CN113333329A - Cigarette defect detection system based on deep learning - Google Patents
Cigarette defect detection system based on deep learning Download PDFInfo
- Publication number
- CN113333329A CN113333329A CN202110293538.6A CN202110293538A CN113333329A CN 113333329 A CN113333329 A CN 113333329A CN 202110293538 A CN202110293538 A CN 202110293538A CN 113333329 A CN113333329 A CN 113333329A
- Authority
- CN
- China
- Prior art keywords
- cigarette
- image
- detection
- deep learning
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 64
- 235000019504 cigarettes Nutrition 0.000 title claims abstract description 62
- 238000013135 deep learning Methods 0.000 title claims abstract description 25
- 230000007547 defect Effects 0.000 title claims abstract description 17
- 239000011159 matrix material Substances 0.000 claims abstract description 13
- 239000013307 optical fiber Substances 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims description 34
- 238000013527 convolutional neural network Methods 0.000 claims description 17
- 238000000034 method Methods 0.000 claims description 10
- 238000012795 verification Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 239000007787 solid Substances 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000011176 pooling Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 241000208125 Nicotiana Species 0.000 description 3
- 235000002637 Nicotiana tabacum Nutrition 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 239000012535 impurity Substances 0.000 description 3
- 238000013526 transfer learning Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/34—Sorting according to other particular properties
- B07C5/342—Sorting according to other particular properties according to optical properties, e.g. colour
- B07C5/3422—Sorting according to other particular properties according to optical properties, e.g. colour using video scanning devices, e.g. TV-cameras
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/02—Measures preceding sorting, e.g. arranging articles in a stream orientating
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/36—Sorting apparatus characterised by the means used for distribution
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
A cigarette defect detection system based on deep learning comprises a computer, a light source, an optical fiber sensor, a camera, an industrial personal computer and an alarm removing device, wherein a control panel controls the light source to flash and simultaneously send a photographing signal to the camera after judging that a cigarette is in place, the camera collects a cigarette image and transmits the image to the computer, the cigarette image is detected through deep learning detection software installed in the computer, and the image is converted into digital information in a matrix form and is put into a deep learning network for detection and judgment; the detection result comprises the number of cigarettes and whether the cigarettes are empty or not, the detection result is transmitted to the industrial personal computer in a serial port mode, and the industrial personal computer drives the alarm removing device to perform corresponding cigarette removing actions.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to the field of tobacco processing detection, in particular to a cigarette defect detection system based on deep learning.
[ background of the invention ]
The tobacco industry has higher requirements on product defect detection, and mainly comprises cigarette missing detection and missing detection, wherein the missing detection detects whether the tobacco of each cigarette is filled with the whole roll paper and whether the end part of the roll paper is empty; the missing cigarette detection detects whether the number of cigarettes in each box meets a preset number; once detection is missed or a large number of false detections occur, great influence is generated on the production efficiency and the brand reputation of the product; in some dangerous working environments which are not suitable for manual operation or occasions which are difficult for manual vision to meet the requirements, machine vision is commonly used to replace the manual vision; meanwhile, in the process of mass industrial production, the efficiency of checking the product quality by using manual vision is low and the precision is not high; in addition, the complex imaging environment of the visual product makes the traditional image detection product difficult to meet the detection requirements, and the setting of various detection frames and fixed thresholds also makes the traditional image product complex to use and poor in precision.
In the traditional cigarette detection, a plurality of small ROI detection frames are marked on an image, preprocessing and detection are carried out in a fixed area, and when a cigarette shakes up and down, left and right and back and forth, the detection is very inaccurate; moreover, because the traditional image processing detection method mainly sets a threshold value for detection through the white pixel ratio and the mean value and the variance of the pixels after the binary segmentation, once the position is greatly changed, the situation that only half of cigarette ends are included may occur, and then a great difference is generated between the position and the preset pixel threshold value, thereby causing false detection; common false detection comprises that the ROI position is unchanged, the cigarette position changes, and the number of white pixels in a detection frame is too large, so that the false detection is carried out to form an empty head; if impurities exist and the color of the impurities is bright, the impurities are mistakenly judged as empty heads when being divided by a fixed threshold value; when the degree and the angle of the cigarette are different, the cigarette ends are different in area due to the fact that the degree and the angle are different, so that the mean value and the variance of the pixels after binarization can change greatly and easily exceed a fixed threshold value, and the cigarette is judged to be empty.
Therefore, the problem to be solved in the field is urgently needed to provide a cigarette defect detection system with high precision and high efficiency.
[ summary of the invention ]
Aiming at the problems, the invention provides a cigarette defect detection system based on deep learning, which comprises a computer, a light source, an optical fiber sensor, a camera, an industrial personal computer and an alarm removing device, wherein a control panel controls the light source to flash and simultaneously sends a photographing signal to the camera after judging that a cigarette is in place, the image of the cigarette is collected by the camera and is transmitted to the computer, the image of the cigarette is detected by deep learning detection software installed in the computer, and the image is converted into digital information in a matrix form and is put into a deep learning network for detection and judgment; the detection result comprises the number of cigarettes and whether the cigarettes are empty, the detection result is transmitted to the industrial personal computer through serial ports and other forms, and the industrial personal computer drives the alarm removing device to execute corresponding cigarette removing actions.
Furthermore, the light source adopts the LED light source to adopt stroboscopic and diffuse reflection mode to carry out even illumination to the cigarette.
Further, the industrial personal computer carries an Intel Celeron J1900 quad-core 2.0GHz processor, an 8G memory and a 1T solid state disk, and adopts stm32 as a control core.
A cigarette defect detection system based on deep learning comprises the following training and detection processes:
the method comprises the following steps: collecting training images, and marking the images as positive samples and negative samples;
step two: performing preprocessing on images participating in training, including changing sizes and enhancing data;
step three: dividing the collected positive and negative sample images into a training set and a verification set;
step four: setting a hyper-parameter of a network, training an image by adopting a convolutional neural network, adjusting the hyper-parameter to train the image for multiple times, and selecting a model with the highest precision from models obtained by multiple times of training through verification set observation precision, recall rate and the like;
step five: and carrying out the same pretreatment on the image to be detected, and then carrying out a trained convolutional neural network model to finally obtain a classification result.
Compared with the prior art, the invention has the following beneficial effects:
1. according to the invention, iterative training is carried out through a large amount of image data, so that a convolutional neural network is formed, not only can the pixel proportion information of the picture be extracted, but also deep texture information and edge information can be obtained, and the detection accuracy is improved through multi-information comprehensive analysis.
2. The invention adopts the transfer learning method to train, so that the training speed is faster, the precision is higher, the training time is saved, and the requirement of industrial detection real-time training is met.
3. The deep learning detection technology adopted by the invention is the most advanced detection technology in the machine vision industry at present, is different from the traditional detection method in that a plurality of detection frames are required to be arranged, the system can realize the full-information detection of all images without drawing the detection frames, has high speed and high precision, and can be matched with various existing medium-speed and high-speed cigarette machines.
[ description of the drawings ]
Fig. 1 is a schematic diagram of the hardware architecture of the present invention.
Fig. 2 is a training and detection flow diagram of the present invention.
FIG. 3 is a schematic diagram of a convolutional neural network of the present invention.
[ detailed description ] embodiments
The directional terms of the present invention, such as "up", "down", "front", "back", "left", "right", "inner", "outer", "side", etc., are only directions in the drawings, and are only used to explain and illustrate the present invention, but not to limit the scope of the present invention.
The invention relates to a cigarette defect detection system based on deep learning, which is based on a convolutional neural network model and mainly comprises three parts, wherein an industrial camera is used for acquiring a front image of a cigarette, then cigarette image information is used as a matrix to be input, the deep learning network model is trained through a large number of images and judges whether the cigarette is missing or empty, and finally the empty cigarette is removed according to a judgment result.
The convolutional neural network is the core of the system, extracts image information, not only extracts surface information such as image brightness and edges, but also utilizes each pixel information of the image; the convolutional neural network comprises a forward propagation process and a backward propagation process, namely a neural network model closest to real judgment is obtained by continuously updating the weight and the bias of the convolutional neural network through iteration of extracting a large amount of image information, and judgment close to a real value is obtained after an image to be detected passes through a trained network.
Referring to fig. 1, hardware of the cigarette defect detection system based on deep learning of the invention comprises a computer, a light source, an optical fiber sensor, a camera, an industrial personal computer and an alarm removing device, wherein the light source adopts an LED light source and a stroboscopic working mode, and can blink when triggered, so that energy consumption is reduced, the cigarettes are uniformly illuminated in a diffuse reflection mode, corner shadows and random light reflection problems are thoroughly eliminated, the occurrence of detection dead angles is avoided, and an effective image basis is established for comprehensive detection; the industrial personal computer carries an Intel Celeron J1900 quad-core 2.0GHz processor, an 8G memory and a 1T solid state disk, and the control board adopts stm32 as a control core.
The working principle is as follows:
the processor in the control panel judges that the cigarette is in place according to the optical fiber sensor, controls the light source to flash and simultaneously sends a photographing signal to the camera, acquires a cigarette image through the camera, transmits the image to the computer, detects the cigarette image through depth learning detection software installed in the computer, converts the image into digital information in a matrix form, and puts the digital information into a depth learning network trained in advance for detection and judgment; the detection result comprises information such as the number (whether the cigarettes lack) of the cigarettes and whether the cigarettes are empty, so that an accurate automatic identification function is realized, the result is finally transmitted to the industrial personal computer through serial ports and the like, and the industrial personal computer drives the alarm removing device to perform corresponding cigarette removing actions.
Referring to fig. 2, the training and detection process of the cigarette defect detection system based on deep learning of the present invention is as follows:
the method comprises the following steps: firstly, acquiring a training image, and marking the image as a positive sample and a negative sample, wherein the positive sample is 3000, and the negative sample is 500;
step two: performing preprocessing on images participating in training, including changing sizes and enhancing data;
step three: dividing the collected positive and negative sample images into a training set and a verification set, wherein the training set is used for training, and the verification set is used for verifying the precision;
step four: setting hyper-parameters (including learning rate, training round number and the like) of a network, training images by adopting a convolutional neural network, adjusting the hyper-parameters to train the images for multiple times, and selecting a model with the highest precision from models obtained by multiple times of training through verification set observation precision, recall rate and the like;
step five: and carrying out the same pretreatment on the image to be detected, and then carrying out a trained convolutional neural network model to finally obtain a classification result.
The algorithm core of the cigarette defect detection system based on deep learning is a convolutional neural network, which comprises an input layer, a convolutional layer, an activation function, a pooling layer and a full-connection layer, wherein the convolutional layer, the pooling layer and the full-connection layer jointly form a hidden layer;
the input layer is used for processing multi-dimensional data including color image information, is input to the whole neural network, and represents a pixel matrix of a picture in the convolutional neural network for processing the image; as shown in fig. 3, the leftmost three-dimensional matrix can represent a picture, wherein the length and width of the three-dimensional matrix represent the size of the image, the depth of the three-dimensional matrix represents the color channel of the image, the depth of the black and white picture is 1, and the depth of the image is 3 in the RGB color mode; starting from an input layer, the convolution neural network converts the three-dimensional matrix of the previous layer into the three-dimensional matrix of the next layer through different neural network structures until the last full-connection layer.
The convolutional layer is the most important part of a convolutional neural network, the input of each node in the convolutional layer is only a small block of the neural network of the previous layer, the common size of the small block is 3 x 3, and the convolutional layer analyzes each small block in the neural network more deeply so as to obtain features with higher abstraction degree.
The pooling layer can reduce the size of the matrix without changing the depth of the three-dimensional matrix, a picture with higher resolution is converted into a picture with lower resolution by pooling operation, and the number of nodes in the final full-connection layer can be further reduced through the pooling layer, so that the aim of reducing parameters in the whole neural network is fulfilled.
The full-connection layer is used for giving a classification result, after image information is processed by the convolution layer and the pooling layer for several times, the image information is abstracted into a characteristic with higher information content, and then a classification task is completed through full connection.
The cigarette loose end detection method is the same as the traditional detection method in that a fixed ROI is also set for shearing, but a large amount of training data can generate different position conditions, a deep learning network model learns more image information, the deep learning network model does not detect the cigarette loose end through single threshold comparison, but carries out iterative training through a large amount of image data, so that a convolutional neural network is formed, and the convolutional neural network comprises a plurality of weights and bias values; the trained images comprise the changes of the distance and the direction of one or more than one image, deep learning training can extract pixel proportion information of the images and can also obtain deep texture information and edge information, and the detection accuracy is improved through multi-information comprehensive analysis; in addition, in order to improve the deep learning training speed, the invention adopts a transfer learning method for training, the transfer learning allows processing scenes by borrowing the label data of some existing related tasks, and stores the knowledge obtained when the related tasks are solved, so that the training speed is higher, the training time is saved, and the requirement of industrial detection real-time training is met.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (4)
1. A cigarette defect detection system based on deep learning is characterized by comprising a computer, a light source, an optical fiber sensor, a camera, an industrial personal computer and an alarm removing device, wherein a control panel controls the light source to flash and simultaneously sends a photographing signal to the camera after judging that a cigarette is in place, the image of the cigarette is collected through the camera and is transmitted to the computer, the cigarette image is detected through deep learning detection software installed in the computer, and the image is converted into digital information in a matrix form and is input into a deep learning network for detection and judgment; the detection result comprises the number of cigarettes and whether the cigarettes are empty, the detection result is transmitted to the industrial personal computer in a serial port mode, and the industrial personal computer drives the alarm removing device to execute corresponding cigarette removing actions.
2. The cigarette defect detection system based on deep learning of claim 1, wherein the light source is an LED light source, and the cigarettes are uniformly illuminated in a stroboscopic and diffuse reflection manner.
3. The cigarette defect detection system based on deep learning of claim 1, characterized in that the industrial personal computer carries an IntelCeleron J1900 quad-core 2.0GHz processor, an 8G memory, and a 1T solid state disk, and adopts stm32 as a control core.
4. The utility model provides a cigarette defect detecting system based on degree of depth study which characterized in that, cigarette defect detecting system's based on degree of depth study training and detection flow includes following step:
the method comprises the following steps: collecting training images, and marking the images as positive samples and negative samples;
step two: performing preprocessing on images participating in training, including changing sizes and enhancing data;
step three: dividing the collected positive and negative sample images into a training set and a verification set;
step four: setting a hyper-parameter of a network, training an image by adopting a convolutional neural network, adjusting the hyper-parameter to train the image for multiple times, and selecting a model with the highest precision from models obtained by multiple times of training through verification set observation precision, recall rate and the like;
step five: and carrying out the same pretreatment on the image to be detected, and then carrying out a trained convolutional neural network model to finally obtain a classification result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110293538.6A CN113333329A (en) | 2021-08-04 | 2021-08-04 | Cigarette defect detection system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110293538.6A CN113333329A (en) | 2021-08-04 | 2021-08-04 | Cigarette defect detection system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113333329A true CN113333329A (en) | 2021-09-03 |
Family
ID=77467712
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110293538.6A Pending CN113333329A (en) | 2021-08-04 | 2021-08-04 | Cigarette defect detection system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113333329A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114155676A (en) * | 2021-11-29 | 2022-03-08 | 山东中烟工业有限责任公司 | Logistics system damaged wood pallet detection alarm system and working method thereof |
| CN114577816A (en) * | 2022-01-18 | 2022-06-03 | 广州超音速自动化科技股份有限公司 | Hydrogen fuel bipolar plate detection method |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102499471A (en) * | 2011-10-25 | 2012-06-20 | 中国电子科技集团公司第四十一研究所 | Cigarette loose-end detection method based on machine vision technology |
| CN102697176A (en) * | 2012-06-23 | 2012-10-03 | 常德烟草机械有限责任公司 | Cigarette quality on-line detection device and detection method based on machine vision |
| CN103385539A (en) * | 2013-08-02 | 2013-11-13 | 南京文采科技有限责任公司 | Single cigarette empty head detection method based on machine vision and special equipment |
| CN109829895A (en) * | 2019-01-09 | 2019-05-31 | 武汉精立电子技术有限公司 | A kind of AOI defect inspection method based on GAN |
| CN111189837A (en) * | 2020-01-08 | 2020-05-22 | 征图新视(江苏)科技股份有限公司 | Cigarette appearance online detection method and device |
| CN111260609A (en) * | 2020-01-08 | 2020-06-09 | 征图新视(江苏)科技股份有限公司 | Cigarette appearance defect detection method based on deep learning |
| CN112184655A (en) * | 2020-09-24 | 2021-01-05 | 东北大学 | Wide and thick plate contour detection method based on convolutional neural network |
| CN112268910A (en) * | 2020-11-13 | 2021-01-26 | 上海奇信机电设备有限公司 | Visual detection system and method for cigarette appearance defects |
| CN112432952A (en) * | 2020-11-20 | 2021-03-02 | 中国电子科技集团公司第四十一研究所 | Cigarette loose end detection method based on machine vision technology |
| CN113034483A (en) * | 2021-04-07 | 2021-06-25 | 昆明理工大学 | Cigarette defect detection method based on deep migration learning |
| CN113066047A (en) * | 2021-02-23 | 2021-07-02 | 青岛科技大学 | Method for detecting impurity defects of tire X-ray image |
-
2021
- 2021-08-04 CN CN202110293538.6A patent/CN113333329A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102499471A (en) * | 2011-10-25 | 2012-06-20 | 中国电子科技集团公司第四十一研究所 | Cigarette loose-end detection method based on machine vision technology |
| CN102697176A (en) * | 2012-06-23 | 2012-10-03 | 常德烟草机械有限责任公司 | Cigarette quality on-line detection device and detection method based on machine vision |
| CN103385539A (en) * | 2013-08-02 | 2013-11-13 | 南京文采科技有限责任公司 | Single cigarette empty head detection method based on machine vision and special equipment |
| CN109829895A (en) * | 2019-01-09 | 2019-05-31 | 武汉精立电子技术有限公司 | A kind of AOI defect inspection method based on GAN |
| CN111189837A (en) * | 2020-01-08 | 2020-05-22 | 征图新视(江苏)科技股份有限公司 | Cigarette appearance online detection method and device |
| CN111260609A (en) * | 2020-01-08 | 2020-06-09 | 征图新视(江苏)科技股份有限公司 | Cigarette appearance defect detection method based on deep learning |
| CN112184655A (en) * | 2020-09-24 | 2021-01-05 | 东北大学 | Wide and thick plate contour detection method based on convolutional neural network |
| CN112268910A (en) * | 2020-11-13 | 2021-01-26 | 上海奇信机电设备有限公司 | Visual detection system and method for cigarette appearance defects |
| CN112432952A (en) * | 2020-11-20 | 2021-03-02 | 中国电子科技集团公司第四十一研究所 | Cigarette loose end detection method based on machine vision technology |
| CN113066047A (en) * | 2021-02-23 | 2021-07-02 | 青岛科技大学 | Method for detecting impurity defects of tire X-ray image |
| CN113034483A (en) * | 2021-04-07 | 2021-06-25 | 昆明理工大学 | Cigarette defect detection method based on deep migration learning |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114155676A (en) * | 2021-11-29 | 2022-03-08 | 山东中烟工业有限责任公司 | Logistics system damaged wood pallet detection alarm system and working method thereof |
| CN114577816A (en) * | 2022-01-18 | 2022-06-03 | 广州超音速自动化科技股份有限公司 | Hydrogen fuel bipolar plate detection method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109239102B (en) | CNN-based flexible circuit board appearance defect detection method | |
| CN111582294B (en) | Method for constructing convolutional neural network model for surface defect detection and application thereof | |
| CN104574353B (en) | The surface defect decision method of view-based access control model conspicuousness | |
| CN111815564B (en) | Method and device for detecting silk ingots and silk ingot sorting system | |
| CN108985170A (en) | Transmission line of electricity hanger recognition methods based on Three image difference and deep learning | |
| CN101726498B (en) | Intelligent detector and method of copper strip surface quality on basis of vision bionics | |
| CN113333329A (en) | Cigarette defect detection system based on deep learning | |
| CN106248680A (en) | A kind of engine commutator quality detecting system based on machine vision and detection method | |
| CN111239142A (en) | Paste appearance defect detection device and method | |
| CN112634237A (en) | Long bamboo strip surface defect detection method and system based on YOLOv3 improved network | |
| Zhang et al. | A novel image detection method for internal cracks in corn seeds in an industrial inspection line | |
| Zhang et al. | Fully automatic system for fish biomass estimation based on deep neural network | |
| CN114577805A (en) | MiniLED backlight panel defect detection method and device | |
| CN112819844A (en) | Image edge detection method and device | |
| CN115311250A (en) | Cigarette appearance detection method based on deep learning and space-time feature fusion | |
| TWI749714B (en) | Method for defect detection, method for defect classification and system thereof | |
| CN117152158B (en) | Textile defect detection method and system based on artificial intelligence | |
| CN116563276A (en) | Chemical fiber filament online defect detection method and detection system | |
| CN117330582A (en) | Polymer PE film surface crystal point detecting system | |
| CN117079125A (en) | Kiwi fruit pollination flower identification method based on improved YOLOv5 | |
| TW202326114A (en) | Fault or defect detection system for long-code zipper capable of meeting the requirement of continuous detection of the production assembly line having transmission speed at 300 m/min or lower, and has high detection precision | |
| CN111340802A (en) | Intelligent cloth inspecting machine capable of automatically inspecting cloth by adopting artificial intelligence technology | |
| CN114235810A (en) | Automatic cloth flaw detection method based on area array industrial camera | |
| JP7089718B1 (en) | Peeled log sorting system, information processing device, peeled log sorting method and program | |
| CN114898148B (en) | Egg offset detection method and system based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210903 |
|
| RJ01 | Rejection of invention patent application after publication |