CN114742811B - SMT production line welding point defect rapid detection method and system based on improved Yolox - Google Patents
SMT production line welding point defect rapid detection method and system based on improved Yolox Download PDFInfo
- Publication number
- CN114742811B CN114742811B CN202210450021.8A CN202210450021A CN114742811B CN 114742811 B CN114742811 B CN 114742811B CN 202210450021 A CN202210450021 A CN 202210450021A CN 114742811 B CN114742811 B CN 114742811B
- Authority
- CN
- China
- Prior art keywords
- model
- production line
- yolox
- improved
- smt production
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 88
- 238000001514 detection method Methods 0.000 title claims abstract description 70
- 238000003466 welding Methods 0.000 title claims abstract description 68
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 62
- 238000012549 training Methods 0.000 claims abstract description 40
- 238000012360 testing method Methods 0.000 claims abstract description 27
- 238000003062 neural network model Methods 0.000 claims abstract description 26
- RVCKCEDKBVEEHL-UHFFFAOYSA-N 2,3,4,5,6-pentachlorobenzyl alcohol Chemical compound OCC1=C(Cl)C(Cl)=C(Cl)C(Cl)=C1Cl RVCKCEDKBVEEHL-UHFFFAOYSA-N 0.000 claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 8
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 210000000988 bone and bone Anatomy 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 229910000679 solder Inorganic materials 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
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
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30152—Solder
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a SMT production line welding point defect rapid detection method and a system based on improved Yolox, wherein the method comprises the steps of collecting welding point defect image data stored in production line AOI equipment for preprocessing, and dividing a training set and a testing set according to a proportion; improving the Yolox neural network structure, and adjusting the size output by the back bone part after the model is improved to match with the back network part to obtain an improved Yolox neural network model; performing model training and testing on the improved Yolox neural network by using a training set and a testing set to obtain an SMT production line welding point defect target detection model; the input of the SMT production line welding point defect target detection model is PCBA image, and the output is defect positioning and classification; the detection model is embedded into the AOI system of the SMT production line, so that the deployment difficulty of the neural network model is reduced. And the target detection model is used for carrying out target detection on PCBA welding spots on the SMT production line, so that the accuracy and the efficiency of quality detection of the welding spots on the SMT production line can be effectively improved.
Description
Technical Field
The invention belongs to the field of SMT quality detection, and particularly relates to a method and a system for rapidly detecting welding point defects of an SMT production line by improving Yolox.
Background
Nowadays, the high demands on the quality of production of electronic products, which are an integral part of the work and life of people, are not neglected. The defect detection of the PCBA welding spots of the SMT production line is improved from initial manual visual inspection to the detection of AOI equipment of machine substitution workers at present. However, the detection algorithm of the AOI equipment is to compare the product with the unique qualified image at the pixel level, so that the standard is single, the efficiency is low, the accuracy is in short, different standard qualified images are required to be set before production of different batches of production lines, and the generalization is extremely poor. With the rapid increase of computer computing power in recent years, the once neglected deep learning algorithm is rising again. The scheme of combining the deep learning technology with the machine vision for defect detection can greatly improve the overall production efficiency and the product quality.
Disclosure of Invention
The invention aims to solve the problems of low detection efficiency and low accuracy of the conventional AOI equipment, and provides a SMT production line welding point defect rapid detection method and system based on improved Yolox.
The technical scheme for realizing the aim of the invention is as follows:
an SMT production line welding point defect rapid detection method based on improved Yolox comprises the following steps:
1) Collecting welding spot defect image data stored in an SMT production line AOI device, preprocessing the image data, and dividing the preprocessed data into a training set and a testing set in proportion;
2) Improving the Yolox neural network model, and adjusting the size output by the back bone part after the model is improved to match with the neg network part to obtain an improved Yolox neural network model; the improvement method comprises the following steps:
2-1) the original backup of the Yolox neural network is CSPDarknet, and the CSPDarknet is combined with a part of the structure of ResNet 50;
2-2) after removal of Stage1 and Stage4 of ResNet50, reserve Resblock body4 of CSPDarknet and access behind Stage3 of ResNet 50;
2-3) matching the size of the Stage3 output matrix with the size of the Resblock body4 input matrix, and taking the output matrices of Stage2, stage3 and Resblock body as the feature map extracted by the backbody part;
2-4) adjusting the output channels of all the parts, specifically:
2-4-1) adjusting the Stage2 input size of ResNet50 to make its output channel 256;
2-4-2) adjusting the Stage3 input size of ResNet50 to an output channel of 512
2-4-3) adjusting the Resblock body4 input size of CSPDarknet to make the output channel 1024;
2-5) respectively connecting the three feature maps into modules corresponding to the neg parts of the Yolox to obtain an improved Yolox neural network model;
3) Inputting the large data set into the improved Yolox neural network model obtained in the step 2) to perform pre-training to obtain a pre-training model, inputting the training set obtained in the step 1) into the pre-training model to perform training on the model, inputting the test set into the trained model to perform testing on the model, and obtaining an SMT production line welding point defect target detection model after training and testing are completed;
4) And 3) performing welding spot defect detection on PCBA images produced by the SMT production line by using the model obtained in the step 3) to obtain a detection result, taking data containing welding spot defects in the detection result as newly collected welding spot defect image data, and repeating the steps 1) to 3).
In the step 1), the pretreatment is specifically:
1-1) extracting PCBA welding spot defect image data collected by an AOI device from a database of an SMT production line;
1-2) primarily screening the images according to the types and the number of defects contained in the collected images, so that the number of defects of each type is approximately equivalent;
1-3) marking the preliminarily screened PCBA welding spot defect image of the SMT production line by using labelimg, rolabelimg, labelme and vott image marking tools, wherein the information required to be marked comprises position positioning information and defect classification information;
1-4) dividing the marked data into a training set and a testing set according to a preset proportion.
An improved Yolox-based SMT production line weld defect rapid detection system comprising:
the data acquisition module is used for acquiring PCBA image data containing welding spot defects stored in the SMT production line AOI equipment, marking the image with information, and dividing the marked data into a training set and a testing set according to a preset proportion;
the model improvement module is used for improving the Yolox neural network model and adjusting the size output by the back bone part after the model improvement so as to match the network part, thereby obtaining an improved Yolox neural network model;
the model training module is used for inputting the training set and the testing set obtained by the data acquisition module into the improved Yolox neural network model for training and testing to obtain an SMT production line welding point defect target detection model;
the defect detection module is used for inputting PCBA images produced by the SMT production line into a welding point defect target detection model of the SMT production line to detect welding point defects and obtain detection results;
the model updating module is used for taking the data containing the welding spot defects in the detection result obtained by the defect detection module as newly collected welding spot defect image data and inputting the newly collected welding spot defect image data into the data acquisition module, and updating the data of the data acquisition module.
According to the SMT production line welding point defect rapid detection method and system based on improved Yolox, provided by the invention, the welding point defect detection is carried out by combining deep learning with a computer vision image, so that the detection accuracy can be greatly improved, and the Yolox neural network is improved, so that the parameter number of a detection model can be reduced, and the detection speed is improved; the detection system is embedded into the AOI system of the SMT production line, so that the deployment difficulty of the neural network model is reduced; by using the SMT production line welding point defect target detection model, the accuracy and efficiency of SMT production line welding point quality detection can be effectively improved.
Drawings
Fig. 1 is a schematic implementation flow chart of a method for rapidly detecting defects of an SMT production line based on improved Yolox according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a basic structure of a CSPDarknet network according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a basic structure of a res net50 network according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of an improved Yolox model according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an SMT production line welding point defect rapid detection system based on improved Yolox according to an embodiment of the present invention.
Detailed Description
The present invention will now be further illustrated with reference to the drawings and examples, but is not limited thereto.
Examples:
as shown in FIG. 1, the SMT production line welding point defect rapid detection method based on improved Yolox comprises the following steps:
1) Collecting welding spot defect image data stored in an SMT production line AOI device, preprocessing the image data, and dividing the preprocessed data into a training set and a testing set in proportion;
the pretreatment is specifically as follows:
1-1) extracting PCBA welding spot defect image data collected by an AOI device from a database of an SMT production line;
1-2) primarily screening the images according to the types and the number of defects contained in the collected images, so that the number of defects of each type is approximately equivalent;
1-3) marking the preliminarily screened PCBA welding spot defect image of the SMT production line by using labelimg, rolabelimg, labelme and vott image marking tools, wherein the information required to be marked comprises position positioning information and defect classification information;
1-4) dividing the marked data into a training set and a testing set according to a preset proportion.
2) Improving the Yolox neural network model, and adjusting the size output by the back bone part after the model is improved to match with the neg network part to obtain an improved Yolox neural network model; the improvement method comprises the following steps:
2-1) the original backup of the Yolox neural network is CSPDarknet, the structure is shown in figure 2, the structure is complex, and the training time is long, so that the structure is required to be simplified; resNet50 is one of the neural networks with simple structure and remarkable feature extraction capability, and the structure is shown in FIG. 3, so that CSPDarknet is combined with part of the structure of ResNet 50;
2-2) after removal of Stage1 and Stage4 of ResNet50, reserve Resblock body4 of CSPDarknet and access behind Stage3 of ResNet 50;
2-3) matching the size of the Stage3 output matrix with the size of the Resblock body4 input matrix, and taking the output matrices of Stage2, stage3 and Resblock body as the feature map extracted by the backbody part;
2-4) adjusting the output channels of all the parts, specifically:
2-4-1) adjusting the Stage2 input size of ResNet50 to make its output channel 256;
2-4-2) adjusting the Stage3 input size of ResNet50 to an output channel of 512
2-4-3) adjusting the Resblock body4 input size of CSPDarknet to make the output channel 1024;
2-5) respectively connecting the three feature maps into modules corresponding to the neg parts of the Yolox to obtain an improved Yolox neural network model, wherein the structure of the improved Yolox neural network model is shown in figure 4;
3) Inputting the large data set such as coco and VOC into the improved Yolox neural network model obtained in the step 2) to perform pre-training to obtain a pre-training model, inputting the training set obtained in the step 1) into the pre-training model to perform training, inputting the test set into the trained model to perform testing on the model, and obtaining the SMT production line welding point defect target detection model after training and testing are completed;
4) And 3) performing welding spot defect detection on PCBA images produced by the SMT production line by using the model obtained in the step 3) to obtain a detection result, taking data containing welding spot defects in the detection result as newly collected welding spot defect image data, and repeating the steps 1) to 3).
Performing defect detection by using an SMT production line welding point defect target detection model, specifically:
4-1) embedding the SMT production line welding point defect target detection model into an AOI equipment detection system of an SMT production line to replace the original traditional detection algorithm;
4-2) the operation of the SMT production line is unchanged along the original path, the real-time PCBA image acquired by the AOI equipment is transmitted into the system, and a welding point defect target detection model of the SMT production line is operated to obtain a detection result;
4-3) the detection result comprises position information and classification information of the welding spot defects on the PCBA.
As shown in fig. 5, a system for rapidly detecting defects of SMT production line solder joints based on improved Yolox, comprising:
the data acquisition module is used for acquiring PCBA image data containing welding spot defects stored in the SMT production line AOI equipment, marking the image with information, and dividing the marked data into a training set and a testing set according to a preset proportion;
the model improvement module is used for improving the Yolox neural network model and adjusting the size output by the back bone part after the model improvement so as to match the network part, thereby obtaining an improved Yolox neural network model;
the model training module is used for inputting the training set and the testing set obtained by the data acquisition module into the improved Yolox neural network model for training and testing to obtain an SMT production line welding point defect target detection model;
the defect detection module is used for inputting PCBA images produced by the SMT production line into a welding point defect target detection model of the SMT production line to detect welding point defects and obtain detection results;
the model updating module is used for taking the data containing the welding spot defects in the detection result obtained by the defect detection module as newly collected welding spot defect image data and inputting the newly collected welding spot defect image data into the data acquisition module, updating the data of the data acquisition module, and performing migration learning on the improved Yolox neural network model by the new data to obtain a new SMT production line welding spot defect target detection model; the original SMT welding spot defect target detection model is replaced to adapt to model incompatibility caused by new defect characteristics generated by production environment changes.
Claims (3)
1. The SMT production line welding point defect rapid detection method based on improved Yolox is characterized by comprising the following steps of:
1) Collecting welding spot defect image data stored in an SMT production line AOI device, preprocessing the image data, and dividing the preprocessed data into a training set and a testing set in proportion;
2) Improving the Yolox neural network model, and adjusting the size output by the back bone part after the model is improved to match with the neg network part to obtain an improved Yolox neural network model; the improvement method comprises the following steps:
2-1) the original backup of the Yolox neural network is CSPDarknet, and the CSPDarknet is combined with a part of the structure of ResNet 50;
2-2) after removal of Stage1 and Stage4 of ResNet50, reserve Resblock body4 of CSPDarknet and access behind Stage3 of ResNet 50;
2-3) matching the size of the Stage3 output matrix with the size of the Resblock body4 input matrix, and taking the output matrices of Stage2, stage3 and Resblock body as the feature map extracted by the backbody part;
2-4) adjusting the output channels of all the parts, specifically:
2-4-1) adjusting the Stage2 input size of ResNet50 to make its output channel 256;
2-4-2) adjusting the Stage3 input size of ResNet50 to an output channel of 512
2-4-3) adjusting the Resblock body4 input size of CSPDarknet to make the output channel 1024;
2-5) respectively connecting the three feature maps into modules corresponding to the neg parts of the Yolox to obtain an improved Yolox neural network model;
3) Inputting the large data set into the improved Yolox neural network model obtained in the step 2) to perform pre-training to obtain a pre-training model, inputting the training set obtained in the step 1) into the pre-training model to perform training on the model, inputting the test set into the trained model to perform testing on the model, and obtaining an SMT production line welding point defect target detection model after training and testing are completed;
4) And 3) performing welding spot defect detection on PCBA images produced by the SMT production line by using the model obtained in the step 3) to obtain a detection result, taking data containing welding spot defects in the detection result as newly collected welding spot defect image data, and repeating the steps 1) to 3).
2. The method for rapidly detecting the defects of the SMT production line welding points based on the improved Yolox according to claim 1, wherein in the step 1), the pretreatment is specifically as follows:
1-1) extracting PCBA welding spot defect image data collected by an AOI device from a database of an SMT production line;
1-2) primarily screening the images according to the types and the number of defects contained in the collected images, so that the number of defects of each type is approximately equivalent;
1-3) marking the preliminarily screened PCBA welding spot defect image of the SMT production line by using labelimg, rolabelimg, labelme and vott image marking tools, wherein the information required to be marked comprises position positioning information and defect classification information;
1-4) dividing the marked data into a training set and a testing set according to a preset proportion.
3. A rapid SMT line weld defect detection system based on improved Yolox for performing the method of claim 1 or 2, characterized in that the system comprises:
the data acquisition module is used for acquiring PCBA image data containing welding spot defects stored in the SMT production line AOI equipment, marking the image with information, and dividing the marked data into a training set and a testing set according to a preset proportion;
the model improvement module is used for improving the Yolox neural network model and adjusting the size output by the back bone part after the model improvement so as to match the network part, thereby obtaining an improved Yolox neural network model;
the model training module is used for inputting the training set and the testing set obtained by the data acquisition module into the improved Yolox neural network model for training and testing to obtain an SMT production line welding point defect target detection model;
the defect detection module is used for inputting PCBA images produced by the SMT production line into a welding point defect target detection model of the SMT production line to detect welding point defects and obtain detection results;
the model updating module is used for taking the data containing the welding spot defects in the detection result obtained by the defect detection module as newly collected welding spot defect image data and inputting the newly collected welding spot defect image data into the data acquisition module, and updating the data of the data acquisition module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210450021.8A CN114742811B (en) | 2022-04-27 | 2022-04-27 | SMT production line welding point defect rapid detection method and system based on improved Yolox |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210450021.8A CN114742811B (en) | 2022-04-27 | 2022-04-27 | SMT production line welding point defect rapid detection method and system based on improved Yolox |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114742811A CN114742811A (en) | 2022-07-12 |
| CN114742811B true CN114742811B (en) | 2024-03-29 |
Family
ID=82283635
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210450021.8A Active CN114742811B (en) | 2022-04-27 | 2022-04-27 | SMT production line welding point defect rapid detection method and system based on improved Yolox |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114742811B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020048119A1 (en) * | 2018-09-04 | 2020-03-12 | Boe Technology Group Co., Ltd. | Method and apparatus for training a convolutional neural network to detect defects |
| 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 |
| CN113409250A (en) * | 2021-05-26 | 2021-09-17 | 杭州电子科技大学 | Solder joint detection method based on convolutional neural network |
| WO2022053001A1 (en) * | 2020-09-10 | 2022-03-17 | 上海航天精密机械研究所 | Weld seam internal defect intelligent detection device and method, and medium |
| CN114240821A (en) * | 2021-10-25 | 2022-03-25 | 上海电机学院 | Weld defect detection method based on improved YOLOX |
-
2022
- 2022-04-27 CN CN202210450021.8A patent/CN114742811B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020048119A1 (en) * | 2018-09-04 | 2020-03-12 | Boe Technology Group Co., Ltd. | Method and apparatus for training a convolutional neural network to detect defects |
| 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 |
| WO2022053001A1 (en) * | 2020-09-10 | 2022-03-17 | 上海航天精密机械研究所 | Weld seam internal defect intelligent detection device and method, and medium |
| CN113409250A (en) * | 2021-05-26 | 2021-09-17 | 杭州电子科技大学 | Solder joint detection method based on convolutional neural network |
| CN114240821A (en) * | 2021-10-25 | 2022-03-25 | 上海电机学院 | Weld defect detection method based on improved YOLOX |
Non-Patent Citations (2)
| Title |
|---|
| 基于残差神经网络的转子焊点图像检测方法;邓仕超;张延儒;高兴宇;刘平;黄海兰;杨家欣;;科学技术与工程;20200708(第19期);全文 * |
| 微芯片焊点快速识别与定位方法;黄知超;梁国丽;朱芳来;;半导体技术;20200303(第03期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114742811A (en) | 2022-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106952250B (en) | Metal plate strip surface defect detection method and device based on fast R-CNN network | |
| CN111080622B (en) | Neural network training method, workpiece surface defect classification and detection method and device | |
| CN110455822A (en) | A kind of detection method of pcb board defect | |
| CN109239102A (en) | A kind of flexible circuit board open defect detection method based on CNN | |
| CN107966454A (en) | A kind of end plug defect detecting device and detection method based on FPGA | |
| CN108802041B (en) | Method for rapidly changing small sample set of screen detection | |
| CN112053318A (en) | Two-dimensional PCB defect real-time automatic detection and classification device based on deep learning | |
| CN108596880A (en) | Weld defect feature extraction based on image procossing and welding quality analysis method | |
| CN112651966A (en) | Printed circuit board micro-defect detection method based on ACYOLOV4_ CSP | |
| CN101576956A (en) | On-line character detection method based on machine vision and system thereof | |
| CN108764134A (en) | A kind of automatic positioning of polymorphic type instrument and recognition methods suitable for crusing robot | |
| CN115082401B (en) | SMT production line chip mounter fault prediction method based on improved YOLOX and PNN | |
| CN110349125A (en) | A kind of LED chip open defect detection method and system based on machine vision | |
| CN110186375A (en) | Intelligent high-speed rail white body assemble welding feature detection device and detection method | |
| CN111077223A (en) | Additive manufacturing method with three-dimensional display, online detection and repair functions | |
| CN107977686A (en) | A kind of sorting technique of industrial X-ray image | |
| CN114627082A (en) | Photovoltaic module EL defect detection method based on improved YOLOv5 network | |
| CN112288741A (en) | Product surface defect detection method and system based on semantic segmentation | |
| CN112070712B (en) | Printing defect detection method based on self-encoder network | |
| CN114066848A (en) | FPCA appearance defect visual inspection system | |
| CN116258682A (en) | PCB solder paste defect detection method based on PSPNet and improved YOLOv7 | |
| CN114742811B (en) | SMT production line welding point defect rapid detection method and system based on improved Yolox | |
| CN112907562A (en) | MobileNet-based SMT defect classification algorithm | |
| CN108765391A (en) | A kind of plate glass foreign matter image analysis methods based on deep learning | |
| CN117114420A (en) | Image recognition-based industrial and trade safety accident risk management and control system and method |
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 |