202010673882.3A CN202010673882A CN111652883A CN 111652883 A CN111652883 A CN 111652883A CN 202010673882 A CN202010673882 A CN 202010673882A CN 111652883 A CN111652883 A CN 111652883A
- Authority
- CN
- China
- Prior art keywords
- defect
- network
- training
- model
- deep learning
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 146
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 238000013135 deep learning Methods 0.000 title claims abstract description 22
- 239000011521 glass Substances 0.000 title claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 46
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000002372 labelling Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 9
- 238000003384 imaging method Methods 0.000 claims description 6
- 239000000428 dust Substances 0.000 abstract description 16
- 230000000694 effects Effects 0.000 abstract description 8
- 238000012805 post-processing Methods 0.000 abstract description 3
- 239000000047 product Substances 0.000 description 20
- 238000013145 classification model Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 240000004093 Mitragyna parvifolia Species 0.000 description 1
- 241000519995 Stachys sylvatica Species 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000004209 hair Anatomy 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 210000002268 wool Anatomy 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- 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
- G06T2207/30168—Image quality inspection
-
- 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 relates to a glass surface defect detection method based on deep learning, which comprises the steps of training a network model by utilizing a residual error network ResNet18 and a residual error network ResNet101, detecting and judging defects of a sample image through the trained network model after the training of the network model is finished, and managing and controlling the defect specification. Compared with the original method for classifying dirt, dust and defects by using the traditional algorithm, the method has the advantages that the classification accuracy of the model is improved to more than 90% from 70%; compared with the defect specification controlled by the score value of the deep learning classification result, the defect specification controlled by the method has better control effect by utilizing the extracted parameters such as length, width, length-width ratio, contrast ratio and the like to perform parameter control on the defect based on the post-processing of the deep learning detection result by the traditional image algorithm.
Description
Technical Field
The invention relates to the technical field of visual image processing, in particular to a glass surface defect detection method based on deep learning.
Background
The ResNet full name Residual Network. The best article for CVPR was obtained from Deep research for Image Recognition by Kaim He. The depth residual error network provided by the user reaches the depth of 152 layers on the premise of ensuring the network precision, and then is further added to the depth of 1000 layers, the characteristic grade is increased along with the deepening of the network, and the expression capacity of the network is greatly improved.
The existing case of applying deep learning to industrial defect detection is mainly divided into the following steps, wherein the first step is to collect a sample picture and label the defect, the second step is to set network weight parameters to train a model, and the third step is to use a test picture to detect OK and NG of the picture.
Therefore, the prior art has the following disadvantages:
1. most of the existing deep learning detection models are used for judging OK/NG of products, and the actual situation is that not only the OK/NG of the products needs to be judged, but also the NG types need to be classified, and defect specification management and control are carried out based on classification results. For glass products, dust and dirt belong to dirt defects and can be repaired by cleaning means, and scratches and bubbles belong to damaged defects and cannot be repaired, so that the two types of products need to be distinguished. Because the imaging characteristics of dust and white spots, linear dirt and scratches are similar, the existing application cases of the traditional algorithm and the deep learning have poor distinguishing effect on the two defects.
2. The existing deep learning detection model lacks of means of quantitative analysis, and the output detection result is not beneficial to parameter control of defect specifications. The existing application case can only output the heat map score of the NG product, the higher the score is, the higher the probability of the NG product is, but the heat map score does not have practical physical significance, and the management and control of defect specifications are not convenient for a manufacturer. The fabricator often wants to define and manage the defect specifications with intuitive parameters such as length, width, aspect ratio, contrast, number of defects in a certain area, etc.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the difficulty of detecting the defects on the surface of the glass is that dirt and dust are difficult to avoid and are easy to be confused with real defects; in addition, the control scale of the defect specification is difficult to control, the control is too tight, a large amount of over killing can be caused, the product detection yield is too low, if the parameter control is too loose, excessive missing detection can be caused, and the shipment quality is influenced; in order to solve the problems, a glass surface defect detection method based on deep learning is provided.
The technical scheme adopted by the invention for solving the technical problems is as follows: a glass surface defect detection method based on deep learning comprises the steps of training a network model by using a residual error network ResNet18 and a residual error network ResNet101, detecting and judging defects of a sample image through the trained network model after the training of the network model is finished, and managing and controlling the defect specification.
Further, the training process of the network model according to the present invention includes the following steps,
1) collecting various defect samples according to the defect detection requirement, and carrying out image acquisition on the samples by using an imaging system;
2) segmenting a region to be detected from the image, setting the gray value of a background region to be 0, and adding the gray value to a training sample image set for defect marking;
3) labeling the sample image;
4) model training, including training of a defect detection network and training of a defect discrimination network;
5) testing a sample picture by using a network model generated by training, and counting detected data and classified data of the model;
6) and collecting undetected pictures and pictures with wrong classification, adding the pictures into a training sample to mark defects, training the model again, testing the model, and iteratively optimizing the model once and again.
Still further, in the step 3), in the labeling process, a method of body projection integrated labeling is adopted, specifically, the body and the projection are connected and labeled into a whole by using a labeling brush.
Furthermore, in the step 4), firstly training the defect detection network, and then training the defect discrimination network; the defect detection network uses a residual error network ResNet18 to segment all defects from the image, and the defect discrimination network uses a residual error network ResNet101 to classify the defects.
Further, the invention utilizes the trained defect detection model to detect the defect of the image with the gray value of 0 in the background area; if the defects are detected, transmitting the images of the defect areas to a defect judgment network for next defect classification; if no suspicious defect is detected, the product is judged to be an OK product.
Further, the defect discrimination network of the present invention discriminates the defect type of the incoming suspicious defect region image, and each suspicious defect can obtain a classification result.
Furthermore, the defect specification management and control method of the invention is to calculate the mathematically definable defect parameters by using an image processing algorithm; if the defect parameters are within the control specification, determining as an OK product; and if the standard is out of the regulation specification, judging as an NG product.
The invention has the beneficial effects that:
1. compared with the original method for classifying dirt, dust and defects by using the traditional algorithm, the classification accuracy of the model is improved to more than 90% from 70%.
2. Compared with the defect specification controlled by the score value of the deep learning classification result, the defect specification controlled by the method has better control effect by utilizing the extracted parameters such as length, width, length-width ratio, contrast ratio and the like to perform parameter control on the defect based on the post-processing of the deep learning detection result by the traditional image algorithm. These parameters have practical physical meanings and are more in line with human perception of the defect specification.
Drawings
FIG. 1 is a flow chart of the network training of the present invention;
FIG. 2 is a flow chart of the present invention for detection using trained networks.
Detailed Description
The invention will now be described in further detail with reference to the drawings and preferred embodiments. These drawings are simplified schematic views illustrating only the basic structure of the present invention in a schematic manner, and thus show only the constitution related to the present invention.
As shown in fig. 1-2, a glass surface defect detection method based on deep learning includes the steps of firstly performing image segmentation by using a traditional image processing algorithm to extract a detection area, then performing defect detection on an image in the detection area by using a deep learning model based on ResNet18, classifying detected defects by using ResNet101, distinguishing dust, dirt and real defects, and finally performing defect specification control by combining with the traditional image processing algorithm to achieve the effects of defect classification detection and independent specification control of each type of defects.
The model comprises the following steps in the training process:
1. and collecting samples, namely collecting various defect samples according to the defect detection requirements, and acquiring images of the samples by using an imaging system. The defect samples in the glass detection comprise dust, dirt, bubbles, scratches and broken edges. In the stage of sample collection, a sample library needs to be established, balance of various samples in the training samples is kept, and overfitting caused by sample unbalance is prevented.
2. And extracting a detection area, namely segmenting an area to be detected from the image by using a method based on traditional image processing, setting the gray value of the background area to be 0 for eliminating the interference of the background area, and finally adding the image with the gray value of the background area set to be 0 to a training sample image set for defect marking.
3. And (4) defect marking, namely, creating 7 types of labels including dust, dirt, broken filaments, bubbles, scratches, broken edges and OK by using marking software, and marking the sample image. In the marking process, a method of body projection integral marking and imaging characteristics of the transparent glass material are adopted, so that when light irradiates the surface of the glass from a certain inclination angle, objects with protruding characteristics such as dust and broken filaments meet, a projection phenomenon exists in imaging, and other defects do not have the characteristics. Therefore, in the stage of sample labeling, the dust body, the projection, the hair body and the projection are connected and labeled into a whole by using the labeling brush in the embodiment. The practical test shows that the model trained by the body projection integrated labeling mode is higher in classification capability on dust, bubbles, broken filaments, scratches and dirt than the model trained by the body projection separation labeling mode, the classification accuracy is high by 15% when the initial training samples are fewer, and the classification accuracy required by customers can be achieved more quickly.
4. And model training, wherein the model training comprises a defect detection network and a defect judgment network, the defect detection network is trained firstly, and then the defect judgment network is trained. The defect detection network uses ResNet18 as a 2-class model, the model is responsible for segmenting all defects from images, and the improvement of the model effect mainly depends on the supplement of training samples, particularly the supplement training of undetected samples. The defect discrimination network is a multi-classification model, sufficient defect samples can not be collected frequently in a product proofing stage for model training, overfitting phenomena can easily occur on the basis of the model trained, the specific expression is that the trained classification model has good test effect on a training set, the test effect on the test set is poor, and the generalization capability of the model is weak. In order to solve the over-fitting problem, Early Stopping is performed in the training process of the classification model in the embodiment, at the end of each Epoch, the accurray of the validity data is calculated and recorded to the best validity, and when 10 consecutive epochs do not reach the best validity, iteration is stopped. The classification effect of the model trained in the way is good in the proofing stage and the small-batch trial production stage. After the product enters the stages of capacity climbing and formal mass production, along with the continuous supplementary training of the training samples, the means for preventing the overfitting phenomenon is changed into the management of the training sample library, and the quantity balance of the training samples with various defects is maintained.
5. And (4) model testing, namely utilizing a model network test sample picture generated by training to count the detected data and the classified data of the model.
6. And (4) optimizing the model, namely collecting undetected pictures and pictures with wrong classification, adding the pictures into a training sample to mark defects, and training the model again, testing the model and iteratively optimizing the model once and again.
As shown in fig. 2, the product inspection process based on the defect detection network and the defect determination network includes the following steps:
(1) and extracting the detection area, namely segmenting the area to be detected from the image by using a method based on traditional image processing, setting the gray value of the background area to be 0 for eliminating the interference of the background area, and finally transmitting the image with the gray value of the background area set to be 0 into a model for detection.
(2) And the defect detection network detects the defects of the transmitted image with the background area gray value set to be 0 by utilizing a trained defect detection model, transmits the image of the defect area to the defect judgment network for next defect classification if the defects are detected, and judges the product as an OK product if no suspicious defects are detected.
(3) The defect identification network carries out defect type identification on the transmitted suspicious defect area image, each suspicious defect can obtain a classification result, the verification project of the embodiment contains 6 types of defects such as dust, dirt, broken filaments, bubbles, scratches and broken edges, so that the classification result is the score of the defect belonging to each type of defect, and the defect type with the highest score is selected as the classification result.
(4) The defect specification management and control method includes that defect parameters which are interesting to customers and can be defined mathematically are calculated by utilizing a traditional image processing algorithm, such as region length (the length of a minimum circumscribed rectangle of a region) and region width (the width of the minimum circumscribed rectangle of the region) 2, region contrast (the difference value of a region average Gray value Gray1 and a background region average Gray value Gray2 in 5 pixels outside the region) and the like, if the defect parameters are within the management and control specifications, a product is judged to be an OK product, and if the defect parameters are outside the management and control specifications, the product is judged to be an NG product (dust, dirt and wool are judged to be dirty NG, and scratch, bubble and edge breakage are judged to be NG).
Therefore, the present invention uses a model based on ResNet18 for defect detection and a model based on ResNet101 for defect classification. The model based on the ResNet18 is excellent in defect segmentation and extraction capability for a glass-like flat background region, and can extract defects such as dust, dirt, broken filaments, bubbles, scratches, edge chipping, and the like. The model classification capability based on ResNet101 is excellent, and under the conditions that training samples such as dust, dirt, broken filaments, air bubbles, scratches and broken edges are sufficient and the samples are balanced, the trained classification model can accurately classify the dust, dirt, broken filaments, air bubbles, scratches and broken edges.
In addition, the invention utilizes the traditional image processing algorithm to carry out post-processing on the detection result of the deep learning model, carries out blob analysis and Cluster analysis, calculates the Length, Width, Aspect ratio Aspect, Contrast and Cluster result of the defects of each defect, and utilizes the parameters to carry out independent control on various defects by process personnel, thereby meeting the requirement of the specification of classifying and controlling the defects of customers.
While particular embodiments of the present invention have been described in the foregoing specification, various modifications and alterations to the previously described embodiments will become apparent to those skilled in the art from this description without departing from the spirit and scope of the invention.
Claims (7)
1. A glass surface defect detection method based on deep learning is characterized in that: the method comprises the steps of training a network model by using a residual error network ResNet18 and a residual error network ResNet101, detecting and judging the defects of a sample image through the trained network model after the training of the network model is finished, and managing and controlling the defect specification.
2. The deep learning-based glass surface defect detection method of claim 1, wherein: the training process of the network model comprises the following steps,
1) collecting various defect samples according to the defect detection requirement, and carrying out image acquisition on the samples by using an imaging system;
2) segmenting a region to be detected from the image, setting the gray value of a background region to be 0, and adding the gray value to a training sample image set for defect marking;
3) labeling the sample image;
4) model training, including training of a defect detection network and training of a defect discrimination network;
5) testing a sample picture by using a network model generated by training, and counting detected data and classified data of the model;
6) and collecting undetected pictures and pictures with wrong classification, adding the pictures into a training sample to mark defects, training the model again, testing the model, and iteratively optimizing the model once and again.
3. The deep learning-based glass surface defect detection method of claim 2, wherein: in the step 3), in the labeling process, a body projection integral labeling method is adopted, specifically, the body and the projection are connected and labeled into a whole by using a labeling brush.
4. The deep learning-based glass surface defect detection method of claim 2, wherein: in the step 4), firstly training a defect detection network, and then training a defect judgment network; the defect detection network uses a residual error network ResNet18 to segment all defects from the image, and the defect discrimination network uses a residual error network ResNet101 to classify the defects.
5. The deep learning-based glass surface defect detection method of claim 2, wherein: the trained defect detection model is used for detecting the defects of the image with the gray value of 0 in the background area; if the defects are detected, transmitting the images of the defect areas to a defect judgment network for next defect classification; if no suspicious defect is detected, the product is judged to be an OK product.
6. The deep learning-based glass surface defect detection method of claim 2, wherein: the defect discrimination network discriminates the defect type of the transmitted suspicious defect area image, and each suspicious defect can obtain a classification result.
7. The deep learning-based glass surface defect detection method of claim 1, wherein: the defect specification control mode is that a defect parameter which can be defined by mathematics is calculated by utilizing an image processing algorithm; if the defect parameters are within the control specification, determining as an OK product; and if the standard is out of the regulation specification, judging as an NG product.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010673882.3A CN111652883B (en) | 2020-07-14 | 2020-07-14 | Glass surface defect detection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010673882.3A CN111652883B (en) | 2020-07-14 | 2020-07-14 | Glass surface defect detection method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111652883A true CN111652883A (en) | 2020-09-11 |
| CN111652883B CN111652883B (en) | 2024-02-13 |
Family
ID=72348818
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010673882.3A Active CN111652883B (en) | 2020-07-14 | 2020-07-14 | Glass surface defect detection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111652883B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298765A (en) * | 2021-05-19 | 2021-08-24 | 广东工业大学 | Glass panel surface defect detection method based on image processing and deep learning |
| CN113658108A (en) * | 2021-07-22 | 2021-11-16 | 西南财经大学 | Glass defect detection method based on deep learning |
| CN114441547A (en) * | 2022-04-11 | 2022-05-06 | 深圳市睿阳精视科技有限公司 | Intelligent household appliance cover plate defect detection method |
| CN114577816A (en) * | 2022-01-18 | 2022-06-03 | 广州超音速自动化科技股份有限公司 | Hydrogen fuel bipolar plate detection method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2018100321A4 (en) * | 2018-03-15 | 2018-04-26 | Chen, Jinghan Mr | Person ReID method based on metric learning with hard mining |
| CN109754392A (en) * | 2018-12-20 | 2019-05-14 | 上海集成电路研发中心有限公司 | A kind of device and method that defect image automatically analyzes |
| CN110175988A (en) * | 2019-04-25 | 2019-08-27 | 南京邮电大学 | Cloth defect inspection method based on deep learning |
| CN110400315A (en) * | 2019-08-01 | 2019-11-01 | 北京迈格威科技有限公司 | A kind of defect inspection method, apparatus and system |
| CN110458019A (en) * | 2019-07-10 | 2019-11-15 | 哈尔滨工程大学 | The waterborne target detection method for excluding inverted image interference under rare cognition sample conditions |
| CN110530875A (en) * | 2019-08-29 | 2019-12-03 | 珠海博达创意科技有限公司 | A kind of FPCB open defect automatic detection algorithm based on deep learning |
| CN111179251A (en) * | 2019-12-30 | 2020-05-19 | 上海交通大学 | Defect detection system and method based on twin neural network and by utilizing template comparison |
-
2020
- 2020-07-14 CN CN202010673882.3A patent/CN111652883B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2018100321A4 (en) * | 2018-03-15 | 2018-04-26 | Chen, Jinghan Mr | Person ReID method based on metric learning with hard mining |
| CN109754392A (en) * | 2018-12-20 | 2019-05-14 | 上海集成电路研发中心有限公司 | A kind of device and method that defect image automatically analyzes |
| CN110175988A (en) * | 2019-04-25 | 2019-08-27 | 南京邮电大学 | Cloth defect inspection method based on deep learning |
| CN110458019A (en) * | 2019-07-10 | 2019-11-15 | 哈尔滨工程大学 | The waterborne target detection method for excluding inverted image interference under rare cognition sample conditions |
| CN110400315A (en) * | 2019-08-01 | 2019-11-01 | 北京迈格威科技有限公司 | A kind of defect inspection method, apparatus and system |
| CN110530875A (en) * | 2019-08-29 | 2019-12-03 | 珠海博达创意科技有限公司 | A kind of FPCB open defect automatic detection algorithm based on deep learning |
| CN111179251A (en) * | 2019-12-30 | 2020-05-19 | 上海交通大学 | Defect detection system and method based on twin neural network and by utilizing template comparison |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298765A (en) * | 2021-05-19 | 2021-08-24 | 广东工业大学 | Glass panel surface defect detection method based on image processing and deep learning |
| CN113298765B (en) * | 2021-05-19 | 2023-11-07 | 广东工业大学 | Glass panel surface defect detection method based on image processing and deep learning |
| CN113658108A (en) * | 2021-07-22 | 2021-11-16 | 西南财经大学 | Glass defect detection method based on deep learning |
| CN114577816A (en) * | 2022-01-18 | 2022-06-03 | 广州超音速自动化科技股份有限公司 | Hydrogen fuel bipolar plate detection method |
| CN114441547A (en) * | 2022-04-11 | 2022-05-06 | 深圳市睿阳精视科技有限公司 | Intelligent household appliance cover plate defect detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111652883B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111652883A (en) | Glass surface defect detection method based on deep learning | |
| CN113989279B (en) | Plastic film quality detection method based on artificial intelligence and image processing | |
| CN110097034B (en) | Intelligent face health degree identification and evaluation method | |
| CN106952258B (en) | A kind of bottle mouth defect detection method based on gradient orientation histogram | |
| CN107771212B (en) | Colony contrast collection | |
| CN109490316A (en) | A kind of surface defects detection algorithm based on machine vision | |
| CN109508664B (en) | Dish identification pricing method based on deep learning | |
| CN110490842B (en) | Strip steel surface defect detection method based on deep learning | |
| CN105844621A (en) | Method for detecting quality of printed matter | |
| CN104537651B (en) | Proportion detecting method and system for cracks in road surface image | |
| CN114972356B (en) | Plastic product surface defect detection and identification method and system | |
| CN115311294B (en) | Glass bottle body flaw identification and detection method based on image processing | |
| CN108318494B (en) | The red online vision detection and classification devices and methods therefor for proposing fruit powder | |
| CN108052886A (en) | A kind of puccinia striiformis uredospore programming count method of counting | |
| CN103852034B (en) | A kind of method for measuring perendicular | |
| CN111126393A (en) | Vehicle appearance refitting judgment method and device, computer equipment and storage medium | |
| CN111160451A (en) | Flexible material detection method and storage medium thereof | |
| CN109115775A (en) | A kind of betel nut level detection method based on machine vision | |
| CN114998205A (en) | Method for detecting foreign matters in bottle in liquid filling process based on optical means | |
| CN102298780B (en) | Method for detecting shadow of color image | |
| CN114004826A (en) | Visual sense-based method for detecting appearance defects of medical injection molding part | |
| CN117330582A (en) | Polymer PE film surface crystal point detecting system | |
| WO2023098187A1 (en) | Processing method, processing apparatus, and processing system | |
| CN116740449A (en) | Shaving form detection method and system based on AI (advanced technology attachment) computer vision technology | |
| CN105973910A (en) | Structure and texture characteristic-based lamp tube quality detection and flaw classification method and system thereof |
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 |