CN110186929A - A kind of real-time product defect localization method - Google Patents
A kind of real-time product defect localization method Download PDFInfo
- Publication number
- CN110186929A CN110186929A CN201910333188.4A CN201910333188A CN110186929A CN 110186929 A CN110186929 A CN 110186929A CN 201910333188 A CN201910333188 A CN 201910333188A CN 110186929 A CN110186929 A CN 110186929A
- Authority
- CN
- China
- Prior art keywords
- flaw
- image
- gray
- real
- pixel
- 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
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)
Abstract
The invention proposes a kind of real-time product defect localization methods, comprising the following steps: Step 1: image preprocessing, including image gray processing and image binaryzation；Step 2: selected processing region；Step 3: flaw slightly defines；Step 4: flaw essence defines.The present invention realize product defect it is real-time, quick and precisely detect, operand is small, and detection accuracy is high, can also the requirement according to client to product quality set, to meet different demands required for different product positions product defect.
Description
Technical field
The present invention relates to a kind of real-time product defect localization methods, belong to product quality detection technique field.
Background technique
With the progress of production and technology, the requirement of product quality also further increased, product surface quality for
The influence of final product quality has been to be concerned by more and more people.The flaws such as optical cable, wafer, chip compare final product quality influence
Big product, in production, to the detection of its product defect with regard to particularly significant.
Currently, the domestic method for product defect positioning also rests on original artificial detection mostly.Artificial detection exists
The required basic demand of industry is all not achieved in the epoch of current production flow line production, accuracy and speed.
At this stage to the research of product defect positioning, mainly by the feature extraction to flaw, then image pixel is carried out
Traversal, matches the region for meeting unwanted visual characteristic.Such method scope of application is not wide, algorithm too complex cannot achieve real-time inspection
It surveys.
Therefore there is an urgent need to explore a kind of objective, effective, flexible, high speed, reliable, applied widely, real-time for many enterprises
Product defect localization method replace traditional artificial detection.
Summary of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of real-time product defect localization method, with
Realize product defect it is real-time, quick and precisely detect, this method operand is small, and detection accuracy is high, can also be according to client to production
The requirement of quality is set, and then can examine various sizes of flaw, is positioned with meeting different product to product defect
Required different demands.
The present invention provides a kind of real-time product defect localization method, comprising the following steps:
Step 1: image preprocessing, including image gray processing and image binaryzation；
Step 2: selected processing region；
Step 3: flaw slightly defines,
Flaw is slightly defined to coefficient matrix and image makees planar convolution, obtains the shade of gray value of each pixel；
Grads threshold is set, the shade of gray value of each pixel is compared with Grads threshold, when shade of gray value is big
When Grads threshold, assert that the pixel is flaw marginal point；
The region that multiple flaw marginal points surround is abnormal area；
Step 4: flaw essence defines, flaw minimum dimension is set, the whether minimum greater than flaw of abnormal area is compared in judgement
Size, if so, then determining abnormal area for flaw.
As further technical solution of the present invention, described image gray processing is specially to substitute into pixel each in image
Following formula:
Gray=R*0.3+G*0.59+B*0.11,
Wherein, the unprocessed form of image is RGB565, and R, G, B are respectively three component values of each pixel in formula,
Gray is the gray value of each pixel.
Further, described image binaryzation specifically: a gray threshold μ is taken again to the image after gray processing, it will
The gray value indirect assignment of pixel of the Gray value greater than μ is 255, the gray value whole assignment 0 of other pixels.
Further, the selected processing region is specially the relative position according to product and camera, right in the picture
It is chosen product region.
Further, it includes that lateral flaw slightly defines coefficient matrix and longitudinal flaw is thick that the flaw, which slightly defines coefficient matrix,
Define coefficient matrix.
Further, in the step 3 shade of gray value specific calculating process are as follows:
Wherein: A represents original image, GxAnd GyIt respectively represents and slightly defines coefficient matrix through transverse direction and longitudinal direction flaw and detected
Complete image gradient value, G are shade of gray value.
Further, the flaw essence definition specifically:
Flaw size threshold value m is set,
According to the proportionate relationship of shooting image and actual object, m is converted into the pixel number n on image, is judged different
Whether the pixel number in normal region is greater than n；
If so, then determining the abnormal area for flaw；
If not, determining the abnormal area for pseudo- flaw.
The invention adopts the above technical scheme compared with prior art, has following technical effect that
1, the present invention first pre-processes original image, and the information after making it contain only binaryzation eliminates major part
Useless information, the operand of image procossing after greatly reducing, improves arithmetic speed, while also keeping original flaw
Locating effect is constant.
2, method proposed by the present invention does not need all pixels of traversal image, as long as locating in the picture to product surface
Region handled, then count the number that abnormal area includes pixel, then be compared to each other with the threshold value provided, so that it may
To determine when product surface whether there is flaw in preceding camera scanning area.Cast out and has found product surface in image procossing
Process, and then shorten the time of image procossing.
3, the whole processing step of the present invention, the fortune other than having used convolution at flaw slightly definition, in other steps
It calculates and there was only addition, be significantly reduced the calculation amount for the treatment of process, while also ensuring detection mass conservation, be fully able to matching work
The speed of production assembly line in industry.
4, the present invention is provided with threshold window, can the requirement according to client to product quality set, and then can be with
Various sizes of flaw is examined, to meet different demands required for different product positions product defect.
Detailed description of the invention
Fig. 1 is a kind of overall flow figure of real-time product defect localization method of the present invention.
Image of the original image of optical cable image information collecting after gray processing processing in Fig. 2 embodiment of the present invention.
Fig. 3 is to carry out the schematic diagram after binary conversion treatment in the embodiment of the present invention to grayscale image.
Fig. 4 is the schematic diagram selected in processing region process in the embodiment of the present invention.
Fig. 5 is the schematic diagram for giving up the image information outside selection area in the embodiment of the present invention.
Fig. 6 is to carry out the schematic diagram after flaw slightly defines to defined area in the embodiment of the present invention.
Fig. 7 is by finding out the schematic diagram of flaw after carrying out the definition of flaw essence to defined area in the embodiment of the present invention.
Specific embodiment
As shown in Figure 1, a kind of real-time product defect localization method disclosed by the invention, mainly includes image preprocessing,
Selected processing region, flaw slightly defines and flaw essence defines this four steps.
Step 1, image preprocessing mainly include image gray processing and image binaryzation.
Image gray processing is mainly to convert grayscale image for the original color image that camera acquisition comes.
The format of original image is RGB565, and by R, tri- values of G, B determine each pixel in image.By each picture
The R of vegetarian refreshments, G, B, which substitutes into formula (1), can calculate the gray value Gray of the pixel:
Gray=R*0.3+G*0.59+B*0.11 (1)
Each pixel of original image is inserted into its gray value Gray, the image after gray processing can be obtained.
Image binaryzation is mainly further to compress the amount of image information after gray processing, with the fortune of image procossing after reduction
Calculation amount.
The specific implementation process of binaryzation is to take a threshold value μ again to the image after gray processing, and the value of μ can be according to reality
Border situation carries out adjustment appropriate, and the gray value indirect assignment of the pixel by Gray value greater than μ is 255, other pixels
Gray value whole assignment 0.
Step 2, the purpose of selected processing region are mainly to utilize the special nature of product production line in industrial production, will be adopted
The image collected carries out the selection of product region, only handles the region, reduce to product with the fortune of exterior domain
It calculates, reduces the operand in image processing process, improve the speed of product defect positioning.
In general industrial products production process, product always with the mode of manufacture of assembly line, is acquired in camera and is produced
When product information, the relative position of product and camera will not have greatly changed.Therefore, product is in the collected original of camera
A relatively fixed position is always in beginning image.The present invention can select the region where product by this characteristic
As processing region, processing region can be adjusted accordingly according to specific industrial requirement.
Therefore before the present invention carries out flaw positioning, product can directly be neglected with the image information of exterior domain, only
Flaw positioning is carried out to fixed area locating for product in image.
Step 3, the flaw slightly define part and mainly slightly define coefficient matrix using flaw provided by the invention:
Flaw slightly defines the matrix that coefficient matrix includes two groups of 3*3, and respectively lateral flaw slightly defines coefficient matrix and indulges
Coefficient matrix is slightly defined to flaw, two matrixes and image are made into planar convolution, can obtain the luminance difference of transverse direction and longitudinal direction respectively
Divide approximation.
If representing original image, G with AxAnd GyIt respectively represents and slightly defines coefficient matrix through transverse direction and longitudinal direction flaw and detected
Complete image gradient value, G are shade of gray, formula such as (2) (3):
The transverse direction and longitudinal direction gray value of each pixel of image is combined by formula (4), to calculate the big of the gradient
It is small；
Gradient G is greater than given threshold, then it is assumed that the point is flaw marginal point.The region that multiple flaw marginal points surround is different
Normal region.
Step 4, the definition of flaw essence mainly include that the setting of flaw threshold value and threshold value comparison determine two parts:
Assuming that the minimum dimension for the flaw that we need to find is m, then m is exactly to carry out smart definition to flaw in the present invention
Threshold value.By shooting the proportionate relationship of image and actual object size, pixel number n m being converted on image.The present invention
Automatically retrieval is greater than the abnormal area of n pixel, and if there is such abnormal area, the present invention decides that product surface is deposited
In flaw.For being less than the abnormal area of n pixel on image, the present invention is regarded as pseudo- flaw, is not considered.
The number for meeting the flaw abnormal area of condition contained in image is recorded, q is denoted as.
In conclusion the data q being disposed by above-mentioned steps, is exactly the quantity of product defect.
Below by taking the positioning of optical cable flaw as an example, technical solution of the present invention is described in further detail in conjunction with attached drawing.
The present embodiment proposes on the assembly line of industrial fiber optic cable manufacture, and the transmission speed of optical cable is generally 3 to 5m/s.For
Guarantee that the flaw of optical cable in this speed is still detected, the acquisition rate of image needs to reach 500fps.
Each part of optical cable is likely to occur flaw, and due to taking the method for choosing processing region that can lose one
The information at optic cable edge in a certain image.
Therefore using 4 high-speed cameras, frame is in 4 directions up and down of optical cable simultaneously, and every place's image is in the present invention
It only intercepts above and below at optical cable central axesThe region of times diameter.
Image handled by every video camera in this way is 1/4 of information content contained by optical cable.And it is handled required for 1 second
Amount of images be 2000 frames, every frame image time to be dealt be 0.5ms.
Therefore in order to keep higher computation rate, it is necessary to use and simplify effective detection and recognizer.
Fig. 2 is image of the original image of optical cable image information collecting in the present invention after gray processing processing, wherein
White rectangle region is cable region, is non-optical cable outside white rectangle region, and gray area is to identify in white rectangle region
Flaw.Although also containing on image as can be seen that have passed through the processing of gray processing and largely positioning unrelated letter with flaw
Breath.Therefore the processing for needing to carry out binaryzation, reduces the operand of image procossing.
Fig. 3 is that the present invention carries out the schematic diagram after binary conversion treatment to grayscale image, since flaw and product surface are to light
Reflectivity be very different, therefore there is obviously color difference in the image of flaw and product surface, and pass through two-value
After change, flaw and product surface are by more obvious separated, after progress that can be easier Defect Detection.In Fig. 2
In, the rectangle white area in the middle part of image is cable region, and black irregular shape region thereon is flaw.
Fig. 4 is the schematic diagram that processing region is selected in the present invention, wherein the region that two dark parallel lines surround is final
Determine selected processing region, remaining region is all inactive area.The one of product and flaw is contained only in selected processing region
A little information all give up the information except this, reduce the information content that must be analyzed substantially.In Fig. 4, also can
The reduction of calculation amount at intuitive performance.
Fig. 5 is the schematic diagram for giving up the image information outside selection area, and all areas mark outside defined area is black, at this
Image information later only carries out the positioning of flaw without acquisition in selection area.
Fig. 6 is that the schematic diagram after flaw slightly defines is carried out to selection area, passes through one of flaw coarse positioning step, flaw portion
Divide and be indicated in the form of black line enclosing region, certain region has been translated into the presence or absence of the judgement of flaw and has judged the region
Whether the size of middle black line enclosing region is in defined threshold value.
It is assumed that our the flaw size requirements to be identified are it can be concluded that, to assert the flaw by calculating greater than 0.4mm*0.4mm
The pixel that defect is included will be more than 1600, and the pixel of the flaw in Fig. 6 at the encirclement of black surround region is more than 1600, therefore
It can be judged as flaw.
Fig. 7 is the schematic diagram to found out flaw after the progress flaw essence definition of defined area, it can be seen that is less than threshold zone
The part in domain has been filtered out, and leaves behind the flaw for meeting testing conditions.
The flaw location algorithm of mainstream is all based on the flaw matching of active at this stage, these algorithms are mostly all relatively more multiple
It is miscellaneous, such as the flaw location algorithm based on uncertainty principle, it needs to carry out calculating judgement in terms of time domain and frequency domain two.This just leads
It has caused these algorithms to take long time, cannot achieve real-time detection product defect.And in contrast, the present invention is with flaw and product table
Reflectivity in face of light is not all point of penetration, and the flaw of camera acquisition and the image of product surface can exist clearly
Color difference, therefore the threshold value of the area for calculating color difference region and setting is only needed to be compared, can judge whether flaw is deposited
Compared with traditional flaw location algorithm, calculation amount substantially reduces, therefore the real-time detection of flaw may be implemented.
In the production process of industrial products, if product defect positioning is unable to real-time perfoming detection, it is necessary to spend volume
The outer time detects product quality, causes entire product manufacturing process elongated, produces if realized using the method for the present invention
The real-time detection of product flaw, so that it may directly the quality of product be detected during production, can substantially shorten production
The production procedure of product improves production efficiency.
The above, the only specific embodiment in the present invention, but scope of protection of the present invention is not limited thereto, appoints
What is familiar with the people of the technology within the technical scope disclosed by the invention, it will be appreciated that expects transforms or replaces, and should all cover
Within scope of the invention, therefore, the scope of protection of the invention shall be subject to the scope of protection specified in the patent claim.
Claims (7)
1. a kind of real-time product defect localization method, it is characterised in that the following steps are included:
Step 1: image preprocessing, including image gray processing and image binaryzation；
Step 2: selected processing region；
Step 3: flaw slightly defines,
Flaw is slightly defined to coefficient matrix and image makees planar convolution, obtains the shade of gray value of each pixel；
Grads threshold is set, the shade of gray value of each pixel is compared with Grads threshold, when shade of gray value is greater than ladder
When spending threshold value, assert that the pixel is flaw marginal point；
The region that multiple flaw marginal points surround is abnormal area；
Step 4: flaw essence defines,
Set flaw minimum dimension, judgement compare abnormal area whether be greater than flaw minimum dimension, if so, then determining exceptions area
Domain is flaw.
2. a kind of real-time product defect localization method according to claim 1, which is characterized in that in the step 1
Image gray processing is specially that pixel each in image is substituted into following formula:
Gray=R*0.3+G*0.59+B*0.11,
Wherein, the unprocessed form of image is RGB565, and R, G, B are respectively three component values of each pixel in formula, and Gray is
The gray value of each pixel.
3. a kind of real-time product defect localization method according to claim 2, which is characterized in that in the step 1
Image binaryzation specifically:
One gray threshold μ is taken again to the image after gray processing, the gray value indirect assignment of the pixel by Gray value greater than μ
It is 255, the gray value whole assignment 0 of other pixels.
4. a kind of real-time product defect localization method according to claim 1, which is characterized in that the selected treatment region
Domain is specially the relative position according to product and camera, is chosen in the picture to product region.
5. a kind of real-time product defect localization method according to claim 1, which is characterized in that the flaw slightly defines
Coefficient matrix slightly defines coefficient matrix including lateral flaw and longitudinal flaw slightly defines coefficient matrix.
6. a kind of real-time product defect localization method according to claim 5, which is characterized in that grey in the step 3
Spend the specific calculating process of gradient value are as follows:
Wherein: A represents original image, GxAnd GyIt respectively represents and slightly defines what coefficient matrix detection finished through transverse direction and longitudinal direction flaw
Image gradient value, G are shade of gray value.
7. a kind of real-time product defect localization method according to claim 1, which is characterized in that the flaw essence definition
Specifically:
Flaw size threshold value m is set,
According to the proportionate relationship of shooting image and actual object, m is converted into the pixel number n on image, judges exceptions area
Whether the pixel number in domain is greater than n；
If so, then determining the abnormal area for flaw；
If not, determining the abnormal area for pseudo- flaw.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910333188.4A CN110186929A (en) | 2019-04-24 | 2019-04-24 | A kind of real-time product defect localization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910333188.4A CN110186929A (en) | 2019-04-24 | 2019-04-24 | A kind of real-time product defect localization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110186929A true CN110186929A (en) | 2019-08-30 |
Family
ID=67714986
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910333188.4A Pending CN110186929A (en) | 2019-04-24 | 2019-04-24 | A kind of real-time product defect localization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110186929A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111933543A (en) * | 2020-08-13 | 2020-11-13 | 扬州扬杰电子科技股份有限公司 | Appearance detection method of axial diode |
| CN114935802A (en) * | 2022-05-13 | 2022-08-23 | 长飞光纤光缆股份有限公司 | Optical cable outer sheath gas polishing assembly, system and method in optical cable production |
| CN116228752A (en) * | 2023-05-06 | 2023-06-06 | 四川广安承平投资开发有限公司 | Surface flaw identification method and processing robot |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102680478A (en) * | 2012-04-25 | 2012-09-19 | 华南农业大学 | Detection method and device of surface defect of mechanical part based on machine vision |
| CN105158268A (en) * | 2015-09-21 | 2015-12-16 | 武汉理工大学 | Intelligent online detection method, system and device for defects of fine-blanked parts |
| CN108037138A (en) * | 2017-12-23 | 2018-05-15 | 陕西科技大学 | A kind of web inspection system and detection method for being used to detect the two-sided defect of paper |
| CN108154510A (en) * | 2018-01-17 | 2018-06-12 | 深圳市亿图视觉自动化技术有限公司 | Method for detecting surface defects of products, device and computer readable storage medium |
| CN109507192A (en) * | 2018-11-02 | 2019-03-22 | 江苏理工学院 | A kind of magnetic core detection method of surface flaw based on machine vision |
| CN109523541A (en) * | 2018-11-23 | 2019-03-26 | 五邑大学 | A kind of metal surface fine defects detection method of view-based access control model |
-
2019
- 2019-04-24 CN CN201910333188.4A patent/CN110186929A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102680478A (en) * | 2012-04-25 | 2012-09-19 | 华南农业大学 | Detection method and device of surface defect of mechanical part based on machine vision |
| CN105158268A (en) * | 2015-09-21 | 2015-12-16 | 武汉理工大学 | Intelligent online detection method, system and device for defects of fine-blanked parts |
| CN108037138A (en) * | 2017-12-23 | 2018-05-15 | 陕西科技大学 | A kind of web inspection system and detection method for being used to detect the two-sided defect of paper |
| CN108154510A (en) * | 2018-01-17 | 2018-06-12 | 深圳市亿图视觉自动化技术有限公司 | Method for detecting surface defects of products, device and computer readable storage medium |
| CN109507192A (en) * | 2018-11-02 | 2019-03-22 | 江苏理工学院 | A kind of magnetic core detection method of surface flaw based on machine vision |
| CN109523541A (en) * | 2018-11-23 | 2019-03-26 | 五邑大学 | A kind of metal surface fine defects detection method of view-based access control model |
Non-Patent Citations (1)
| Title |
|---|
| 周智敏等: "《超宽带地表穿透成像雷达》", 31 March 2013 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111933543A (en) * | 2020-08-13 | 2020-11-13 | 扬州扬杰电子科技股份有限公司 | Appearance detection method of axial diode |
| CN111933543B (en) * | 2020-08-13 | 2024-07-09 | 扬州扬杰电子科技股份有限公司 | Appearance detection method of axial diode |
| CN114935802A (en) * | 2022-05-13 | 2022-08-23 | 长飞光纤光缆股份有限公司 | Optical cable outer sheath gas polishing assembly, system and method in optical cable production |
| CN116228752A (en) * | 2023-05-06 | 2023-06-06 | 四川广安承平投资开发有限公司 | Surface flaw identification method and processing robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114937055B (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
| CN110349126B (en) | Convolutional neural network-based marked steel plate surface defect detection method | |
| CN111209952B (en) | Underwater target detection method based on improved SSD and migration learning | |
| CN108876768B (en) | Shadow defect detection method for light guide plate | |
| CN115684176B (en) | Online visual detection system for film surface defects | |
| CN112308854B (en) | Automatic detection method and system for chip surface flaws and electronic equipment | |
| CN112288726B (en) | Method for detecting foreign matters on belt surface of underground belt conveyor | |
| CN115272303B (en) | Textile fabric defect degree evaluation method, device and system based on Gaussian blur | |
| CN110186929A (en) | A kind of real-time product defect localization method | |
| CN115841434A (en) | Infrared image enhancement method for gas concentration analysis | |
| CN113657335A (en) | Mineral phase identification method based on HSV color space | |
| CN112330613B (en) | Evaluation method and system for cytopathology digital image quality | |
| CN117274113A (en) | Broken silicon wafer cleaning effect visual detection method based on image enhancement | |
| CN116843687A (en) | Communication optical cable surface flaw detection method and device | |
| CN114565563B (en) | Color steel plate surface abnormity detection method based on artificial intelligence | |
| CN117078608B (en) | Double-mask guide-based high-reflection leather surface defect detection method | |
| CN113052794A (en) | Image definition recognition method based on edge features | |
| Prabha et al. | Defect detection of industrial products using image segmentation and saliency | |
| CN118329910A (en) | Printing paper defect detection system and method based on machine vision | |
| CN111667509A (en) | Method and system for automatically tracking moving target under condition that target is similar to background color | |
| CN117011291A (en) | Watch shell quality visual detection method | |
| CN109448012A (en) | A kind of method for detecting image edge and device | |
| CN109934817A (en) | The external contouring deformity detection method of one seed pod | |
| CN104809712B (en) | A kind of image fast repairing method based on rough set | |
| CN108765365A (en) | A kind of rotor winding image qualification detection 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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190830 |