CN116309564B - Method and system for detecting appearance defects of battery cells based on artificial intelligent image recognition - Google Patents
Method and system for detecting appearance defects of battery cells based on artificial intelligent image recognition Download PDFInfo
- Publication number
- CN116309564B CN116309564B CN202310553482.2A CN202310553482A CN116309564B CN 116309564 B CN116309564 B CN 116309564B CN 202310553482 A CN202310553482 A CN 202310553482A CN 116309564 B CN116309564 B CN 116309564B
- Authority
- CN
- China
- Prior art keywords
- battery cell
- identification
- appearance
- current
- standard
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000007547 defect Effects 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000001514 detection method Methods 0.000 claims abstract description 28
- 206010052428 Wound Diseases 0.000 claims description 45
- 208000027418 Wounds and injury Diseases 0.000 claims description 45
- 238000011156 evaluation Methods 0.000 claims description 41
- 230000002159 abnormal effect Effects 0.000 claims description 39
- 230000001953 sensory effect Effects 0.000 claims description 19
- 238000012360 testing method Methods 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 13
- 238000009826 distribution Methods 0.000 claims description 12
- 238000004806 packaging method and process Methods 0.000 claims description 11
- 238000013473 artificial intelligence Methods 0.000 claims description 9
- 208000014674 injury Diseases 0.000 claims description 9
- 230000008733 trauma Effects 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 8
- 238000005314 correlation function Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 238000012795 verification Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000012800 visualization Methods 0.000 claims description 5
- 238000012512 characterization method Methods 0.000 claims description 4
- 238000013441 quality evaluation Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000005859 cell recognition Effects 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 5
- 210000004027 cell Anatomy 0.000 description 208
- 239000000047 product Substances 0.000 description 20
- 230000009286 beneficial effect Effects 0.000 description 11
- 238000013461 design Methods 0.000 description 8
- 230000000007 visual effect Effects 0.000 description 8
- 238000004519 manufacturing process Methods 0.000 description 5
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 229910052744 lithium Inorganic materials 0.000 description 4
- 238000012797 qualification Methods 0.000 description 3
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 2
- 239000004020 conductor Substances 0.000 description 2
- 210000005069 ears Anatomy 0.000 description 2
- 229910001416 lithium ion Inorganic materials 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 229910052751 metal Inorganic materials 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 229920000642 polymer Polymers 0.000 description 2
- 239000002994 raw material Substances 0.000 description 2
- 230000000472 traumatic effect Effects 0.000 description 2
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 206010047571 Visual impairment Diseases 0.000 description 1
- 229910052782 aluminium Inorganic materials 0.000 description 1
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 description 1
- 229910052802 copper Inorganic materials 0.000 description 1
- 239000010949 copper Substances 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a battery cell appearance defect detection method and system based on artificial intelligent image recognition, wherein the method comprises the following steps: standard size information and standard appearance information of the battery cell of the target specification are obtained, and battery cell identification characteristics are extracted according to the standard size information and the standard appearance information; constructing an identification model of the battery cell with the target specification according to the battery cell identification characteristics and a preset learning network model, and a plurality of identification strategies of the identification model; acquiring a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model; and determining each appearance index defect of the current battery cell according to the identification result of the identification model. The appearance detection can be carried out by using the model instead of manpower, the appearance detection is carried out on each processing battery cell without errors through the pre-recorded accurate appearance parameters, the detection precision and efficiency are improved, the manpower use cost is reduced, the processing quality of battery cell products is ensured, and the practicability is improved.
Description
Technical Field
The invention relates to the technical field of battery cell detection, in particular to a battery cell appearance defect detection method and system based on artificial intelligent image recognition.
Background
The battery cell is a raw material of a lithium ion polymer battery product, the battery cell is provided with electrode lugs, the electrode lugs are metal conductors which are led out from the battery cell, namely the ears of the positive electrode and the negative electrode of the battery cell in common, and one battery cell can extend out of the two electrode lugs.
Whether the electrode lug of the battery core is bent or folded is an important factor directly related to the quality of a lithium battery product in the production process of the lithium battery, so that whether the electrode lug of the battery core is bent or folded needs to be detected; in addition, flaws can occur in the surface appearance of some cells: if the upper surface and the lower surface of the battery cell, the upper edge and the lower edge can have defects such as scratches, pits, protruding points and the like, the appearance is influenced, the production and the use safety of the battery cell are also greatly influenced, the appearance of the battery cell needs to be detected after the battery cell is produced, the traditional detection mode is to observe and identify through human eyes, the human labor is wasted, meanwhile, detection errors can be caused, the quality of products and the subsequent use experience cannot be guaranteed, and the practicability is reduced.
Disclosure of Invention
Aiming at the problems displayed above, the invention provides a battery cell appearance defect detection method and system based on artificial intelligent image recognition, which are used for solving the problems that human labor is wasted, detection errors exist at the same time, the quality and subsequent use experience of products cannot be ensured, and the practicability is reduced by manually observing and recognizing the battery cell appearance defects by human eyes in the conventional technology in the background art.
A battery cell appearance defect detection method based on artificial intelligent image recognition comprises the following steps:
standard size information and standard appearance information of the battery cell of the target specification are obtained, and battery cell identification characteristics are extracted according to the standard size information and the standard appearance information;
constructing an identification model of the battery cell with the target specification according to the battery cell identification characteristics and a preset learning network model, and a plurality of identification strategies of the identification model;
acquiring a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model;
and determining each appearance index defect of the current battery cell according to the identification result of the identification model.
Preferably, the obtaining the standard size information and the standard appearance information of the target specification battery cell, and extracting the battery cell identification feature according to the standard size information and the standard appearance information, includes:
Acquiring a standard finished product sample of a target specification battery cell, and acquiring finished product images of the standard finished product sample based on a plurality of shooting angles;
analyzing the finished product images of the shooting angles to determine standard size information, standard shape information, standard color information and standard packaging information of the battery cell of the target specification;
integrating the standard shape information, the standard color information and the standard packaging information of the target specification battery cell to obtain the standard appearance information of the target specification battery cell;
and extracting the cell size identification feature according to the standard size information, extracting the cell appearance identification feature according to the standard appearance information, and integrating the cell size identification feature and the cell appearance identification feature to generate the cell identification feature of the cell with the target specification.
Preferably, the extracting the cell size identifying feature according to the standard size information, and extracting the cell appearance identifying feature according to the standard appearance information at the same time, includes:
acquiring a size description parameter of a battery cell of a target specification according to the standard size information, and extracting a battery cell size identification characteristic based on the size description parameter;
acquiring the color attribute, the packaging attribute and the shape attribute of the battery cell of the target specification according to the standard appearance information;
Acquiring feature description factors corresponding to the color attribute, the packaging attribute and the shape attribute respectively, and performing entity definition on the feature description factors to acquire a definition result;
and obtaining the appearance global characteristics of the battery cell with the target specification according to the definition result, and obtaining the appearance identification characteristics of the battery cell according to the appearance global characteristics.
Preferably, the construction of the identification model of the target specification cell and the multiple identification strategies of the identification model according to the cell identification feature and the preset learning network model includes:
acquiring a plurality of image samples of a target specification cell and dividing the image samples into a training set, a testing set and a verification set;
writing the cell identification characteristics into a preset learning network model to construct an identification model of the cell with the target specification, and respectively carrying out repeated training and testing on the identification model by utilizing a training set, a testing set and a verification set until the identification accuracy of the identification model is greater than or equal to a preset threshold value;
determining gray value intervals of the image with the visualization conditions and dividing all interval values into a plurality of gray levels;
acquiring dominant pixel parameters and recessive pixel parameters of the cell image under each gray level, and setting a recognition strategy of a recognition model for the cell image of each gray level according to the dominant pixel parameters and the recessive pixel parameters.
Preferably, obtaining a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model, including:
judging whether the high-definition image of the current battery cell has necessary identification characteristics according to the display content of the high-definition image, if so, judging that the high-definition image is qualified, and if not, judging that the high-definition image is unqualified;
randomly extracting n pixel points from the high-definition image, determining the current gray level of each pixel point, and determining the current gray level of the high-definition image according to the current gray level;
determining a target recognition strategy of a recognition model according to the current gray level and starting a corresponding current model recognition mode;
the high-definition image is input into a recognition model based on the current model recognition mode.
Preferably, the determining each appearance index defect of the current battery cell according to the identification result of the identification model includes:
determining a current size index, a current shape index and a current color index of the current battery cell according to the identification result of the identification model;
comparing the current size index, the current shape index and the current color index of the current battery cell with the standard size index, the standard shape index and the standard color index of the battery cell of the target specification to obtain a comparison result;
Determining defect parameters of each appearance index according to the comparison result, and determining the defect grade of each appearance index according to the defect parameters;
each appearance index defect is classified into an acceptable appearance index defect and an unacceptable appearance index defect based on the defect level of each appearance index.
Preferably, the method further comprises:
after determining that the appearance index of the current battery cell is defect-free, acquiring the surface texture characteristics of the current battery cell according to the high-definition image;
analyzing the surface texture characteristics, and judging whether scratches, pits or salient point detail wounds exist on the surface of the current battery cell according to the analysis result;
if yes, marking and amplifying the image area with scratches and pits or bumps in the high-definition image;
and generating a process quality evaluation report of the current battery cell according to the distribution condition of the surface scratches and pits or protruding points of the current battery cell.
Preferably, the analyzing the surface texture features, determining whether the surface of the current battery core has scratches and pits or protruding points or not is traumatic according to the analysis result, including:
constructing a linear correlation function between the sensory quality and the texture evaluation index, and determining the sensory parameters of the electric core entity under different texture evaluation index values according to the linear correlation function;
Performing feature characterization on the surface texture features to obtain a first certainty result, and determining class display attributes corresponding to the surface texture features according to the first certainty result;
determining current sensory parameters for the surface texture features according to the class display attributes corresponding to the surface texture features;
determining a feasible texture evaluation index item for the current battery cell based on the current sensory parameter and the battery cell entity sensory parameters under different texture evaluation index values;
determining texture display parameters of each pixel point according to the surface texture characteristics and pixel distribution of the high-definition image, and constructing a texture display parameter set according to the texture display parameters;
evaluating each texture display parameter in the texture display parameter set based on a feasible texture evaluation index item of the current battery cell to obtain an evaluation result;
and judging whether the surface of the current battery cell has scratches, pits or convex points and detail wounds according to the evaluation result.
Preferably, the determining whether the surface of the current cell has scratches and pits or protruding points or not is performed according to the evaluation result, including:
screening abnormal texture display parameters with unqualified texture evaluation indexes according to the evaluation results, and determining parameter abnormal characteristics according to the distribution conditions;
Performing detail trauma qualitative on the abnormal texture display parameters according to the abnormal characteristics of the parameters to obtain a second qualitative result;
determining the wound types corresponding to the abnormal texture display parameters according to the second qualitative results, and obtaining a standard display parameter set corresponding to each wound type;
matching the current display parameter set corresponding to each abnormal texture display parameter with the standard display parameter set corresponding to the wound type of the abnormal texture display parameter, and determining whether the current display parameter set corresponding to each abnormal texture display parameter has a display wound or not according to a matching result;
if yes, determining the wound grade according to the matching interval of the current display parameter set in the standard display parameter set;
and judging whether the surface of the current battery cell has scratches, pits or salient points and detail wounds or not according to the wound grade and the wound type corresponding to each abnormal texture display parameter.
An artificial intelligence image recognition-based cell appearance defect detection system, the system comprising:
the extraction module is used for acquiring standard size information and standard appearance information of the battery cell with the target specification and extracting battery cell identification characteristics according to the standard size information and the standard appearance information;
the construction module is used for constructing an identification model of the battery cell with the target specification and a plurality of identification strategies of the identification model according to the battery cell identification characteristics and a preset learning network model;
The input module is used for acquiring a high-definition image of the current battery cell, determining a target recognition strategy of the recognition model according to the high-definition image and inputting the high-definition image into the recognition model;
and the determining module is used for determining each appearance index defect of the current battery cell according to the identification result of the identification model.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description and drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention.
FIG. 1 is a workflow diagram of a method for detecting appearance defects of a battery cell based on artificial intelligent image recognition;
FIG. 2 is another workflow diagram of a method for detecting appearance defects of a battery cell based on artificial intelligent image recognition provided by the invention;
FIG. 3 is a further workflow diagram of a method for detecting appearance defects of a battery cell based on artificial intelligent image recognition according to the present invention;
fig. 4 is a schematic structural diagram of a system for detecting appearance defects of a battery cell based on artificial intelligent image recognition.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples are not representative of all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present disclosure as detailed in the accompanying claims.
The battery cell is a raw material of a lithium ion polymer battery product, the battery cell is provided with electrode lugs, the electrode lugs are metal conductors which are led out from the battery cell, namely the ears of the positive electrode and the negative electrode of the battery cell in common, and one battery cell can extend out of the two electrode lugs.
Whether the electrode lug of the battery core is bent or folded is an important factor directly related to the quality of a lithium battery product in the production process of the lithium battery, so that whether the electrode lug of the battery core is bent or folded needs to be detected; in addition, flaws can occur in the surface appearance of some cells: if the upper surface and the lower surface of the battery cell, the upper edge and the lower edge can have defects such as scratches, pits, protruding points and the like, the appearance is influenced, the production and the use safety of the battery cell are also greatly influenced, the appearance of the battery cell needs to be detected after the battery cell is produced, the traditional detection mode is to observe and identify through human eyes, the human labor is wasted, meanwhile, detection errors can be caused, the quality of products and the subsequent use experience cannot be guaranteed, and the practicability is reduced. In order to solve the above problems, the present embodiment discloses a method for detecting appearance defects of a battery cell based on artificial intelligence image recognition.
A battery cell appearance defect detection method based on artificial intelligent image recognition, as shown in figure 1, comprises the following steps:
step S101, standard size information and standard appearance information of a battery cell with a target specification are obtained, and battery cell identification features are extracted according to the standard size information and the standard appearance information;
step S102, constructing an identification model of a target specification battery cell and a plurality of identification strategies of the identification model according to the battery cell identification characteristics and a preset learning network model;
step S103, obtaining a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model;
and step S104, determining each appearance index defect of the current battery cell according to the identification result of the identification model.
In this embodiment, the target specification cell is represented as a standard cell of different specifications;
in this embodiment, the standard size information is expressed as a standard process size parameter of the target specification cell;
in this embodiment, the standard appearance information is expressed as appearance parameters such as shape, package, color, etc. of the target specification cell in the absence of defects;
in this embodiment, the cell identification feature is expressed as an image identification feature of the target specification cell;
In this embodiment, the identification model is used to identify whether the shape and size of the current battery cell, the color package, etc. meet the preset design specification;
in this embodiment, the recognition policy is represented as a recognition mode policy of a recognition model for different gray-scale images;
in this embodiment, the appearance index defect is represented as a defect condition of each appearance index of the current cell, for example: the shape is not standard, the size is too large or too small, the package is unqualified, etc.
The working principle of the technical scheme is as follows: standard size information and standard appearance information of the battery cell of the target specification are obtained, and battery cell identification characteristics are extracted according to the standard size information and the standard appearance information; constructing an identification model of the battery cell with the target specification according to the battery cell identification characteristics and a preset learning network model, and a plurality of identification strategies of the identification model; acquiring a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model; and determining each appearance index defect of the current battery cell according to the identification result of the identification model.
The beneficial effects of the technical scheme are as follows: the method has the advantages that the appearance detection can be intelligently carried out on the processing battery cells by constructing the identification model of the battery cells and then utilizing the image identification mode, the model can be used for replacing manual appearance detection, the appearance detection is carried out on each processing battery cell without errors through the accurate appearance parameters input in advance, the detection precision and efficiency are improved, the labor use cost is reduced, the processing quality of battery cell products is guaranteed, the practicability is improved, the problems that the traditional detection mode in the prior art is not only wasteful in labor, but also has detection errors due to human eye observation and identification, the quality of products and subsequent use experience cannot be guaranteed, and the practicability is reduced are solved.
In this embodiment, before determining the target recognition strategy of the recognition model according to the high-definition image and inputting the high-definition image into the recognition model, the method further includes:
acquiring an electric core image contour and an interference image contour in a high-definition image;
respectively acquiring average pixel values in the contour of the cell image and the contour area of the interference image and respective display elements of the average pixel values;
calculating visual interference indexes of the interference image contour according to the average pixel values in the contour of the cell image and the contour area of the interference image and the respective display elements of the average pixel values and the contour area of the interference image:
；
where Q is expressed as a visual disturbance index which disturbs the image profile,visual scale expressed as high definition image, +.>Expressed as natural logarithm>Represented as average pixel value in the cell image profile +.>Expressed as average pixel value in the interference image profile,/->Definition expressed as high definition image, +.>Expressed as logarithm>Represented as display elements in the cell image profile, < >>Represented as display elements in the outline of the interference image, < >>Cell image contour regionColor confusion index with disturbing image contour region, +.>Visual display deviation represented as a high definition image;
determining whether a visual interference index of the interference image contour is larger than or equal to a preset index, if so, cutting the interference image contour in the high-definition image to obtain a processed high-definition image;
And confirming the processed high-definition image as an image sample to determine a target recognition strategy of the recognition model.
The beneficial effects of the technical scheme are as follows: visual influence of image elements except the battery cell image in the high-definition image on the battery cell image can be intuitively evaluated, so that the battery cell image is selectively cut, the identification stability of the high-definition image is ensured, the visual error caused by interference factors is reduced, and the practicability and the stability are further improved.
In one embodiment, as shown in fig. 2, the obtaining the standard size information and the standard appearance information of the target specification cell, and extracting the cell identification feature according to the standard size information and the standard appearance information includes:
step S201, acquiring a standard finished product sample of a target specification battery cell, and acquiring finished product images of the standard finished product sample based on a plurality of shooting angles;
step S202, analyzing finished product images of a plurality of shooting angles to determine standard size information, standard shape information, standard color information and standard package information of a target specification battery cell;
step S203, integrating the standard shape information, the standard color information and the standard packaging information of the battery cell with the target specification to obtain the standard appearance information of the battery cell with the target specification;
And S204, extracting the cell size identification feature according to the standard size information, extracting the cell appearance identification feature according to the standard appearance information, and integrating the cell size identification feature and the cell appearance identification feature to generate the cell identification feature of the cell with the target specification.
In this embodiment, the standard finished product sample is represented as a standard process finished product sample of the target specification cell;
in this embodiment, the final image is represented as a photographed image of a standard process final product of the target specification cell;
in this embodiment, the standard size information is represented as a standard design size parameter of the target specification cell, the standard shape information is represented as a standard design shape parameter of the target specification cell, the standard color information is represented as a standard design material color parameter standard package information of the target specification cell, and the standard design appearance package parameter of the target specification cell;
in this embodiment, the cell size identification feature is expressed as a reference feature that identifies the size of the target specification cell;
in this embodiment, the cell appearance identification feature is represented as a reference feature that identifies the appearance of the target specification cell.
The beneficial effects of the technical scheme are as follows: the accuracy and quality of the reference features can be guaranteed by using the finished product image of the battery cell with the target specification as the reference condition to acquire the identification features, accurate and effective model parameters are provided for subsequent model construction, and the practicability is further improved.
In one embodiment, as shown in fig. 3, the extracting the cell size identification feature according to the standard size information, and simultaneously extracting the cell appearance identification feature according to the standard appearance information, includes:
step S301, acquiring a size description parameter of a battery cell of a target specification according to standard size information, and extracting a battery cell size identification characteristic based on the size description parameter;
step S302, obtaining the color attribute, the packaging attribute and the shape attribute of the battery cell of the target specification according to the standard appearance information;
step S303, acquiring feature description factors corresponding to the color attribute, the packaging attribute and the shape attribute respectively, and performing entity definition on the feature description factors to acquire a definition result;
and S304, obtaining the appearance global characteristics of the battery cells of the target specification according to the definition result, and obtaining the appearance identification characteristics of the battery cells according to the appearance global characteristics.
In this embodiment, the size description parameter is expressed as the three-dimensional size, length, width, height, etc. of the target specification battery cell;
in this embodiment, the size identifying feature is represented as a digital description feature corresponding to the size description parameter of the target specification cell, for example: the length of the battery core is 5cm, the width is 3cm, and the height is 6cm;
in the present embodiment, the color attribute is represented as a display attribute of color, for example: green, red, package properties are expressed as skin material properties of the target specification cell, such as: copper, aluminum, etc., the shape attribute is expressed as a shape description of the target specification cell, for example: cuboid, cube, cylinder, etc.;
In this embodiment, the entity definition is represented by defining the feature descriptors corresponding to the color attribute, the package attribute, and the shape attribute on the same entity;
in this embodiment, the appearance global feature is represented as a global appearance feature of the overall appearance of the target specification cell.
The beneficial effects of the technical scheme are as follows: the global description characteristic of the battery cell with the target specification can be comprehensively obtained through entity definition, so that the accurate battery cell appearance identification characteristic is obtained, the obtained characteristic is more comprehensive and accords with the standard, and the practicability is further improved.
In one embodiment, the construction of the identification model of the target specification cell and the plurality of identification strategies of the identification model according to the cell identification characteristics and the preset learning network model includes:
acquiring a plurality of image samples of a target specification cell and dividing the image samples into a training set, a testing set and a verification set;
writing the cell identification characteristics into a preset learning network model to construct an identification model of the cell with the target specification, and respectively carrying out repeated training and testing on the identification model by utilizing a training set, a testing set and a verification set until the identification accuracy of the identification model is greater than or equal to a preset threshold value;
Determining gray value intervals of the image with the visualization conditions and dividing all interval values into a plurality of gray levels;
acquiring dominant pixel parameters and recessive pixel parameters of the cell image under each gray level, and setting a recognition strategy of a recognition model for the cell image of each gray level according to the dominant pixel parameters and the recessive pixel parameters.
In this embodiment, the preset threshold may be 99%;
in the present embodiment, the visualization condition is represented as a defined condition in which an element in an image can be substantially visualized;
in the present embodiment, the number of divided gradation values within each gradation level is the same;
in this embodiment, the dominant pixel parameter is expressed as a pixel parameter that can be observed by human eyes in the cell image at each gray level;
in this embodiment, the implicit pixel parameter is represented as a pixel parameter that cannot be observed by the human eye, but can be observed by an instrument, in the cell image at each gray level.
The beneficial effects of the technical scheme are as follows: the recognition accuracy of the model can be ensured to meet the expectations by repeatedly training and testing the recognition model, the accuracy and the reliability of appearance detection of the battery cell image are improved, and furthermore, the recognition strategy distribution and configuration can be accurately carried out on the pixel parameters of each gray level by recognizing the recognition strategy of the battery cell image of each gray level according to the dominant pixel parameters and the recessive pixel parameters of the battery cell image of each gray level, and the suitability and the model recognition reliability are improved.
In one embodiment, obtaining a high-definition image of a current cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model, comprises:
judging whether the high-definition image of the current battery cell has necessary identification characteristics according to the display content of the high-definition image, if so, judging that the high-definition image is qualified, and if not, judging that the high-definition image is unqualified;
randomly extracting n pixel points from the high-definition image, determining the current gray level of each pixel point, and determining the current gray level of the high-definition image according to the current gray level;
determining a target recognition strategy of a recognition model according to the current gray level and starting a corresponding current model recognition mode;
the high-definition image is input into a recognition model based on the current model recognition mode.
In this embodiment, the necessary identification feature is represented as a necessary identification medium feature for performing appearance identification on the current cell according to the high-definition image;
in this embodiment, the current model identification pattern is represented as a cell appearance identification pattern that identifies the model under the target identification policy.
The beneficial effects of the technical scheme are as follows: the reliability and the accuracy of the appearance recognition of the recognition model to the battery cell image can be ensured to the greatest extent by determining the recognition strategy corresponding to the high-definition image and then starting the corresponding recognition mode for recognition, so that the stability and the practicability are further improved.
In one embodiment, the determining each appearance index defect of the current cell according to the identification result of the identification model includes:
determining a current size index, a current shape index and a current color index of the current battery cell according to the identification result of the identification model;
comparing the current size index, the current shape index and the current color index of the current battery cell with the standard size index, the standard shape index and the standard color index of the battery cell of the target specification to obtain a comparison result;
determining defect parameters of each appearance index according to the comparison result, and determining the defect grade of each appearance index according to the defect parameters;
each appearance index defect is classified into an acceptable appearance index defect and an unacceptable appearance index defect based on the defect level of each appearance index.
In this embodiment, the defect parameter is represented as a difference parameter between a current index value and a preset index value of each appearance index;
in the present embodiment, the defect level is expressed as a level corresponding to the defect level of each appearance index, for example: the original design size of the battery core is 6 cubic centimeters, the size of the battery core which is actually detected is 8 cubic centimeters, and the size deviation at the moment is 33 percent, namely the medium defect grade; if the size deviation is 50% or more, the defect grade is high;
In this embodiment, the acceptable appearance index defect is expressed as an index defect reference value within a reasonable design error;
in this embodiment, the unacceptable visual index defect is expressed as an index defect reference value within an unreasonable design error.
The beneficial effects of the technical scheme are as follows: the defect parameters of each appearance index can be rapidly determined by utilizing an index comparison mode, the defect grade of each appearance index is further determined, the defects of each current battery cell can be rapidly identified and judged, and the working efficiency and the practicability are further improved.
In one embodiment, the method further comprises:
after determining that the appearance index of the current battery cell is defect-free, acquiring the surface texture characteristics of the current battery cell according to the high-definition image;
analyzing the surface texture characteristics, and judging whether scratches, pits or salient point detail wounds exist on the surface of the current battery cell according to the analysis result;
if yes, marking and amplifying the image area with scratches and pits or bumps in the high-definition image;
and generating a process quality evaluation report of the current battery cell according to the distribution condition of the surface scratches and pits or protruding points of the current battery cell.
In this embodiment, the surface texture features are expressed as texture parameter features of the current cell surface.
The beneficial effects of the technical scheme are as follows: the texture details of the current battery cells can be subjected to qualification detection in a finer manner, the detection precision is improved, the objectivity and the accuracy of the final detection result are ensured, further, the process quality evaluation report of the current battery cells is generated, the process evaluation can be performed on each current battery cell to perform subsequent self-adaptive improvement, the production quality and the qualification degree of the subsequent battery cells are ensured, and the excessive loss of the cost is reduced.
In one embodiment, the analyzing the surface texture features and determining whether the surface of the current cell has scratches and pits or protruding points or not is traumatic according to the analysis result includes:
constructing a linear correlation function between the sensory quality and the texture evaluation index, and determining the sensory parameters of the electric core entity under different texture evaluation index values according to the linear correlation function;
performing feature characterization on the surface texture features to obtain a first certainty result, and determining class display attributes corresponding to the surface texture features according to the first certainty result;
determining current sensory parameters for the surface texture features according to the class display attributes corresponding to the surface texture features;
Determining a feasible texture evaluation index item for the current battery cell based on the current sensory parameter and the battery cell entity sensory parameters under different texture evaluation index values;
determining texture display parameters of each pixel point according to the surface texture characteristics and pixel distribution of the high-definition image, and constructing a texture display parameter set according to the texture display parameters;
evaluating each texture display parameter in the texture display parameter set based on a feasible texture evaluation index item of the current battery cell to obtain an evaluation result;
and judging whether the surface of the current battery cell has scratches, pits or convex points and detail wounds according to the evaluation result.
In this embodiment, the linear correlation function is expressed as a linear correlation function in which the sensory quality varies with the number of texture evaluation indexes;
in this embodiment, the sensory parameters of the cell entity are expressed as sensory effect parameters of the cell image under different texture evaluation index values for human eyes, for example: ambiguity, etc.;
in this embodiment, the characterization table is a table that determines the display characteristics of the surface texture features;
in this embodiment, the class display attribute is expressed as a visual attribute under a characteristic display characteristic corresponding to the surface texture feature;
in this embodiment, the feasible texture evaluation index item is represented as a texture evaluation index item that can collect texture evaluation index parameters in the high-definition image of the current cell;
In the present embodiment, the texture display parameter is expressed as a texture display description parameter for each pixel in the high definition image;
in this embodiment, the evaluation result is expressed as evaluation content corresponding to the index qualification evaluation performed on the texture display parameter of each pixel.
The beneficial effects of the technical scheme are as follows: by determining feasible texture evaluation index items in the high-definition image of the current battery cell to evaluate the texture display parameters of each pixel point, the trauma judgment of the limiting index can be accurately carried out based on the actual sensory effect of the high-definition image, so that the judgment result is more practical and objective, the influence of useless evaluation indexes is avoided, and the judgment accuracy is improved.
In one embodiment, the determining whether the surface of the current cell has scratches and pits or bumps or not is performed according to the evaluation result, including:
screening abnormal texture display parameters with unqualified texture evaluation indexes according to the evaluation results, and determining parameter abnormal characteristics according to the distribution conditions;
performing detail trauma qualitative on the abnormal texture display parameters according to the abnormal characteristics of the parameters to obtain a second qualitative result;
determining the wound types corresponding to the abnormal texture display parameters according to the second qualitative results, and obtaining a standard display parameter set corresponding to each wound type;
Matching the current display parameter set corresponding to each abnormal texture display parameter with the standard display parameter set corresponding to the wound type of the abnormal texture display parameter, and determining whether the current display parameter set corresponding to each abnormal texture display parameter has a display wound or not according to a matching result;
if yes, determining the wound grade according to the matching interval of the current display parameter set in the standard display parameter set;
and judging whether the surface of the current battery cell has scratches, pits or salient points and detail wounds or not according to the wound grade and the wound type corresponding to each abnormal texture display parameter.
In this embodiment, the abnormal texture display parameter is expressed as a texture display parameter having a large deviation from the standard texture display parameter;
in this embodiment, the distribution condition is represented as the regional distribution of the abnormal texture display parameter in the high-definition image;
in the present embodiment, the parameter anomaly characteristic is expressed as a region shape characteristic of a region occupied by an abnormal texture display parameter in a high definition image, for example: a long-strip region or a circular region, etc.;
in this embodiment, the detail wound qualitative indication is that the abnormal texture display parameter is wound qualitative according to the area shape characteristic of the occupied area of the abnormal texture display parameter in the high-definition image, if the abnormal texture display parameter is a strip area, the abnormal texture display parameter is a scratch wound, and if the abnormal texture display parameter is a pit or a bump wound;
In the present embodiment, the wound type is expressed as a wound type corresponding to a qualitative result of the wound, for example: the wound type of the scratch wound is a range wound, the wound type of the concave point or the convex point wound is a single point wound;
in this embodiment, the wound level is expressed as a wound level corresponding to the abnormal texture display parameter.
The beneficial effects of the technical scheme are as follows: by performing detail trauma qualitative on the abnormal texture display parameters according to the abnormal characteristics of the parameters, whether the abnormal texture display parameters accord with the trauma conditions or not can be intuitively determined, and then accurate trauma structure and slight and serious trauma judgment are performed according to the trauma types and the display parameter sets, so that the judgment precision and accuracy are improved, and the comprehensiveness and reliability of appearance detection of the battery cell are further ensured.
The embodiment also discloses a system for detecting the appearance defects of the battery cells based on artificial intelligent image recognition, as shown in fig. 4, the system comprises:
the extraction module 401 is configured to obtain standard size information and standard appearance information of a target specification battery cell, and extract a battery cell identification feature according to the standard size information and the standard appearance information;
a construction module 402, configured to construct an identification model of the target specification cell and a plurality of identification strategies of the identification model according to the cell identification feature and a preset learning network model;
The input module 403 is configured to obtain a high-definition image of a current battery cell, determine a target recognition policy of a recognition model according to the high-definition image, and input the high-definition image into the recognition model;
and the determining module 404 is configured to determine each appearance index defect of the current cell according to the recognition result of the recognition model.
The working principle and the beneficial effects of the above technical solution are described in the method claims, and are not repeated here.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains.
It is to be understood that the present disclosure is not limited to the precise arrangements and instrumentalities shown in the drawings, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims (8)
1. The battery cell appearance defect detection method based on artificial intelligent image recognition is characterized by comprising the following steps of:
Standard size information and standard appearance information of the battery cell of the target specification are obtained, and battery cell identification characteristics are extracted according to the standard size information and the standard appearance information;
constructing an identification model of the battery cell with the target specification according to the battery cell identification characteristics and a preset learning network model, and a plurality of identification strategies of the identification model;
acquiring a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model;
determining each appearance index defect of the current battery cell according to the identification result of the identification model;
the construction of the identification model of the target specification battery cell and a plurality of identification strategies of the identification model according to the battery cell identification characteristics and a preset learning network model comprises the following steps:
acquiring a plurality of image samples of a target specification cell and dividing the image samples into a training set, a testing set and a verification set;
writing the cell identification characteristics into a preset learning network model to construct an identification model of the cell with the target specification, and respectively carrying out repeated training and testing on the identification model by utilizing a training set, a testing set and a verification set until the identification accuracy of the identification model is greater than or equal to a preset threshold value;
determining gray value intervals of the image with the visualization conditions and dividing all interval values into a plurality of gray levels;
Acquiring dominant pixel parameters and recessive pixel parameters of the cell image under each gray level, and setting an identification strategy of an identification model for the cell image of each gray level according to the dominant pixel parameters and the recessive pixel parameters;
obtaining a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model, wherein the method comprises the following steps:
judging whether the high-definition image of the current battery cell has necessary identification characteristics according to the display content of the high-definition image, if so, judging that the high-definition image is qualified, and if not, judging that the high-definition image is unqualified;
randomly extracting n pixel points from the high-definition image, determining the current gray level of each pixel point, and determining the current gray level of the high-definition image according to the current gray level;
determining a target recognition strategy of a recognition model according to the current gray level and starting a corresponding current model recognition mode;
the high-definition image is input into a recognition model based on the current model recognition mode.
2. The method for detecting the appearance defect of the battery cell based on the artificial intelligence image recognition according to claim 1, wherein the step of obtaining the standard size information and the standard appearance information of the battery cell of the target specification, and extracting the battery cell recognition feature according to the standard size information and the standard appearance information comprises the steps of:
Acquiring a standard finished product sample of a target specification battery cell, and acquiring finished product images of the standard finished product sample based on a plurality of shooting angles;
analyzing the finished product images of the shooting angles to determine standard size information, standard shape information, standard color information and standard packaging information of the battery cell of the target specification;
integrating the standard shape information, the standard color information and the standard packaging information of the target specification battery cell to obtain the standard appearance information of the target specification battery cell;
and extracting the cell size identification feature according to the standard size information, extracting the cell appearance identification feature according to the standard appearance information, and integrating the cell size identification feature and the cell appearance identification feature to generate the cell identification feature of the cell with the target specification.
3. The method for detecting the appearance defect of the battery cell based on the artificial intelligence image recognition according to claim 2, wherein the extracting the cell size recognition feature according to the standard size information and the extracting the cell appearance recognition feature according to the standard appearance information simultaneously comprises:
acquiring a size description parameter of a battery cell of a target specification according to the standard size information, and extracting a battery cell size identification characteristic based on the size description parameter;
Acquiring the color attribute, the packaging attribute and the shape attribute of the battery cell of the target specification according to the standard appearance information;
acquiring feature description factors corresponding to the color attribute, the packaging attribute and the shape attribute respectively, and performing entity definition on the feature description factors to acquire a definition result;
and obtaining the appearance global characteristics of the battery cell with the target specification according to the definition result, and obtaining the appearance identification characteristics of the battery cell according to the appearance global characteristics.
4. The method for detecting the appearance defects of the battery cells based on the artificial intelligence image recognition according to claim 1, wherein the determining the appearance index defects of the current battery cells according to the recognition result of the recognition model comprises the following steps:
determining a current size index, a current shape index and a current color index of the current battery cell according to the identification result of the identification model;
comparing the current size index, the current shape index and the current color index of the current battery cell with the standard size index, the standard shape index and the standard color index of the battery cell of the target specification to obtain a comparison result;
determining defect parameters of each appearance index according to the comparison result, and determining the defect grade of each appearance index according to the defect parameters;
each appearance index defect is classified into an acceptable appearance index defect and an unacceptable appearance index defect based on the defect level of each appearance index.
5. The method for detecting the appearance defects of the battery cells based on the artificial intelligence image recognition according to claim 1, wherein the method further comprises:
after determining that the appearance index of the current battery cell is defect-free, acquiring the surface texture characteristics of the current battery cell according to the high-definition image;
analyzing the surface texture characteristics, and judging whether scratches, pits or salient point detail wounds exist on the surface of the current battery cell according to the analysis result;
if yes, marking and amplifying the image area with scratches and pits or bumps in the high-definition image;
and generating a process quality evaluation report of the current battery cell according to the distribution condition of the surface scratches and pits or protruding points of the current battery cell.
6. The method for detecting the appearance defect of the battery cell based on the artificial intelligence image recognition according to claim 5, wherein the analyzing the surface texture features and judging whether scratches and pits or protruding points exist on the surface of the current battery cell according to the analysis result comprises the following steps:
constructing a linear correlation function between the sensory quality and the texture evaluation index, and determining the sensory parameters of the electric core entity under different texture evaluation index values according to the linear correlation function;
performing feature characterization on the surface texture features to obtain a first certainty result, and determining class display attributes corresponding to the surface texture features according to the first certainty result;
Determining current sensory parameters for the surface texture features according to the class display attributes corresponding to the surface texture features;
determining a feasible texture evaluation index item for the current battery cell based on the current sensory parameter and the battery cell entity sensory parameters under different texture evaluation index values;
determining texture display parameters of each pixel point according to the surface texture characteristics and pixel distribution of the high-definition image, and constructing a texture display parameter set according to the texture display parameters;
evaluating each texture display parameter in the texture display parameter set based on a feasible texture evaluation index item of the current battery cell to obtain an evaluation result;
and judging whether the surface of the current battery cell has scratches, pits or convex points and detail wounds according to the evaluation result.
7. The method for detecting the appearance defect of the battery cell based on the artificial intelligence image recognition according to claim 6, wherein the step of judging whether scratches and pits or protruding points are wound on the surface of the current battery cell according to the evaluation result comprises the following steps:
screening abnormal texture display parameters with unqualified texture evaluation indexes according to the evaluation results, and determining parameter abnormal characteristics according to the distribution conditions;
performing detail trauma qualitative on the abnormal texture display parameters according to the abnormal characteristics of the parameters to obtain a second qualitative result;
Determining the wound types corresponding to the abnormal texture display parameters according to the second qualitative results, and obtaining a standard display parameter set corresponding to each wound type;
matching the current display parameter set corresponding to each abnormal texture display parameter with the standard display parameter set corresponding to the wound type of the abnormal texture display parameter, and determining whether the current display parameter set corresponding to each abnormal texture display parameter has a display wound or not according to a matching result;
if yes, determining the wound grade according to the matching interval of the current display parameter set in the standard display parameter set;
and judging whether the surface of the current battery cell has scratches, pits or salient points and detail wounds or not according to the wound grade and the wound type corresponding to each abnormal texture display parameter.
8. An artificial intelligence image recognition-based cell appearance defect detection system is characterized in that the system comprises:
the extraction module is used for acquiring standard size information and standard appearance information of the battery cell with the target specification and extracting battery cell identification characteristics according to the standard size information and the standard appearance information;
the construction module is used for constructing an identification model of the battery cell with the target specification and a plurality of identification strategies of the identification model according to the battery cell identification characteristics and a preset learning network model;
The input module is used for acquiring a high-definition image of the current battery cell, determining a target recognition strategy of the recognition model according to the high-definition image and inputting the high-definition image into the recognition model;
the determining module is used for determining each appearance index defect of the current battery cell according to the identification result of the identification model;
the construction of the identification model of the target specification battery cell and a plurality of identification strategies of the identification model according to the battery cell identification characteristics and a preset learning network model comprises the following steps:
acquiring a plurality of image samples of a target specification cell and dividing the image samples into a training set, a testing set and a verification set;
writing the cell identification characteristics into a preset learning network model to construct an identification model of the cell with the target specification, and respectively carrying out repeated training and testing on the identification model by utilizing a training set, a testing set and a verification set until the identification accuracy of the identification model is greater than or equal to a preset threshold value;
determining gray value intervals of the image with the visualization conditions and dividing all interval values into a plurality of gray levels;
acquiring dominant pixel parameters and recessive pixel parameters of the cell image under each gray level, and setting an identification strategy of an identification model for the cell image of each gray level according to the dominant pixel parameters and the recessive pixel parameters;
Obtaining a high-definition image of a current battery cell, determining a target recognition strategy of a recognition model according to the high-definition image, and inputting the high-definition image into the recognition model, wherein the method comprises the following steps:
judging whether the high-definition image of the current battery cell has necessary identification characteristics according to the display content of the high-definition image, if so, judging that the high-definition image is qualified, and if not, judging that the high-definition image is unqualified;
randomly extracting n pixel points from the high-definition image, determining the current gray level of each pixel point, and determining the current gray level of the high-definition image according to the current gray level;
determining a target recognition strategy of a recognition model according to the current gray level and starting a corresponding current model recognition mode;
the high-definition image is input into a recognition model based on the current model recognition mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310553482.2A CN116309564B (en) | 2023-05-17 | 2023-05-17 | Method and system for detecting appearance defects of battery cells based on artificial intelligent image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310553482.2A CN116309564B (en) | 2023-05-17 | 2023-05-17 | Method and system for detecting appearance defects of battery cells based on artificial intelligent image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116309564A CN116309564A (en) | 2023-06-23 |
| CN116309564B true CN116309564B (en) | 2023-08-11 |
Family
ID=86817119
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310553482.2A Active CN116309564B (en) | 2023-05-17 | 2023-05-17 | Method and system for detecting appearance defects of battery cells based on artificial intelligent image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116309564B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116805311B (en) * | 2023-08-18 | 2023-11-07 | 长春师范大学 | Automobile part surface defect monitoring method based on robot vision |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353413A (en) * | 2020-02-25 | 2020-06-30 | 武汉大学 | Low-missing-report-rate defect identification method for power transmission equipment |
| CN113744270A (en) * | 2021-11-05 | 2021-12-03 | 南京市特种设备安全监督检验研究院 | Unmanned aerial vehicle visual detection and identification method for crane complex steel structure surface defects |
| CN114283111A (en) * | 2020-09-17 | 2022-04-05 | 赢创运营有限公司 | Qualitative or quantitative characterization of the surface of a coating |
| CN114419007A (en) * | 2022-01-21 | 2022-04-29 | 西安交通大学 | Defect type identification method and system of multi-strategy fusion deep learning network model |
| CN114627360A (en) * | 2020-12-14 | 2022-06-14 | 国电南瑞科技股份有限公司 | Substation equipment defect identification method based on cascade detection model |
| CN115078392A (en) * | 2022-07-29 | 2022-09-20 | 深圳市泽信智能装备有限公司 | Semiconductor chip defect detection equipment and method |
| CN115829912A (en) * | 2022-07-29 | 2023-03-21 | 宁德时代新能源科技股份有限公司 | Method and device for detecting surface defects of battery cell |
| CN115937079A (en) * | 2022-09-28 | 2023-04-07 | 贵州电网有限责任公司 | YOLO v 3-based rapid detection method for defects of power transmission line |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220318667A1 (en) * | 2021-03-30 | 2022-10-06 | Accenture Global Solutions Limited | Intelligent real-time defect prediction, detection, and ai driven automated correction solution |
| US20220318672A1 (en) * | 2021-04-01 | 2022-10-06 | Samsung Display Co., Ltd. | Systems and methods for identifying manufacturing defects |
-
2023
- 2023-05-17 CN CN202310553482.2A patent/CN116309564B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353413A (en) * | 2020-02-25 | 2020-06-30 | 武汉大学 | Low-missing-report-rate defect identification method for power transmission equipment |
| CN114283111A (en) * | 2020-09-17 | 2022-04-05 | 赢创运营有限公司 | Qualitative or quantitative characterization of the surface of a coating |
| CN114627360A (en) * | 2020-12-14 | 2022-06-14 | 国电南瑞科技股份有限公司 | Substation equipment defect identification method based on cascade detection model |
| CN113744270A (en) * | 2021-11-05 | 2021-12-03 | 南京市特种设备安全监督检验研究院 | Unmanned aerial vehicle visual detection and identification method for crane complex steel structure surface defects |
| CN114419007A (en) * | 2022-01-21 | 2022-04-29 | 西安交通大学 | Defect type identification method and system of multi-strategy fusion deep learning network model |
| CN115078392A (en) * | 2022-07-29 | 2022-09-20 | 深圳市泽信智能装备有限公司 | Semiconductor chip defect detection equipment and method |
| CN115829912A (en) * | 2022-07-29 | 2023-03-21 | 宁德时代新能源科技股份有限公司 | Method and device for detecting surface defects of battery cell |
| CN115937079A (en) * | 2022-09-28 | 2023-04-07 | 贵州电网有限责任公司 | YOLO v 3-based rapid detection method for defects of power transmission line |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116309564A (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114549522B (en) | Textile quality detection method based on target detection | |
| WO2021143343A1 (en) | Method and device for testing product quality | |
| CN103034838B (en) | A kind of special vehicle instrument type identification based on characteristics of image and scaling method | |
| CN109978822B (en) | Banana maturity judging modeling method and judging method based on machine vision | |
| CN114419025A (en) | Fiberboard quality evaluation method based on image processing | |
| US10169861B2 (en) | Image processing apparatus, non-transitory computer readable medium, and image processing method | |
| CN108090902A (en) | A kind of non-reference picture assessment method for encoding quality based on multiple dimensioned generation confrontation network | |
| CN116309564B (en) | Method and system for detecting appearance defects of battery cells based on artificial intelligent image recognition | |
| CN116645367B (en) | Steel plate cutting quality detection method for high-end manufacturing | |
| CN109215009B (en) | Continuous casting billet surface image defect detection method based on deep convolution neural network | |
| CN111401419A (en) | Improved RetinaNet-based employee dressing specification detection method | |
| CN116188475B (en) | Intelligent control method, system and medium for automatic optical detection of appearance defects | |
| CN105335692A (en) | Tire X-ray image detection and identification method and system | |
| CN107564000B (en) | Hericium erinaceus non-destructive testing stage division based on computer vision | |
| CN111242899B (en) | Image-based flaw detection method and computer-readable storage medium | |
| CN112613454A (en) | Electric power infrastructure construction site violation identification method and system | |
| CN111415339A (en) | Image defect detection method for complex texture industrial product | |
| CN111312023A (en) | Device and method for automatically drawing middle school physics circuit experiment circuit diagram | |
| CN113295690A (en) | Strawberry maturity rapid discrimination method based on machine learning | |
| CN111626358A (en) | Tunnel surrounding rock grading method based on BIM picture recognition | |
| CN109102486B (en) | Surface defect detection method and device based on machine learning | |
| CN116416523A (en) | Machine learning-based rice growth stage identification system and method | |
| CN114820597B (en) | Smelting product defect detection method, device and system based on artificial intelligence | |
| KR101782366B1 (en) | Vision inspection method based on learning data using normalization of sample scale | |
| CN115330721A (en) | Banana fruit comb plumpness detection method and system based on shape and color information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |