CN117197054A - Display panel dead pixel detection method and system based on artificial intelligence - Google Patents
Display panel dead pixel detection method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN117197054A CN117197054A CN202311061538.9A CN202311061538A CN117197054A CN 117197054 A CN117197054 A CN 117197054A CN 202311061538 A CN202311061538 A CN 202311061538A CN 117197054 A CN117197054 A CN 117197054A
- Authority
- CN
- China
- Prior art keywords
- vgg
- display panel
- artificial intelligence
- dead pixel
- network
- 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.)
- Withdrawn
Links
- 238000001514 detection method Methods 0.000 title claims description 26
- 238000013473 artificial intelligence Methods 0.000 title claims description 20
- 238000000034 method Methods 0.000 claims abstract description 23
- 238000012545 processing Methods 0.000 claims description 9
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 230000005012 migration Effects 0.000 claims description 5
- 238000013508 migration Methods 0.000 claims description 5
- 239000013598 vector Substances 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 11
- 238000012423 maintenance Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 6
- 230000006872 improvement Effects 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention belongs to the technical method and system in the field of industrial manufacturing, and mainly relates to a method and system for detecting dead pixels of a display panel. The method is independent of a production system of the display panel, does not depend on circuits and electronic devices of the display panel, can independently discover and identify bad points of the display panel from the perspective of a third party, and has wide application prospects in the fields of production, sales and maintenance of the display panel.
Description
Technical Field
The invention belongs to the technical method and system in the field of industrial manufacturing, and mainly relates to a method and system for detecting dead pixels of a display panel. The method is independent of a production system of the display panel, does not depend on circuits and electronic devices of the display panel, can independently discover and identify bad points of the display panel from the perspective of a third party, and has wide application prospects in the fields of production, sales and maintenance of the display panel.
Background
The display panel is applied to various fields of production and life, and great convenience is brought to the production and life of people. According to the disclosure of related departments, the capacity of the display panel in China reaches 2 hundred million square meters, and the production process of the display panel is greatly improved. Most of the processes have been automated, but in many factories, the final detection of the display panel products is still mainly performed manually, and quality inspection personnel observe bad spots (bright spots, dark spots and the like) in the display panel when the display panel is adjusted to a monochromatic state, so that the detection efficiency is reduced, the labor cost is greatly increased, the manufacturing cost of the products is increased, and the competitiveness of the products is further reduced.
At present, two methods for detecting the display panel exist, the first method is to detect by using instruments and equipment, and the method mainly detects whether a physical device can work normally or not and whether a circuit is normal or not, and the detection method mainly finds out a relatively large physical fault in the display panel; the second detection method is to use manpower to detect the display panel, and is mainly used for detecting the display effect, and the bad points in the display panel are detected, so that the display panel is generally enabled to display a single-color mode, then a quality inspector observes the display effect through amplifying equipment, and the validity of the method mainly depends on the capability of the quality inspector, so that the efficiency is low.
At present, the artificial intelligence has made remarkable progress in the field of image processing, and at present, the human face recognition has been exceeded, and the application of the artificial intelligence and the image processing technology to industrial production can necessarily achieve remarkable effects. The invention realizes the automatic detection of the display panel mainly through image processing and artificial intelligence, thereby improving the detection efficiency and accuracy.
The invention discloses a method and a system for detecting fault points of a display panel by utilizing artificial intelligence and image processing technology.
Disclosure of Invention
The invention comprises the following steps: a display panel dead pixel detection method and system based on artificial intelligence are disclosed, which are realized by an industrial microscope, a high-resolution camera and an artificial intelligence model.
The invention discloses a display panel dead pixel detection method and system based on artificial intelligence.
The data acquisition part mainly comprises an industrial microscope, a high-resolution camera and a control system, the overall structure is shown in figure 1, the display panel is controlled by the control module to display black, white, red, green and blue modes respectively in a pure color mode, then the microscope (magnifier) amplifies the image, and the high-resolution camera shoots pictures in different modes through the magnifier to complete data acquisition.
The data acquisition image is transmitted to a data processing part, and the data processing mainly completes fault detection through a VGG16 network with 5 improved single channels. The 5 improved VGG16 networks are respectively marked as VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5, the five networks respectively detect photos in black, white, red, green and blue pure color modes, the VGG-1, the VGG-2, the VGG-3, the VGG-4 and the VGG-5 are obtained by performing migration learning on the VGG16 network, and the data of the migration learning is a picture data set of a fault screen with labels, which is accumulated in production.
VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 are connected in parallel and are respectively detected, each network has independent output, each network output '1' indicates that the network judges that the display panel corresponding to the current picture has faults, and otherwise '0' indicates that the display panel corresponding to the picture judges that the display panel has no faults. The 5 networks are connected in parallel, the structure is shown in fig. 2, a five-dimensional vector is output, and a ' 1 ' and a ' 0 ' of each dimension respectively indicate whether the corresponding detection panel has a corresponding fault, for example, 10000 ' indicates that VGG-1 detects the fault, namely, when the display panel has the fault in a pure white mode, for example, black spots or bright spots of other colors appear.
VGG-1, VGG-2, VGG-3, VGG-4, and VGG-5 networks used in the present invention are network models based on VGG16 improvements. Improvements to VGG16 have mainly been:
the VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 networks used in the invention have only 1 input channel, and because the input pictures are all shot in a single-color mode, three channels are not needed, and a plurality of networks respectively process single-color images, a better effect can be obtained even in the case of single channels.
And finally, the second VGG16 is provided with three full connection layers, the last two full connection layers are modified in VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 networks, the last two full connection layers are changed into 128 neurons, corresponding weight vectors are correspondingly modified, and other structures are not modified. The VGG16 is used for multi-object identification, and because of various identified objects, the neurons of the final full-connection layer are more, the VGG16 is mainly used for identifying faults of the display panel, the image is relatively simple, so that the VGG16 does not need more neurons, after a part of neurons are simplified, the complexity of a model can be reduced, the operation efficiency of the system is improved, and the faults of the model identification display panel are not reduced.
Another object of the present invention is to provide a computer-readable storage medium, which includes instructions that, when executed on a computer and a corresponding system, cause the computer and the related system to execute the display panel fault point detection implementation method.
Another object of the present invention is to provide a display panel fault detection implementation control system for implementing the display panel fault detection implementation method.
In summary, the invention has the advantages and effects that:
the method is different from the traditional method, the faults of the display panels are not observed and identified by people, the display panels are amplified by an industrial microscope, then the images are shot by a high-resolution camera, then the images are transmitted to an automatic identification system based on artificial intelligence, the identification system identifies the images by 5 optimized VGG16 networks, each network identifies a defect in a solid color state, and finally the display faults of each display panel, namely the fault types, are comprehensively obtained.
The invention is characterized in that the invention can realize comprehensive automatic fault detection without manual intervention, and secondly, the fault detection of the display panel is not dependent on the detection of the fault of the display panel electronic device, the fault of the display panel is detected only from the display effect of the display panel, which is closer to the actual demand of products, and furthermore, the modified VGG16 network has better image recognition capability, can well recognize various fault types of the display panel, and has high detection precision.
Drawings
FIG. 1 data acquisition architecture
FIG. 2 data processing architecture
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The work of the data acquisition part is uniformly managed under the control module, under the uniform management of the control module, the display panel is firstly adjusted to display pure white, then the display panel is enlarged by an industrial microscope (magnifying glass), the image enlarged by the microscope is shot by a high-resolution camera, and as the view field of the microscope is smaller, all contents of the display panel can not be shot at one time, the image can be finally synthesized into a complete image by shooting for many times and then splicing, then the spliced image is subjected to edge cutting treatment, and the area displayed by the non-display panel around the display panel is cut off to form the image under the condition of pure white; and then the control module adjusts the display panel into other solid-color modes, the industrial microscope and the high-resolution camera work in a combined mode, six pictures of the same display panel are collected, the six pictures are sent to six improved VGG16 neural network models of the identification module together, the improved deep learning model identifies the image, and then bad points of the display panel are found.
The VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 networks used in the invention are network models based on VGG16 improvement, the modification mainly comprises two parts, the first modification part is that the number of input channels is changed to 1, the second modification part is that the number of neurons of the last two full-connection layers is changed to 128, the network structures of VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 are identical, but after the network is modified, migration learning is respectively carried out on the 5 networks respectively and independently, for example, the network of VGG-1 is obtained by adopting pure white marked data for migration learning; when VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 are trained through the corresponding data sets, the images are combined together, and the images under different pure color conditions are respectively identified, so that whether the corresponding dead pixels exist in the display panel is judged.
VGG-1, VGG-2, VGG-3, VGG-4, and VGG-5 each network outputs a string of "0" and "1" by the last softmax layer, five separate networks output 5 bits of "0" and "1", and the occurrence of a "1" at the corresponding position of the 5 bits of string indicates that the display panel has a corresponding failure, such as: the first bit of 10000 is 1, which indicates that the VGG-1 network detects the pure white picture to find the fault, and indicates that the corresponding display panel has bad points, namely the display panel has corresponding faults.
In the present invention, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When used in whole or in part, is implemented in the form of a computer program product comprising one or more computer instructions. When loaded or executed on a computer, produces a flow or function in accordance with embodiments of the present invention, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.
The above description should not be taken as limiting the invention, but rather should be construed to cover all modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention.
Claims (10)
1. The invention discloses a display panel dead pixel detection method and system based on artificial intelligence, and discloses a method and system for realizing display panel dead pixel detection through an industrial microscope, a high-resolution camera and an artificial intelligence model.
2. The display panel dead pixel detection method and system data acquisition part based on artificial intelligence as claimed in claim 1 comprises an industrial microscope, a high-resolution camera and a control system, wherein the overall structure is shown in figure 1.
3. According to the artificial intelligence-based display panel dead pixel detection method and system as claimed in claim 1 and claim 2, the display panel respectively displays black, white, red, green and blue modes in a pure color mode under the control of the control module, then a microscope (magnifier) amplifies the image, and a high-resolution camera shoots pictures in different modes through the magnifier to complete data acquisition.
4. The method and system for detecting the dead pixel of the display panel based on the artificial intelligence according to claim 1, wherein the data acquisition image is transmitted to a data processing part, and the data processing mainly completes the fault detection through a VGG16 network with 5 improved single channels.
5. The method and system for detecting dead pixel of display panel based on artificial intelligence according to claim 1 and claim 4, wherein 5 improved VGG16 networks are respectively designated as VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5, and the five networks respectively detect photos in black, white, red, green and blue pure color modes; VGG-1, VGG-2, VGG-3, VGG-4, and VGG-5 are all obtained by performing migration learning on the VGG16 network.
6. The method and system for detecting dead pixel of display panel according to claim 1 and claim 4, wherein the VGG-1, VGG-2, VGG-3, VGG-4, and VGG-5 networks are used with only 1 input channel.
7. The method and system for detecting a dead pixel of a display panel based on artificial intelligence according to claim 1 and claim 4, wherein in the VGG-1, VGG-2, VGG-3, VGG-4, and VGG-5 networks, the last two full connection layers of VGG16 are modified, the last two full connection layers become only 128 neurons, and the corresponding weight vectors are correspondingly modified.
8. According to the artificial intelligence based display panel dead pixel detection method and system as claimed in claim 1 and claim 4, VGG-1, VGG-2, VGG-3, VGG-4 and VGG-5 are connected in parallel to detect each network, each network has independent output, each network output '1' indicates that the network judges that the display panel corresponding to the current picture has a fault, otherwise '0' indicates that the display panel corresponding to the picture has no fault.
9. A computer readable storage medium comprising instructions that when executed on a computer cause the computer to perform the artificial intelligence based method and system for detecting a dead pixel of a display panel of claim 1.
10. A control system for realizing the artificial intelligence-based display panel dead pixel detection method and system as claimed in claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311061538.9A CN117197054A (en) | 2023-08-22 | 2023-08-22 | Display panel dead pixel detection method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311061538.9A CN117197054A (en) | 2023-08-22 | 2023-08-22 | Display panel dead pixel detection method and system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117197054A true CN117197054A (en) | 2023-12-08 |
Family
ID=89002623
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311061538.9A Withdrawn CN117197054A (en) | 2023-08-22 | 2023-08-22 | Display panel dead pixel detection method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117197054A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118521575A (en) * | 2024-07-19 | 2024-08-20 | 深圳市蔚来芯科技有限公司 | Data display system and method based on image processing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000346802A (en) * | 1999-03-26 | 2000-12-15 | Sony Corp | Inspecting device for inside of element and its method |
| CN1712944A (en) * | 2004-06-25 | 2005-12-28 | 鸿富锦精密工业（深圳）有限公司 | System and method for inspecting LCD panel |
| CN104200775A (en) * | 2014-09-22 | 2014-12-10 | 西安电子科技大学 | LED defective pixel treatment method |
| CN106645181A (en) * | 2017-02-24 | 2017-05-10 | 清华大学 | Microscopic vision-based roller grinding surface defect detecting system |
| US20220343481A1 (en) * | 2020-05-29 | 2022-10-27 | Boe Technology Group Co., Ltd. | Detection device of display panel and detection method thereof, electronic device and readable medium |
-
2023
- 2023-08-22 CN CN202311061538.9A patent/CN117197054A/en not_active Withdrawn
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000346802A (en) * | 1999-03-26 | 2000-12-15 | Sony Corp | Inspecting device for inside of element and its method |
| CN1712944A (en) * | 2004-06-25 | 2005-12-28 | 鸿富锦精密工业（深圳）有限公司 | System and method for inspecting LCD panel |
| CN104200775A (en) * | 2014-09-22 | 2014-12-10 | 西安电子科技大学 | LED defective pixel treatment method |
| CN106645181A (en) * | 2017-02-24 | 2017-05-10 | 清华大学 | Microscopic vision-based roller grinding surface defect detecting system |
| US20220343481A1 (en) * | 2020-05-29 | 2022-10-27 | Boe Technology Group Co., Ltd. | Detection device of display panel and detection method thereof, electronic device and readable medium |
Non-Patent Citations (1)
| Title |
|---|
| 蔡湘云;: "利用机器视觉的检测液晶屏幕坏点分析", 电子测试, no. 09, 5 May 2017 (2017-05-05) * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118521575A (en) * | 2024-07-19 | 2024-08-20 | 深圳市蔚来芯科技有限公司 | Data display system and method based on image processing |
| CN118521575B (en) * | 2024-07-19 | 2024-09-20 | 深圳市蔚来芯科技有限公司 | Data display system and method based on image processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN201489795U (en) | Display flaw automatic detecting device | |
| CN109544204B (en) | Shopping guide behavior analysis method based on lightweight multitask convolutional neural network | |
| CN116245962B (en) | Data extraction system and method for wireless transmission to blockchain server | |
| CN117197054A (en) | Display panel dead pixel detection method and system based on artificial intelligence | |
| CN114240939A (en) | Method, system, equipment and medium for detecting appearance defects of mainboard components | |
| CN117130186B (en) | LCD display screen flaw defect intelligent detection method | |
| CN116990993B (en) | LCD display panel quality detection method | |
| JP4143660B2 (en) | Image analysis method, image analysis apparatus, inspection apparatus, image analysis program, and computer-readable recording medium | |
| CN109856824A (en) | Display picture detection method and detection system thereof | |
| CN115578585A (en) | Industrial image anomaly detection method, system, computer device and storage medium | |
| WO2022158628A1 (en) | System for determining defect in display panel on basis of machine learning model | |
| CN106960188B (en) | Weather image classification method and device | |
| CN118093350A (en) | Remote management method and system for outdoor display screen data | |
| US7330580B2 (en) | System and method for inspecting an LCD panel | |
| WO2022255518A1 (en) | Device for determining defect in panel to be inspected using deep learning neural network model | |
| WO2022250190A1 (en) | Defect determination system for image inspection object by using deep learning model | |
| CN112257514B (en) | Infrared vision intelligent detection shooting method for equipment fault inspection | |
| CN116310953A (en) | Pontoon off-grid intelligent detection method of cold source interception net | |
| WO2019073615A1 (en) | Worksite image assessment device and worksite image assessment program | |
| CN115655666A (en) | Camera-based cascaded display module detection device and method | |
| CN113112487A (en) | Method, device, equipment and storage medium for identifying thermal defects | |
| CN110991361A (en) | Multi-channel multi-modal background modeling method for high-definition high-speed video | |
| CN109459872A (en) | A kind of mould group detection method and device using FPGA and multiple PCIe video cards | |
| US20240288718A1 (en) | Method of detecting defective pixels in electronic displays | |
| US20240029414A1 (en) | Methods and systems for inspecting components |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20231208 |
|
| WW01 | Invention patent application withdrawn after publication |