CN113781458A - Artificial intelligence based identification method - Google Patents
Artificial intelligence based identification method Download PDFInfo
- Publication number
- CN113781458A CN113781458A CN202111086196.7A CN202111086196A CN113781458A CN 113781458 A CN113781458 A CN 113781458A CN 202111086196 A CN202111086196 A CN 202111086196A CN 113781458 A CN113781458 A CN 113781458A
- Authority
- CN
- China
- Prior art keywords
- flaw
- image
- artificial intelligence
- detection
- intelligence based
- 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
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 15
- 230000007547 defect Effects 0.000 claims abstract description 35
- 238000001514 detection method Methods 0.000 claims abstract description 18
- 238000003708 edge detection Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 230000011218 segmentation Effects 0.000 claims abstract description 6
- 238000000576 coating method Methods 0.000 claims description 29
- 239000011248 coating agent Substances 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 14
- 230000002950 deficient Effects 0.000 claims description 11
- 230000000877 morphologic effect Effects 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 3
- 230000007797 corrosion Effects 0.000 claims description 3
- 238000005260 corrosion Methods 0.000 claims description 3
- 238000000926 separation method Methods 0.000 claims description 3
- 238000005253 cladding Methods 0.000 abstract description 11
- 238000013136 deep learning model Methods 0.000 abstract description 8
- 238000004445 quantitative analysis Methods 0.000 abstract description 4
- 238000009776 industrial production Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 abstract description 2
- PXHVJJICTQNCMI-UHFFFAOYSA-N Nickel Chemical compound [Ni] PXHVJJICTQNCMI-UHFFFAOYSA-N 0.000 description 14
- 229910052759 nickel Inorganic materials 0.000 description 7
- 238000013135 deep learning Methods 0.000 description 6
- 239000000463 material Substances 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 238000004372 laser cladding Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 238000003754 machining Methods 0.000 description 4
- 239000002184 metal Substances 0.000 description 4
- 229910052751 metal Inorganic materials 0.000 description 4
- 230000035882 stress Effects 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000000227 grinding Methods 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005204 segregation Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000008646 thermal stress Effects 0.000 description 1
- 238000007514 turning Methods 0.000 description 1
Images
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
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration by the use of local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
Abstract
The invention discloses an identification method based on artificial intelligence, belonging to the technical field of identification and detection; the method comprises the following steps: collecting an image; preprocessing an image; identifying flaws; extracting edges; classifying defects; the method has the advantages that firstly, an artificial intelligence flaw identification method is established based on a deep learning model, secondly, the area of the flaw is further calculated by utilizing an edge detection algorithm, and further the flaw classification is realized; compared with the image semantic segmentation algorithm for analysis, the method has the advantages that the condition that the image semantic segmentation algorithm is wrongly segmented due to complex flaws is avoided; the method realizes qualitative detection of the flaw through the deep learning model, realizes quantitative analysis of the flaw through the edge detection algorithm, and can greatly improve the detection efficiency, facilitate industrial production and detection and greatly reduce the labor cost compared with the traditional method of analyzing the flaw of the cladding layer through a microscope.
Description
Technical Field
The invention belongs to the technical field of identification and detection, and particularly relates to an identification method based on artificial intelligence.
Background
Laser cladding, also known as laser cladding or laser cladding, is a new surface modification technique. The method is characterized in that a cladding material is added on the surface of a base material, and the cladding material and a thin layer on the surface of the base material are fused together by utilizing a laser beam with high energy density, so that a metallurgically bonded cladding layer is formed on the surface of a base layer. However, the defects of the cladding layer are widely existed at present, and the development and the application of the laser cladding technology are seriously hindered. The defects of the cladding layer are caused by stress generated by the reasons of different thermal expansion coefficients of the cladding layer and the base material after laser cladding processing and cooling, element segregation in the cladding layer and the like, and the stress is mainly divided into constraint stress, thermal stress and structural stress. The current researchers mostly keep the research direction on how to improve the process for inhibiting the defects, and an effective technical scheme on how to detect the defects of the cladding layer is not available; the traditional detection is mostly carried out by adopting modes such as an electron microscope, a microscope and the like, but the mode has low efficiency and high labor cost, is difficult to use in industrial production, and is more difficult to detect when facing a large-area coating.
Disclosure of Invention
Technical problem to be solved
The invention aims to provide an artificial intelligence-based identification method for rapidly detecting defects of a cladding layer.
(II) technical scheme
The invention is realized by the following technical scheme: an artificial intelligence based identification method; the method is applied to detection of coating defects; the method comprises the following steps:
step 100: collecting an image;
step 200: preprocessing an image;
step 300: identifying flaws;
step 400: extracting edges;
step 500: and (4) classifying defects.
As a further explanation of the above scheme, in step 100, before image acquisition, a sample is machined and cleaned; and putting the processed sample piece into an image acquisition device, dividing the coating into at least two areas, and acquiring each area respectively.
As a further illustration of the above scheme, the step 200: the image preprocessing further comprises the following steps;
step 210: image registration;
step 220: graying processing and noise reduction.
As a further illustration of the above, the defect of step 300 comprises: cracks, bubbles, bubble-crack mixed defects; after the image is preprocessed in the step 200, extracting output characteristics, namely flaws; the aspect ratio of the flaw features is extracted by semantic segmentation,
if the length-width ratio is smaller than a first preset value, judging the flaw to be a crack;
if the length-width ratio is larger than the first preset value and smaller than the second preset value, judging the flaw to be a bubble crack mixed flaw;
if the length-width ratio is larger than a second preset value, the defect is judged to be a bubble.
As a further illustration of the above scheme, the step 400 specifically includes the following steps
Step 410: carrying out edge detection on the image and extracting edge details;
step 420: processing the flaw edge by adopting morphological expansion and morphological corrosion;
step 430: carrying out region separation by pixel point communication;
step 440: and calculating the area pixel value, and extracting the defective area when the area pixel value is greater than a third preset value.
As a further explanation of the above scheme, the step 500 specifically calculates a ratio of the area pixel value to the total pixels of the image, so as to obtain a ratio X; and (4) grading the flaw degree based on the comparison of the numerical value of X and the threshold value.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
FIG. 1 is a diagram illustrating a step 200 of image preprocessing according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating an edge extraction process in step 400 according to an embodiment of the present invention;
FIG. 3 is a schematic view of the installation of the device after model selection according to the embodiment of the present invention;
FIG. 4 is a schematic diagram of a minimum bounding rectangle of area No. 1 in the embodiment of the present invention;
in the figure: the detection device comprises a No. 1 area (1), a No. 2 area (2), a No. 3 area (3), a No. 4 area (4), a No. 5 area (5), a No. 6 area (6), a No. 7 area (7), a No. 8 area (8), a No. 9 area (9), bubbles (11), a bubble and crack mixed defect (12), cracks (13), a CCD camera (14), a lens (15), an annular diffuse reflection light source (16), a sample piece to be detected (17) and a moving device (18).
(III) advantageous effects
Compared with the prior art, the invention has the following beneficial effects: the method has the advantages that firstly, an artificial intelligence flaw identification method is established based on a deep learning model, secondly, the area of the flaw is further calculated by utilizing an edge detection algorithm, and further the flaw classification is realized; compared with the image semantic segmentation algorithm for analysis, the method has the advantages that the condition that the image semantic segmentation algorithm is wrongly segmented due to complex flaws is avoided; the method realizes qualitative detection of the flaw through the deep learning model, realizes quantitative analysis of the flaw through the edge detection algorithm, and can greatly improve the detection efficiency, facilitate industrial production and detection and greatly reduce the labor cost compared with the traditional method of analyzing the flaw of the cladding layer through a microscope.
Detailed Description
Examples
The invention provides an identification method based on artificial intelligence; the method is mainly applied to the detection of surface flaws of the coating after the secondary processing of the coating in the additive manufacturing is finished; the detection method only aims at the flat surface after secondary processing, and is not suitable for the coating surface which is not subjected to secondary processing;
step 100: collecting an image;
machining and cleaning a sample piece before image acquisition; and putting the processed sample piece into an image acquisition device, dividing the coating into at least two areas, and acquiring each area respectively. Firstly, it is to be noted that the roughness of the surface of the coating made of different materials after secondary machining is different, and particularly, the influence of the roughness of the surface of the coating after precision machining on image acquisition is large; therefore, in the embodiment, the metal coating with the widest application range is exemplified, and the standard of the acquisition device for image acquisition has the following requirements:
1) since the metal coating has many fine defects, it is necessary to secure sufficient brightness;
2) the surface of the metal coating after secondary processing is smooth, so that the surface of the metal coating is easy to reflect light, and image acquisition is further influenced;
3) the image acquisition part is realized by adopting a multi-region acquisition subsequent splicing step, so that tiny flaws are prevented from being missed, and therefore, the fact that the images are overlapped needs to be ensured, and the subsequent image registration is facilitated;
based on the above requirements, the image capturing device was chosen to state that the human eye, according to the common knowledge, discriminates a distance of 0.291mm, i.e. equal to about 0.3mm, at a distance of 1m, taking 0.3mm as the threshold for the minimum defects of the coating; because the invention adopts multi-image splicing detection to realize the detection of the flaws, one pixel is relatively considered to correspond to 0.3mm, and the precision of the image acquisition system is 0.3 mm/pixel; in this embodiment, the detected nickel-based coating is 60mm by 60mm, in this embodiment, the nickel-based coating is divided into 3 by 3 grid regions for respective detection, so the field of view of the camera in the horizontal direction is not less than 20mm, so the resolution in the horizontal direction is not less than 20mm/0.3 mm/pixel is 67 pixels, and the vertical resolution is not less than 67 pixels, which is the same as the resolution in the horizontal direction, therefore, in this embodiment, a CCD camera manufactured by BASLER company and having a resolution of acA640-90gc is used, the resolution is 658 × 492, the sensor is ICX424, the sensor size is 1/3 inches, the sensor type is CCD, the pixel size is 7.4 μm x 7.4.4 μm, the maximum full frame rate is 90fts, the image mode is color, and the pixels are 32 ten thousand pixels; according to the practical application requirements, a Basler Premium C-mouth lens manufactured by BASLER company is adopted, the fixed focal length is 25mm, the aperture range is F2.2-F22, and the resolution is 500 ten thousand pixels. After the nickel-based coating is secondarily processed, the surface is smooth, a diffuse reflection light source is selected to avoid light reflection, and meanwhile, an annular light source is more suitable to ensure the brightness, so that the annular diffuse reflection light source is adopted; after the model selection is finished, in order to ensure the shooting precision, the sample piece is moved, and the image acquisition device is carried out in a static mode; therefore, the sample piece is placed on the moving device, the sample piece is controlled to move, and the overlapping area is 30% -40% when shooting is carried out each time.
It should be noted that the type selection of the camera and the lens in this embodiment is not limited to the sample detected in this embodiment, and the type selection can be performed for the sample with a larger size, the shooting area per time, and the pixel requirement.
Step 200: preprocessing an image;
this step includes image registration; graying processing and noise reduction. The steps aim at constructing flaw identification, namely flaw quantitative analysis; it should be noted that the quantitative analysis refers to the number of defects, not the area of the defects.
The specific image registration is carried out by the following steps
Step 211: extracting and matching the characteristic points;
step 212: image registration;
step 213: and (5) image fusion.
It should be noted that step 200 is completed by using halcon software in this embodiment;
the image registration aspect accomplishes this using the operators proj _ match _ points _ ranging _ identified and gen _ project _ mosaic, which are self-contained using halcon software.
After the work is finished, a complete 60 mm-60 mm nickel-based coating image is obtained, then gray processing is carried out, the contrast degree of the defects of the coating is enhanced, the information content in the image is reduced, and the positions and the number of the defects can be better identified; the noise reduction part adopts a Gaussian filtering denoising method;
step 300: identifying flaws;
identifying the collected image by using the constructed deep learning model;
the deep learning model establishing method comprises the following steps:
step 310: establishing a learning network and experimental data;
step 320: training a deep learning model;
step 330: evaluating the model;
it should be noted that, in the present embodiment, the step 300 is implemented by using halcon software,
in this step, a pretrained network trained _ dl _ classifier _ compact of the deep learning classification method of the halcon software is called to establish the deep learning network in the embodiment; as can be seen from the model selection in step 100, the resolution of a single acquired image in this embodiment is 658 × 492; when a deep learning model is constructed, in order to overcome the problem that the GPU computational power is limited after image splicing is completed, a mode of separately identifying and training each region is adopted, so that the imported image resolution is 658 multiplied by 492; in the embodiment, 1800 images are collected on the surface of the nickel-based coating, namely nine area images are extracted on the surface of 200 nickel-based coatings, so that the basic amount during training is ensured, and meanwhile, a data set is further expanded to 3600 images in a horizontal overturning mode; the image is manually classified into two parts, namely a defective part and a non-defective part, wherein the number of the defective part is 1500, and the number of the non-defective part is 2100. And storing the two parts respectively; in order to ensure the training precision, the stored data is subjected to gray level processing and Gaussian filtering denoising processing; after the steps are completed, the data set is divided into a training set of 70%, a verification set of 15% and a verification set of 15%; finally, the model is trained and evaluated, which is not described herein again. The flaws of the nickel-based coating are divided into three conditions of cracks, bubbles and bubble crack mixed defects in the embodiment; after the defects are identified, the type of the defects is judged according to the aspect ratio of the minimum circumscribed rectangle of the defects.
It should be noted that there are various deep learning training networks, which are suitable for implementation of the present invention, and for convenience of explanation in this embodiment, a deep learning network provided in the halcon software is used for explanation; the content of the completion of the previous steps only comprises the identification and the number statistics of the flaws, and the extraction of the flaw area is completed in the next step; meanwhile, compared with image semantic recognition, the method adopting deep learning and manual work has better precision because the methods of machining the coating are more, such as grinding, polishing, turning, grinding and the like, if the image semantic recognition is adopted to automatically extract features, if the edge information after the defect noise reduction treatment is similar to the information of the coating, the algorithm cannot well segment the defect and the matrix, the result is not accurate enough, the image semantic recognition cannot accurately extract the defect, and the detection precision is further reduced.
Step 400: extracting edges;
after the steps are completed and corresponding flaws are extracted, edge extraction can be carried out on the positions of the flaws; specifically comprises the following steps
Step 410: carrying out edge detection on the image of the flaw part, and extracting edge details;
step 420: processing the flaw edge by adopting morphological expansion and morphological corrosion;
step 430: carrying out region separation by pixel point communication;
step 440: and calculating the area pixel value, and extracting the defective area when the area pixel value is greater than a third preset value.
It should be noted that the size of the coating defect needs to be determined according to the actual situation, that is, when the pixel value of the defect is less than a small value, the defect can be regarded as qualified, the third preset value here is to set a threshold value according to needs, extract features according to the threshold value, and easily cause excessive noise when there is a foreign object on the surface of the coating, which affects the judgment of the result, and if the defect is not eliminated by setting the threshold value, the obtained result does not conform to the actual situation, and the accuracy is not sufficient.
Step 500: and (4) classifying defects. Specifically calculating the ratio of the pixel value of the area to the total pixels of the image to obtain a ratio X; and (4) grading the flaw degree based on the comparison of the numerical value of X and the threshold value. The area pixel value is the pixel value of the flaw position, the percentage X of the area of the flaw in the total coating area can be known through the ratio of the pixel value to the total pixels of the image, and then samples of different flaw conditions are classified through percentage. The defective part is 1500 sheets, and the non-defective part
The confusion matrix fed back from the test set in this embodiment is shown in the following table:
TABLE 1 confusion matrix for test set
As can be seen from the above table, the accuracy of the determination of the defective part category in the test set is (218/225) × 100% — 96%; the accuracy rate of judging the non-defective part is (307/315) × 100% > (97%), and the probability of judging the correctness of all the test set samples is (525/540) × 100% > (97%);
and calculating the accuracy X of the model by the result fed back by the confusion matrix, wherein the formula is as follows:
X＝TP/(TP+FP)
wherein TP indicates positive prediction and is marked as positive; FP indicates positive prediction and negative marker;
further, X is found to be 218/(218+8) 0.96;
and calculating the recall Y of the model according to the result fed back by the confusion matrix, wherein the formula is as follows:
Y＝TP/(TP+FN)
wherein TP indicates positive prediction and is marked as positive; FN indicates negative prediction, marked positive;
further, Y is found to be 218/(218+7) is found to be 0.97;
the harmonic mean F, which is a measure of the accuracy of the classifier, is calculated from the values of X and Y. The formula is as follows:
F＝(2*X*Y)/(X+Y)＝(2*0.96*0.97)/(0.96+0.97)＝0.96；
by combining the above calculations, it can be seen that each calculation result in this embodiment is higher than 95%, the network performance is sufficient to meet the actual application requirements, and meanwhile, according to the network characteristics of deep learning, as the number of learning samples increases, the accuracy is higher, and the subsequent samples are continuously increased, which can be continuously improved.
While there have been shown and described what are at present considered the fundamental principles and essential features of the invention and its advantages, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing exemplary embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
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 (6)
1. An artificial intelligence based identification method; the method is applied to detection of coating defects; the method is characterized in that: the method comprises the following steps:
step 100: collecting an image;
step 200: preprocessing an image;
step 300: identifying flaws;
step 400: extracting edges;
step 500: and (4) classifying defects.
2. The artificial intelligence based recognition method of claim 1, wherein: in step 100, before image acquisition, a sample is machined and cleaned; and putting the processed sample piece into an image acquisition device, dividing the coating into at least two areas, and acquiring each area respectively.
3. The artificial intelligence based recognition method of claim 1, wherein: the step 200: the image preprocessing further comprises the following steps;
step 210: image registration;
step 220: graying processing and noise reduction.
4. The artificial intelligence based recognition method of claim 1, wherein: the step 300 of the defect includes: cracks, bubbles, bubble-crack mixed defects; after the image is preprocessed in the step 200, extracting output characteristics, namely flaws; the aspect ratio of the flaw features is extracted by semantic segmentation,
if the length-width ratio is smaller than a first preset value, judging the flaw to be a crack;
if the length-width ratio is larger than the first preset value and smaller than the second preset value, judging the flaw to be a bubble crack mixed flaw;
if the length-width ratio is larger than a second preset value, the defect is judged to be a bubble.
5. The artificial intelligence based recognition method of claim 1, wherein: the step 400 specifically includes the following steps
Step 410: carrying out edge detection on the image and extracting edge details;
step 420: processing the flaw edge by adopting morphological expansion and morphological corrosion;
step 430: carrying out region separation by pixel point communication;
step 440: and calculating the area pixel value, and extracting the defective area when the area pixel value is greater than a third preset value.
6. The artificial intelligence based recognition method of claim 1, wherein: specifically, in the step 500, a ratio of the area pixel value to the total pixels of the image is calculated to obtain a ratio X; and (4) grading the flaw degree based on the comparison of the numerical value of X and the threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111086196.7A CN113781458A (en) | 2021-09-16 | 2021-09-16 | Artificial intelligence based identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111086196.7A CN113781458A (en) | 2021-09-16 | 2021-09-16 | Artificial intelligence based identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113781458A true CN113781458A (en) | 2021-12-10 |
Family
ID=78851372
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111086196.7A Pending CN113781458A (en) | 2021-09-16 | 2021-09-16 | Artificial intelligence based identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113781458A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115018829A (en) * | 2022-08-03 | 2022-09-06 | 创新奇智(成都)科技有限公司 | Glass flaw positioning method and device |
| CN115239725A (en) * | 2022-09-22 | 2022-10-25 | 张家港大裕橡胶制品有限公司 | Visual monitoring method and system for rubber dipping coating of rubber glove |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102879401A (en) * | 2012-09-07 | 2013-01-16 | 西安工程大学 | Method for automatically detecting and classifying textile flaws based on pattern recognition and image processing |
| CN105931246A (en) * | 2016-05-05 | 2016-09-07 | 东华大学 | Fabric flaw detection method based on wavelet transformation and genetic algorithm |
| CN110889837A (en) * | 2019-11-25 | 2020-03-17 | 东华大学 | Cloth flaw detection method with flaw classification function |
| WO2020206861A1 (en) * | 2019-04-08 | 2020-10-15 | 江西理工大学 | Yolo v3-based detection method for key object at transportation junction |
| CN111862064A (en) * | 2020-07-28 | 2020-10-30 | 桂林电子科技大学 | Silver wire surface flaw identification method based on deep learning |
-
2021
- 2021-09-16 CN CN202111086196.7A patent/CN113781458A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102879401A (en) * | 2012-09-07 | 2013-01-16 | 西安工程大学 | Method for automatically detecting and classifying textile flaws based on pattern recognition and image processing |
| CN105931246A (en) * | 2016-05-05 | 2016-09-07 | 东华大学 | Fabric flaw detection method based on wavelet transformation and genetic algorithm |
| WO2020206861A1 (en) * | 2019-04-08 | 2020-10-15 | 江西理工大学 | Yolo v3-based detection method for key object at transportation junction |
| CN110889837A (en) * | 2019-11-25 | 2020-03-17 | 东华大学 | Cloth flaw detection method with flaw classification function |
| CN111862064A (en) * | 2020-07-28 | 2020-10-30 | 桂林电子科技大学 | Silver wire surface flaw identification method based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| 中国机械工程学会焊接学会: "《焊接手册 第3卷 焊接结构 第2版》", 31 August 2001, 北京：机械工业出版社, pages: 1110 * |
| 刘宏志: "《推荐系统》", 31 May 2020, 北京：机械工业出版社, pages: 88 - 89 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115018829A (en) * | 2022-08-03 | 2022-09-06 | 创新奇智(成都)科技有限公司 | Glass flaw positioning method and device |
| CN115239725A (en) * | 2022-09-22 | 2022-10-25 | 张家港大裕橡胶制品有限公司 | Visual monitoring method and system for rubber dipping coating of rubber glove |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113362326B (en) | Method and device for detecting defects of welding spots of battery | |
| CN111650210B (en) | Burr detection method and detection system for high-speed high-precision lithium ion battery pole piece | |
| CN102175700B (en) | Method for detecting welding seam segmentation and defects of digital X-ray images | |
| CN113781458A (en) | Artificial intelligence based identification method | |
| CN101670533A (en) | Cutting-tool wear state evaluating method based on image analysis of workpiece machining surface | |
| CN109376773A (en) | Crack detecting method based on deep learning | |
| CN115760654B (en) | Industrial microscope image processing system | |
| CN113325001A (en) | Automatic distinguishing and detecting equipment and method for surface appearance flaws of copper foil | |
| CN104792794A (en) | Machine vision based optical film surface defect detecting method | |
| CN111523540A (en) | Metal surface defect detection method based on deep learning | |
| CN114910480A (en) | Wafer surface defect detection method based on machine vision | |
| CN113177924A (en) | Industrial production line product flaw detection method | |
| CN113393426A (en) | Method for detecting surface defects of rolled steel plate | |
| CN111260617A (en) | Solar cell panel defect detection method based on deep learning | |
| CN110426395B (en) | Method and device for detecting surface of solar EL battery silicon wafer | |
| CN115471466A (en) | Steel surface defect detection method and system based on artificial intelligence | |
| CN115601359A (en) | Welding seam detection method and device | |
| CN111415330A (en) | Copper foil appearance defect detection method based on deep learning | |
| CN113962929A (en) | Photovoltaic cell assembly defect detection method and system and photovoltaic cell assembly production line | |
| CN116091505B (en) | Automatic defect detection and classification method and system for sapphire substrate | |
| CN112581424B (en) | Classification extraction method for surface and subsurface defects of optical element | |
| CN114813748A (en) | Steel surface defect detection method and system based on machine vision | |
| CN113469988A (en) | Defect identification method | |
| CN114494142A (en) | Mobile terminal middle frame defect detection method and device based on deep learning | |
| CN114113116A (en) | Accurate detection process method for micro-defects on surface of large-diameter element |
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 |