CN108445011B - Defect detection system and method based on deep learning - Google Patents
Defect detection system and method based on deep learning Download PDFInfo
- Publication number
- CN108445011B CN108445011B CN201810200744.6A CN201810200744A CN108445011B CN 108445011 B CN108445011 B CN 108445011B CN 201810200744 A CN201810200744 A CN 201810200744A CN 108445011 B CN108445011 B CN 108445011B
- Authority
- CN
- China
- Prior art keywords
- image
- deep learning
- flaw detection
- edge
- flaw
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a flaw detection system based on deep learning, which comprises: the system comprises a belt transmission device, a line scanning camera, a lens connected with the line scanning camera, a PC host and a cloud end; image processing software is arranged in the PC host, flaw detection and processing are carried out through the image processing software, processing results are displayed and uploaded to the cloud end, and big data analysis is carried out through the cloud end. The invention also discloses a flaw detection method based on deep learning, and the method introduces an AI algorithm of deep learning to identify and extract the flaw characteristics in the picture with high efficiency and high identification rate. The labor investment is low, and the maintenance cost of the detection performance of the image processing software is low. The compatibility is strong, and under the condition of facing product updating, the image processing software algorithm is not required to be developed additionally, and the flaw detection application of a new product can be met quickly only by collecting more product samples for learning and training.
Description
Technical Field
The invention relates to the technical field of flaw detection, in particular to a flaw detection system and method based on deep learning.
Background
In the industrial production process, flaw detection is an important step of a plurality of product quality detection links. The flaw detection device carries out flaw identification processing on a product surface image acquired by the industrial camera through image processing software, finds out flaws, and meanwhile effectively classifies and carries out subsequent processing on the flaws. The conventional image processing software has several problems:
firstly, image processing software has a plurality of open parameters, and can achieve better detection performance only by investing a great deal of effort in debugging.
Secondly, the universality and the function expansibility of the bottom algorithm of the image processing software are weak, and the new product and the new requirements of customers need to be redeveloped by personnel.
Therefore, the present inventors have earnestly demanded to conceive a new technology to improve the problems thereof.
Disclosure of Invention
In order to solve the technical problem, the invention provides a flaw detection system and method based on deep learning.
The technical scheme of the invention is as follows:
a deep learning based flaw detection system comprising:
the belt conveying device is used for conveying a product to be detected;
the line scanning camera and the lens connected with the line scanning camera are used for scanning a product to be detected on the belt transmission device and sending a collected surface image of the product to the PC host;
image processing software is arranged in the PC host, flaw detection and processing are carried out through the image processing software, and processing results are displayed and uploaded to the cloud end;
and the cloud end analyzes the big data.
Preferably, the image processing software comprises the following modules:
the image preprocessing module is used for preprocessing the image, and the preprocessing comprises but is not limited to gray level transformation and image cropping;
the prediction module is used for predicting through the convolutional neural network degree image to obtain a prediction result;
the processing module is used for processing the prediction result to obtain a processed picture;
and the display module is used for displaying the processed picture.
Preferably, the system further comprises a model base connected with the prediction module, and the model base is internally provided with a convolutional neural network model trained offline.
Preferably, the prediction result is a single-channel gray-scale map, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw.
Preferably, two parameters, namely a brightness threshold and an area threshold, are introduced into the processing module to process the prediction result.
A flaw detection method based on deep learning comprises the following steps:
s1: the line scanning camera scans a tested product on the belt transmission device and sends the collected surface image of the product to the image processing software;
s2: flaw detection and processing are carried out through image processing software, and meanwhile, processing results are displayed and uploaded to a cloud end;
s3: and the cloud end analyzes the big data.
Preferably, the step S2 specifically includes:
s21: preprocessing a picture, wherein the preprocessing comprises but is not limited to gray scale transformation of an image and cropping of the image;
s22: predicting through a convolutional neural network degree image to obtain a prediction result;
s23: processing the prediction result to obtain a processed picture;
s24: and displaying the processed picture.
Preferably, the step S22 specifically includes: and loading the convolution neural network model trained offline, predicting the preprocessed image, and obtaining a test result.
Preferably, the prediction result is a single-channel gray-scale map, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw.
Preferably, in the step S23, two parameters, namely, a brightness threshold and an area threshold, are introduced to process the prediction result.
By adopting the technical scheme, the invention at least comprises the following beneficial effects:
the flaw detection system and method based on deep learning introduce an AI algorithm of deep learning to identify and extract flaw features in pictures with high efficiency and high identification rate. And automatically acquiring, uploading the detection sorting data and the equipment state data to the cloud in real time, and analyzing the big data. The labor investment is low, and the maintenance cost of the detection performance of the image processing software is low. The compatibility is strong, and under the condition of facing product updating, the image processing software algorithm is not required to be developed additionally, and the flaw detection application of a new product can be met quickly only by collecting more product samples for learning and training.
Drawings
FIG. 1 is a schematic diagram of a fault detection system based on deep learning according to the present invention;
FIG. 2 is a flowchart of a defect detection method based on deep learning according to the present invention.
Wherein: the method comprises the following steps of 1-a line scanning camera, 2-a lens, 3-a light source, 4-a measuring center line, 5-a belt transmission device, 6-a tested product, 7-a PC host and 8-a cloud.
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 derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
As shown in fig. 1, a defect detection system based on deep learning according to the present embodiment includes:
a belt transmission device 5 for transmitting the product (silicon wafer) to be tested;
the line scanning camera 1 and the lens 2 connected with the line scanning camera are used for scanning a tested product 6 on the belt transmission device 5 and sending a collected surface image of the product to a PC host 7;
preferably, the device further comprises a light source 3 arranged on one side of the measuring center line 4 and used for supplementing light for the camera and ensuring the collection precision of the image.
Image processing software is arranged in the PC host 7, flaw detection and processing are carried out through the image processing software, and processing results are displayed and uploaded to the cloud end 8;
and the cloud end 8 analyzes the big data.
Preferably, the image processing software comprises the following modules:
the image preprocessing module is used for preprocessing the image, and the preprocessing comprises but is not limited to gray level transformation and image cropping;
the prediction module is used for predicting through the convolutional neural network degree image to obtain a prediction result;
the processing module is used for processing the prediction result to obtain a processed picture;
and the display module is used for displaying the processed picture.
Preferably, the system further comprises a model base connected with the prediction module, and the model base is internally provided with a convolutional neural network model trained offline.
Preferably, the prediction result is a single-channel gray-scale map, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw.
Preferably, two parameters, namely a brightness threshold and an area threshold, are introduced into the processing module to process the prediction result.
In this embodiment, the image processing software acquires the picture acquired by the line scan camera 1, and the picture needs to be preprocessed. The software image preprocessing module mainly comprises gray level transformation and image cutting of the image. The software adopts different cropping modes according to the distribution condition of the flaws on the image, so that the calculation amount is reduced, and the prediction time is shortened. The first is that the defects are distributed only at the edge positions of the image, and the second is that the defects are distributed only in a certain area of the image. The defects are distributed on the whole image, and then cutting is not performed.
In the first case, the software cuts the image into four small images, which are the images at the top, right, left and bottom edge positions on the image, while performing affine transformations to the images of the specified size.
In the second case, the software cuts out a specific region in the image, and performs affine transformation on the cut-out image to an image of a specified size.
And loading the convolution neural network model trained offline by the image processing software, predicting the cut image, wherein the prediction result is a single-channel gray image, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw. The larger the pixel value, the larger the score, indicating a greater probability that the location is a flaw.
And the image processing software introduces two parameters of a brightness threshold and an area threshold to process the prediction result. The brightness threshold is used to suppress the region with the lower score value. Pixel values less than this brightness threshold are not considered defective at that location. A pixel value greater than the brightness threshold is considered defective at that location. Some small area defect regions can be filtered out by the area threshold. Through the operation, whether the prediction result graph has defects or not can be judged, and therefore whether the piece of product is an OK piece or an NG piece is judged. For the NG piece, the position of the flaw on the original image before cropping can be converted by using the affine transformation matrix during cropping. For the OK picture, the software interface only displays the original picture, and for the NG picture, the software interface marks the position and the area size of the flaw while displaying the original picture.
The detection result of each product can be uploaded to the cloud end 8, a large amount of data can be integrated by the cloud end 8, and the NG and OK quantities of the whole batch of chips are analyzed, so that software users can perform data analysis.
In the embodiment, an AI algorithm of deep learning is introduced into image processing software to detect the surface flaws of the product; the flaw detection capability of the image processing software is greatly improved, and the false detection rate is reduced; the detection performance of the image processing software is easy to maintain, a large number of parameters do not need to be debugged, and only a model file trained by a deep learning technology needs to be maintained by a person; the universality of the image processing software is improved, after the product is updated, the defect detection application of a new product can be quickly met only by collecting more product samples for learning training and updating the model file without consuming manpower for redevelopment.
Example 2
As shown in fig. 2, on the basis of the embodiment, the embodiment provides a flaw detection method based on deep learning, which includes the following steps:
s1: the line scanning camera 1 scans a tested product 6 on the belt transmission device 5 and sends a collected product surface image to an image processing software;
s2: flaw detection and processing are carried out through image processing software, and meanwhile, processing results are displayed and uploaded to a cloud end 8;
s3: and the cloud end 8 performs big data analysis.
Preferably, the step S2 specifically includes:
s21: preprocessing a picture, wherein the preprocessing comprises but is not limited to gray scale transformation of an image and cropping of the image;
s22: predicting through a convolutional neural network degree image to obtain a prediction result;
s23: processing the prediction result to obtain a processed picture;
s24: and displaying the processed picture.
Preferably, the step S22 specifically includes: and loading the convolution neural network model trained offline, predicting the preprocessed image, and obtaining a test result.
Preferably, the prediction result is a single-channel gray-scale map, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw.
Preferably, in the step S23, two parameters, namely, a brightness threshold and an area threshold, are introduced to process the prediction result.
In this embodiment, the image processing software acquires the picture acquired by the line scan camera 1, and the picture needs to be preprocessed. The software image preprocessing module mainly comprises gray level transformation and image cutting of the image. The software adopts different cropping modes according to the distribution condition of the flaws on the image, so that the calculation amount is reduced, and the prediction time is shortened. The first is that the defects are distributed only at the edge positions of the image, and the second is that the defects are distributed only in a certain area of the image. The defects are distributed on the whole image, and then cutting is not performed.
In the first case, the software cuts the image into four small images, which are the images at the top, right, left and bottom edge positions on the image, while performing affine transformations to the images of the specified size.
In the second case, the software cuts out a specific region in the image, and performs affine transformation on the cut-out image to an image of a specified size.
And loading the convolution neural network model trained offline by the image processing software, predicting the cut image, wherein the prediction result is a single-channel gray image, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw. The larger the pixel value, the larger the score, indicating a greater probability that the location is a flaw.
And the image processing software introduces two parameters of a brightness threshold and an area threshold to process the prediction result. The brightness threshold is used to suppress the region with the lower score value. Pixel values less than this brightness threshold are not considered defective at that location. A pixel value greater than the brightness threshold is considered defective at that location. Some small area defect regions can be filtered out by the area threshold. Through the operation, whether the prediction result graph has defects or not can be judged, and therefore whether the piece of product is an OK piece or an NG piece is judged. For the NG piece, the position of the flaw on the original image before cropping can be converted by using the affine transformation matrix during cropping. For the OK picture, the software interface only displays the original picture, and for the NG picture, the software interface marks the position and the area size of the flaw while displaying the original picture.
The detection result of each product can be uploaded to the cloud end 8, a large amount of data can be integrated by the cloud end 8, and the NG and OK quantities of the whole batch of chips are analyzed, so that software users can perform data analysis.
In the embodiment, an AI algorithm of deep learning is introduced into image processing software to detect the surface flaws of the product; the flaw detection capability of the image processing software is greatly improved, and the false detection rate is reduced; the detection performance of the image processing software is easy to maintain, a large number of parameters do not need to be debugged, and only a model file trained by a deep learning technology needs to be maintained by a person; the universality of the image processing software is improved, after the product is updated, the defect detection application of a new product can be quickly met only by collecting more product samples for learning training and updating the model file without consuming manpower for redevelopment.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A system for fault detection based on deep learning, comprising:
the belt conveying device is used for conveying a product to be detected;
the line scanning camera and the lens connected with the line scanning camera are used for scanning a product to be detected on the belt transmission device and sending a collected surface image of the product to the PC host;
image processing software is arranged in the PC host, flaw detection and processing are carried out through the image processing software, and processing results are displayed and uploaded to the cloud end;
wherein the image processing software comprises the following modules:
the image preprocessing module is used for preprocessing the image, and the preprocessing comprises but is not limited to gray level transformation and image cropping; the preprocessing module cuts four small edge pictures from the upper edge, the lower edge, the left edge and the right edge of the image with flaws on the edge, and cuts the inner part of the image with flaws on the inner side of the edge into small inner pictures; converting the cut small pictures into pictures with processing sizes according to an affine transformation matrix;
the prediction module is used for predicting through the convolutional neural network degree image to obtain a prediction result;
the processing module is used for processing the prediction result, and converting the position and the area of the flaw on the original image by using the affine transformation matrix during cutting of the NG picture to obtain a processed picture;
the display module is used for displaying the processed picture; and the cloud end analyzes the big data.
2. The deep learning based flaw detection system of claim 1, wherein: the system also comprises a model base which is connected with the prediction module, and a convolutional neural network model which is trained off line is arranged in the model base.
3. The deep learning based flaw detection system of claim 1, wherein: the prediction result is a single-channel gray image, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw.
4. The deep learning based flaw detection system of claim 1, wherein: and the processing module introduces two parameters of a brightness threshold and an area threshold to process the prediction result.
5. A flaw detection method based on deep learning is characterized by comprising the following steps:
s1: the line scanning camera scans a tested product on the belt transmission device and sends the collected surface image of the product to the image processing software;
s2: flaw detection and processing are carried out through image processing software, and meanwhile, processing results are displayed and uploaded to a cloud end; the image processing software introduces two parameters of a brightness threshold and an area threshold for processing; the step S2 specifically includes:
s21: preprocessing a picture, wherein the preprocessing comprises but is not limited to gray scale transformation of an image and cropping of the image; cutting four small edge pictures from the upper edge, the lower edge, the left edge and the right edge of the image with the flaws on the edge, and cutting internal small pictures from the interior of the image with the flaws on the inner side of the edge by a preprocessing module; converting the cut small pictures into pictures with processing sizes according to an affine transformation matrix;
s22: predicting through a convolutional neural network degree image to obtain a prediction result;
s23: processing the prediction result, and converting the position and the area of the flaw on the original image by using an affine transformation matrix during cutting of the NG picture to obtain a processed picture;
s24: displaying the processed picture;
s3: and the cloud end analyzes the big data.
6. The method for detecting flaws based on deep learning of claim 5, wherein the step S22 specifically comprises: and loading the convolution neural network model trained offline, predicting the preprocessed image, and obtaining a test result.
7. The flaw detection method based on deep learning of claim 5 or 6, wherein: the prediction result is a single-channel gray image, the pixel values of the image are distributed in the range of 0-255, and the size of each pixel value represents the score of the position as a flaw.
8. The flaw detection method based on deep learning of claim 5 or 6, wherein: in step S23, the two parameters of the brightness threshold and the area threshold are introduced to process the prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810200744.6A CN108445011B (en) | 2018-03-12 | 2018-03-12 | Defect detection system and method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810200744.6A CN108445011B (en) | 2018-03-12 | 2018-03-12 | Defect detection system and method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108445011A CN108445011A (en) | 2018-08-24 |
| CN108445011B true CN108445011B (en) | 2021-06-11 |
Family
ID=63194102
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810200744.6A Active CN108445011B (en) | 2018-03-12 | 2018-03-12 | Defect detection system and method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108445011B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109903267B (en) * | 2019-01-22 | 2021-01-05 | 江苏恒力化纤股份有限公司 | Method for testing network wire network degree based on image processing technology |
| CN111598827A (en) * | 2019-02-19 | 2020-08-28 | 富泰华精密电子（郑州）有限公司 | Appearance flaw detection method, electronic device and storage medium |
| CN109738595A (en) * | 2019-03-07 | 2019-05-10 | 福建工程学院 | A kind of system and method based on Internet of Things detection transformer fault |
| CN109975308B (en) * | 2019-03-15 | 2023-03-24 | 维库（厦门）信息技术有限公司 | Surface detection method based on deep learning |
| CN110070521A (en) * | 2019-03-19 | 2019-07-30 | 广东工业大学 | A kind of 3D printing model flaw anticipation system and method for view-based access control model nerve study |
| CN110349108B (en) * | 2019-07-10 | 2022-07-26 | 北京字节跳动网络技术有限公司 | Method, apparatus, electronic device, and storage medium for processing image |
| CN113358020A (en) * | 2020-03-05 | 2021-09-07 | 青岛海尔工业智能研究院有限公司 | Machine vision detection system and method |
| CN111340798A (en) * | 2020-03-16 | 2020-06-26 | 浙江一木智能科技有限公司 | Application of deep learning in product appearance flaw detection |
| CN111784644A (en) * | 2020-06-11 | 2020-10-16 | 上海布眼人工智能科技有限公司 | Printing defect detection method and system based on deep learning |
| KR102259519B1 (en) * | 2020-08-28 | 2021-06-03 | 주식회사농심 | cutting error detecting system for pack unti |
| CN113189109A (en) * | 2021-01-15 | 2021-07-30 | 深圳锦绣创视科技有限公司 | Flaw judgment system and flaw judgment method based on artificial intelligence |
| KR102670982B1 (en) * | 2023-12-15 | 2024-05-30 | 재단법인 포항산업과학연구원 | Defect detecting method, defect detecting device about color steel plate, and computer program |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103411974A (en) * | 2013-07-10 | 2013-11-27 | 杭州赤霄科技有限公司 | Cloud big data-based planar material detection remote system and cloud big data-based planar material detection method |
| CN106290388A (en) * | 2016-08-03 | 2017-01-04 | 国网山东省电力公司电力科学研究院 | A kind of insulator breakdown automatic testing method |
| CN107123114A (en) * | 2017-04-21 | 2017-09-01 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of cloth defect inspection method and device based on machine learning |
-
2018
- 2018-03-12 CN CN201810200744.6A patent/CN108445011B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103411974A (en) * | 2013-07-10 | 2013-11-27 | 杭州赤霄科技有限公司 | Cloud big data-based planar material detection remote system and cloud big data-based planar material detection method |
| CN106290388A (en) * | 2016-08-03 | 2017-01-04 | 国网山东省电力公司电力科学研究院 | A kind of insulator breakdown automatic testing method |
| CN107123114A (en) * | 2017-04-21 | 2017-09-01 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of cloth defect inspection method and device based on machine learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108445011A (en) | 2018-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108445011B (en) | Defect detection system and method based on deep learning | |
| CN110148130B (en) | Method and device for detecting part defects | |
| CN111080622B (en) | Neural network training method, workpiece surface defect classification and detection method and device | |
| US8139841B2 (en) | Visual inspection method and apparatus and image analysis system | |
| CN113095438B (en) | Wafer defect classification method, device and system thereof, electronic equipment and storage medium | |
| CN109671058B (en) | Defect detection method and system for large-resolution image | |
| JP5168215B2 (en) | Appearance inspection device | |
| CN107966454A (en) | A kind of end plug defect detecting device and detection method based on FPGA | |
| CN113096119B (en) | Method and device for classifying wafer defects, electronic equipment and storage medium | |
| KR102168724B1 (en) | Method And Apparatus for Discriminating Normal and Abnormal by using Vision Inspection | |
| CN112862770B (en) | Defect analysis and diagnosis system, method and device based on artificial intelligence | |
| CN111652098B (en) | Product surface defect detection method and device | |
| CN117635565B (en) | Semiconductor surface defect detection system based on image recognition | |
| CN111242899B (en) | Image-based flaw detection method and computer-readable storage medium | |
| CN114742818A (en) | Visual defect detection system based on industrial area array camera and detection method thereof | |
| CN117392112A (en) | Method, device and system for detecting wafer surface defects | |
| CN113111903A (en) | Intelligent production line monitoring system and monitoring method | |
| US20240095983A1 (en) | Image augmentation techniques for automated visual inspection | |
| KR102390058B1 (en) | Data generation device and method for led panel defect detection | |
| CN113177925A (en) | Method for nondestructive detection of fruit surface defects | |
| CN112967224A (en) | Electronic circuit board detection system, method and medium based on artificial intelligence | |
| CN111458344A (en) | Mask defect visual detection method, equipment and storage medium | |
| CN111028250A (en) | Real-time intelligent cloth inspecting method and system | |
| CN116258908A (en) | Ground disaster prediction evaluation classification method based on unmanned aerial vehicle remote sensing image data | |
| CN112967258B (en) | Display defect detection method and device for watch and computer readable storage medium |
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 |