CN115082483A - Glass fiber board surface defect identification method based on optical camera - Google Patents
Glass fiber board surface defect identification method based on optical camera Download PDFInfo
- Publication number
- CN115082483A CN115082483A CN202211009352.4A CN202211009352A CN115082483A CN 115082483 A CN115082483 A CN 115082483A CN 202211009352 A CN202211009352 A CN 202211009352A CN 115082483 A CN115082483 A CN 115082483A
- Authority
- CN
- China
- Prior art keywords
- image
- defect
- glass fiber
- pixel points
- value
- 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.)
- Granted
Links
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- 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)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of defect detection, in particular to a glass fiber plate data identification method based on an optical camera. Acquiring a grey image of a glass fiber board by using an optical camera, and acquiring a dictionary matrix of the grey image to change to obtain a plurality of column change dictionary matrices; acquiring a plurality of reconstructed images, and acquiring a difference image between the gray level image of the glass fiber plate and each reconstructed image; calculating the distribution range of each difference image pixel point; acquiring defect characteristic images with distribution ranges smaller than a threshold value, setting weight for each pixel point, and calculating the distribution of the pixel points; acquiring the average distribution density of all pixel points in each defect characteristic image; calculating the defect degree of each defect characteristic image, and acquiring a defect characteristic heat map corresponding to each defect characteristic image; and superposing all the defect characteristic heat maps to obtain a defect heat map of the glass fiber board, and identifying the defect area in the glass fiber board. According to the invention, the defect identification effect is improved by analyzing the optical image.
Description
Technical Field
The invention relates to the field of defect detection, in particular to a method for identifying surface defects of a glass fiber board based on an optical camera.
Background
The glass fiber partition plate is provided with inherent edge information and defect edge information, a conventional defect identification method is interfered by the inherent edge information, the defect identification effect cannot be well achieved, the neural network is used for identifying the defects, the quality is further evaluated, the neural network needs a large amount of training data and a labeling process, and cost is high.
Based on the method, the defect characteristics are extracted through the K-SVD technology, so that defect information can be identified through a conventional defect identification method under the condition that an inherent edge exists, and a good identification effect is obtained.
Disclosure of Invention
The invention provides a method for identifying surface defects of a glass fiber board based on an optical camera, which aims to solve the existing problems and comprises the following steps:
collecting a gray level image of the glass fiber board, and obtaining a dictionary matrix of the gray level image for changing to obtain a plurality of column change dictionary matrices; acquiring a plurality of reconstructed images, and acquiring a difference image between the gray level image of the glass fiber plate and each reconstructed image; calculating the distribution range of each difference image pixel point; acquiring defect characteristic images with distribution ranges smaller than a threshold value, setting weight for each pixel point, and calculating the distribution of the pixel points; acquiring the average distribution density of all pixel points in each defect characteristic image; calculating the defect degree of each defect characteristic image, and acquiring a defect characteristic heat map corresponding to each defect characteristic image; and superposing all the defect characteristic heat maps to obtain a defect heat map of the glass fiber board, and identifying the defect area in the glass fiber board.
According to the technical means provided by the invention, the dictionary matrix of the glass fiber board is obtained through the K-SVD technology, the dictionary matrix is subjected to column change, the difference image of the glass fiber board is further obtained according to the column change dictionary matrix for defect feature analysis, the pixel points in each defect feature image are subjected to feature calculation, so that the defect heat map representing the defect features in each defect feature image is obtained, all the defect heat maps are superposed, the feature enhancement of the defect region in the glass fiber board image is realized, the defect region in the glass fiber board can be well identified, and the identification accuracy and the identification efficiency are improved.
The invention adopts the following technical scheme that a glass fiber plate surface defect identification method based on an optical camera comprises the following steps:
collecting a gray level image of a glass fiber board, obtaining a dictionary matrix and a sparse coding matrix of the gray level image by utilizing K-SVD, and sequentially changing each column vector in the dictionary matrix of the gray level image to obtain a plurality of column change dictionary matrices.
And acquiring a plurality of corresponding reconstructed images according to the plurality of column change dictionary matrixes and the sparse coding matrix of the gray image, and acquiring a difference image between the gray image of the glass fiber plate and each reconstructed image.
Calculating the distribution range of each difference image pixel point according to the area of the minimum external rectangle of all pixel points in each difference image and the area of the grey image of the glass fiber plate; and acquiring all difference images with the distribution range smaller than the threshold value as defect characteristic images.
And calculating the distribution of the pixel points in each defect characteristic image according to the gray value of the pixel points in each defect characteristic image.
Clustering the pixel points in each defect characteristic image, calculating the distribution density of the pixel points in each category according to the number of the pixel points in each category, and obtaining the average distribution density of the pixel points in each defect characteristic image.
And calculating the defect degree of each defect characteristic image according to the distribution and the average distribution density of the pixel points in each defect characteristic image, and acquiring a defect characteristic heat map corresponding to each defect characteristic image according to the defect degree of each defect characteristic image.
And superposing all the defect characteristic heat maps to obtain a defect heat map of the glass fiber board.
Furthermore, the method for identifying the surface defects of the glass fiber board based on the optical camera sequentially replaces each column vector in the dictionary matrix of the gray level image with a 0 vector to obtain a plurality of column change dictionary matrices.
Further, a method for identifying surface defects of a glass fiber board based on an optical camera, which obtains a plurality of corresponding reconstructed images according to a plurality of column change dictionary matrixes, comprises the following steps:
decomposing the glass fiber board gray level image into a product of a dictionary matrix and a sparse coding matrix, obtaining a plurality of column change dictionary matrices, and multiplying each column change dictionary matrix with the sparse coding matrix of the glass fiber gray level image to obtain a plurality of corresponding reconstructed images.
Further, the method for identifying the defects of the surface of the glass fiber plate based on the optical camera is used as the distribution range of the pixel points of each difference image according to the ratio of the area of the minimum circumscribed rectangle of all the pixel points in each difference image to the area of the gray image of the glass fiber plate.
Further, a method for identifying the surface defects of the glass fiber plate based on the optical camera comprises the following steps of:
performing threshold segmentation on each defect characteristic image, and acquiring a gray average value of all pixel points with gray values larger than a threshold value as a numerical value 1; acquiring a gray average value of all pixel points with gray values smaller than a threshold value as a numerical value 2;
calculating the sum of the value 1 and the value 2 to obtain a value 3, and taking the ratio of the value 1 to the value 3 as the weight values of all pixel points of which the gray value is smaller than the threshold;
taking the ratio of the value 2 to the value 3 as the weighted value of all pixel points with the gray value larger than the threshold value;
and carrying out weighted summation according to the weighted values of all pixel points in each defect characteristic image to obtain the distributivity of the pixel points in the corresponding defect characteristic image.
Further, a method for identifying the surface defect of the glass fiber board based on the optical camera, which is used for obtaining the average distribution density of all pixel points in each defect characteristic image, comprises the following steps:
obtaining an updating circle of the category by taking the center point of each category as a circle center and the maximum Euclidean distance from the circle center of the category to the pixel points in other categories as a radius, and obtaining the distribution density of the pixel points of each category according to the ratio of the area of the updating circle to the number of the pixel points in the category in the updating circle;
and calculating the distribution density average value of all the types of pixel points in each defect characteristic image to obtain the average distribution density of the corresponding defect characteristic image.
Further, the method for identifying the surface defects of the glass fiber board based on the optical camera is characterized in that the method for acquiring the defect characteristic heat map corresponding to each defect characteristic image comprises the following steps:
and calculating the cosine similarity mean value of each column vector and all other column vectors in the dictionary matrix of the glass fiber board as a correlation value, and obtaining a defect heat map corresponding to each defect characteristic image according to the product of the defect degree of each defect characteristic image and the correlation value.
The invention has the beneficial effects that: according to the technical means provided by the invention, the dictionary matrix of the glass fiber board is obtained through the K-SVD technology, the dictionary matrix is subjected to column change, the difference image of the glass fiber board is further obtained according to the column change dictionary matrix for defect feature analysis, the pixel points in each defect feature image are subjected to feature calculation, so that the defect heat map representing the defect features in each defect feature image is obtained, all the defect heat maps are superposed, the identification of the defect area in the glass fiber board image is realized, and the identification accuracy and the identification efficiency are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a schematic structural diagram of a method for identifying surface defects of a glass fiber board based on an optical camera according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a method for identifying surface defects of a glass fiber board based on an optical camera according to an embodiment of the present invention includes:
101. the method comprises the steps of collecting a grey image of a glass fiber board, obtaining a dictionary matrix of the grey image, and sequentially changing each column vector in the dictionary matrix of the grey image to obtain a plurality of column change dictionary matrices.
Firstly, data are acquired through equipment, and then semantic segmentation is carried out on image data.
The present invention identifies objects in segmented images by means of DNN semantic segmentation.
The relevant content of the DNN network is as follows:
the data set used is a fibreglass panel image data set acquired from a top view, the fibreglass panel being of a variety of formats.
The pixels needing to be segmented are divided into two types, namely the labeling process of the corresponding labels of the training set is as follows: and in the semantic label of the single channel, the pixel at the corresponding position belongs to the background class and is marked as 0, and the pixel belonging to the glass fiber board is marked as 1.
The task of the network is to classify, and all the used loss functions are cross entropy loss functions.
The 0-1 mask image obtained by semantic segmentation is multiplied by the original image, and the obtained image only contains the image of the glass fiber board, so that the interference of the background is removed.
And obtaining a dictionary matrix of the glass fiber plate image through a K-SVD algorithm, wherein the K-SVD is a dictionary representation method and decomposes the original image matrix into a product of the dictionary matrix and a sparse coding matrix.
The dictionary matrix represents the characteristics of the image, the sparse coding matrix represents that the information in the image is obtained by a combination method of different characteristics, the three-dimensional coordinate system (X, Y and Z unit vectors) can be used for representing, the three vectors can represent all three-dimensional vectors which are dictionaries through combination, and the three vectors can be restored into the three-dimensional vectors by matching with an X, Y and Z combination mode (sparse coding) of each three-dimensional vector.
In the process of obtaining the dictionary matrix by the K-SVD decomposition, the dictionary matrix is updated column by column, so that the column vectors of the dictionary matrix can be used for representing the common features between the image line data.
And representing the column number of the dictionary matrix by using n, and sequentially replacing n column vectors of the dictionary matrix by using corresponding 0 vectors to obtain n modified dictionary matrices to obtain a plurality of column modified dictionary matrices.
102. And acquiring a plurality of corresponding reconstructed images according to the column change dictionary matrixes, and acquiring a difference image between the gray image of the glass fiber plate and each reconstructed image.
The method for acquiring the corresponding multiple reconstructed images according to the multiple column change dictionary matrixes comprises the following steps:
decomposing the glass fiber board gray level image into a product of a dictionary matrix and a sparse coding matrix, obtaining a plurality of column change dictionary matrices, and multiplying each column change dictionary matrix with the sparse coding matrix of the glass fiber gray level image to obtain a plurality of corresponding reconstructed images.
And multiplying the n column change dictionary matrixes by the sparse coding matrix of the original glass fiber plate gray image to obtain n reconstructed images, and calculating a difference image of the original glass fiber plate gray image and the reconstructed image to obtain n difference images.
Each reconstructed image corresponds to an image obtained by removing the features represented by the corresponding column vector from the source image, so that the difference image can represent the features represented by the corresponding column vector.
103. Calculating the distribution range of each difference image pixel point according to the area of the minimum external rectangle of all pixel points in each difference image and the area of the grey image of the glass fiber plate; and acquiring all difference images with the distribution range smaller than the threshold value as defect characteristic images.
And taking the ratio of the area of the minimum circumscribed rectangle of all the pixel points in each difference image to the area of the gray image of the glass fiber board as the distribution range p3 of the pixel points of each difference image, and taking the difference image with the p3 smaller than a preset first threshold (the value is 0.8 in the invention) as a defect feature image.
104. Setting weight for each pixel point according to the gray value of the pixel point in each defect characteristic image, and calculating the distribution of the pixel points in each defect characteristic image according to the weight of each pixel point in each defect characteristic image.
For the fiber board, the defects are often distributed locally, while the inherent texture of the fiber board is distributed globally, so that the probability that the difference image is the defect feature can be obtained by calculating the distribution of different difference image features.
The distributivity refers to that the features are local and global, the greater the distributivity, the closer the features are to the global features, and the smaller the distributivity, the closer the features are to the local features.
The more the number of non-zero-value pixel points in the difference image is, the larger the distribution is, the more the pixel points with smaller gray values are represented as defects, and the more serious the defects are, so that corresponding importance weights are given to the pixel points with different gray values. In a gray image, the corresponding weighted value with a small pixel point should be larger.
Therefore, the method for setting the weight according to the gray value of the pixel point in each defect characteristic image comprises the following steps:
performing threshold segmentation on each defect characteristic image, and acquiring a gray average value of all pixel points with gray values larger than a threshold value as a numerical value 1; acquiring a gray average value of all pixel points with gray values smaller than a threshold value as a numerical value 2;
calculating the sum of the value 1 and the value 2 to obtain a value 3, and taking the ratio of the value 1 to the value 3 as the weighted value 1 of all pixel points of which the gray value is smaller than the threshold;
and taking the ratio of the value 2 to the value 3 as the weighted value 2 of all pixel points of which the gray value is greater than the threshold value.
And weighting and summing the weight value 1 and the weight value 2 of the pixel point to obtain the distributivity of the pixel point in the corresponding defect characteristic image.
The more the distribution of the difference image tends to be global, the greater the probability that the characteristic represented by the difference image is an inherent texture characteristic; the more the distribution of the difference image tends to be local, the greater the probability that the difference image represents a feature that is a defective feature.
105. Clustering the pixel points in each defect characteristic image, calculating the distribution density of the pixel points in each category according to the number of the pixel points in each category, and acquiring the average distribution density of all the pixel points in each defect characteristic image.
The invention clusters pixel points at different positions by a mean shift clustering method, then obtains an updating circle by taking the center point of each category as the center of a circle and the Euclidean distance from the center point of the category with the largest distance from the center of the circle as the radius, and obtains the distribution density of each category by calculating the updating circle.
The specific method for obtaining the average distribution density of all pixel points in each defect characteristic image comprises the following steps:
obtaining an updating circle of the category by taking the center point of each category as a circle center and the maximum Euclidean distance from the circle center of the category to the pixel points in other categories as a radius, and obtaining the distribution density of the pixel points of each category according to the ratio of the area of the updating circle to the number of the pixel points in the category in the updating circle;
and calculating the distribution density average value of all the types of pixel points in each defect characteristic image to obtain the average distribution density of the corresponding defect characteristic image.
In the invention, when the distribution densities of different defect characteristic images are calculated, the clustering radiuses are kept consistent.
106. And calculating the defect degree of each defect characteristic image according to the distribution and the average distribution density of the pixel points in each defect characteristic image, and acquiring a defect characteristic heat map corresponding to each defect characteristic image according to the defect degree of each defect characteristic image.
The method for acquiring the defect characteristic heat map corresponding to each defect characteristic image comprises the following steps:
and calculating the cosine similarity mean value of each column vector and all other column vectors in the dictionary matrix of the glass fiber board as a correlation value, and obtaining a defect heat map corresponding to each defect characteristic image according to the product of the defect degree of each defect characteristic image and the correlation value.
107. And superposing all the defect characteristic heat maps to obtain a defect heat map of the glass fiber board, and identifying the defect area in the glass fiber board according to the defect heat map of the glass fiber board.
The defect heat map indicates the severity of the defect, and identifying more severe defects requires more enhancement to make the defect signature apparent and easily distinguishable.
According to the method, the obtained defect characteristic heat maps are superposed, so that the defect characteristics in the glass fiber board are enhanced, the defect heat map of the glass fiber board is obtained, and the defect area in the glass fiber board is identified according to the defect heat map, so that the accurate defect area is obtained.
According to the technical means provided by the invention, the dictionary matrix of the glass fiber board is obtained through the K-SVD technology, the dictionary matrix is subjected to column change, the difference image of the glass fiber board is further obtained according to the column change dictionary matrix for defect feature analysis, the pixel points in each defect feature image are subjected to feature calculation, so that the defect heat map representing the defect features in each defect feature image is obtained, all the defect heat maps are superposed, the feature enhancement of the defect region in the glass fiber board image is realized, the defect region in the glass fiber board can be well identified, and the identification accuracy and the identification efficiency are improved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and should not be taken as limiting the scope of the present invention, which is intended to cover any modifications, equivalents, improvements, etc. within the spirit and scope of the present invention.
Claims (7)
1. A method for identifying surface defects of a glass fiber board based on an optical camera is characterized by comprising the following steps:
collecting a grey-scale image of a glass fiber board, acquiring a dictionary matrix and a sparse coding matrix of the grey-scale image by utilizing K-SVD (K-singular value decomposition), and sequentially changing each column vector in the dictionary matrix of the grey-scale image to obtain a plurality of column change dictionary matrices;
acquiring a plurality of corresponding reconstructed images according to the plurality of column change dictionary matrixes and the sparse coding matrix of the gray image, and acquiring a difference image between the gray image of the glass fiber plate and each reconstructed image;
calculating the distribution range of each difference image pixel point according to the area of the minimum external rectangle of all pixel points in each difference image and the area of the grey image of the glass fiber plate; acquiring all difference images with distribution ranges smaller than a threshold value as defect characteristic images;
calculating the distribution of pixel points in each defect characteristic image according to the gray value of the pixel points in each defect characteristic image;
clustering the pixel points in each defect characteristic image, calculating the distribution density of the pixel points in each category according to the number of the pixel points in each category, and acquiring the average distribution density of the pixel points in each defect characteristic image;
calculating the defect degree of each defect characteristic image according to the distribution and the average distribution density of the pixel points in each defect characteristic image, and acquiring a defect characteristic heat map corresponding to each defect characteristic image according to the defect degree of each defect characteristic image;
and superposing all the defect characteristic heat maps to obtain the defect heat map of the glass fiber board.
2. The method for identifying the surface defects of the glass fiber board based on the optical camera as claimed in claim 1, wherein each column vector in the dictionary matrix of the gray-scale image is sequentially replaced by a 0 vector, so as to obtain a plurality of column change dictionary matrices.
3. The method for identifying the surface defects of the glass fiber board based on the optical camera as claimed in claim 1, wherein the method for obtaining the corresponding plurality of reconstructed images according to the plurality of column change dictionary matrixes comprises the following steps:
decomposing the glass fiber board gray level image into a product of a dictionary matrix and a sparse coding matrix, obtaining a plurality of column change dictionary matrices, and multiplying each column change dictionary matrix by the sparse coding matrix of the glass fiber gray level image to obtain a plurality of corresponding reconstructed images.
4. The method for identifying the surface defects of the glass fiber board based on the optical camera as claimed in claim 1, wherein the ratio of the area of the minimum circumscribed rectangle of all the pixel points in each difference image to the area of the gray image of the glass fiber board is used as the distribution range of the pixel points of each difference image.
5. The method for identifying the surface defects of the glass fiber board based on the optical camera as claimed in claim 1, wherein the method for calculating the distribution of the pixel points in each defect feature image comprises the following steps:
performing threshold segmentation on each defect characteristic image, and acquiring a gray average value of all pixel points with gray values larger than a threshold value as a numerical value 1; acquiring a gray average value of all pixel points with gray values smaller than a threshold value as a numerical value 2;
calculating the sum of the value 1 and the value 2 to obtain a value 3, and taking the ratio of the value 1 to the value 3 as the weight values of all pixel points of which the gray value is smaller than the threshold;
taking the ratio of the value 2 to the value 3 as the weighted value of all pixel points with the gray value larger than the threshold value;
and carrying out weighted summation according to the weight values of all pixel points in each defect characteristic image to obtain the distributivity of the pixel points in the corresponding defect characteristic image.
6. The method for identifying the surface defects of the glass fiber plate based on the optical camera as claimed in claim 1, wherein the method for obtaining the average distribution density of all pixel points in each defect feature image comprises the following steps:
obtaining an updating circle of the category by taking the center point of each category as a circle center and the maximum Euclidean distance from the circle center of the category to the pixel points in other categories as a radius, and obtaining the distribution density of the pixel points of each category according to the ratio of the area of the updating circle to the number of the pixel points in the category in the updating circle;
and calculating the mean distribution density of all the pixel points in each defect characteristic image to obtain the mean distribution density of the corresponding defect characteristic image.
7. The method for identifying the surface defects of the glass fiber board based on the optical camera as claimed in claim 1, wherein the method for acquiring the defect feature heat map corresponding to each defect feature image comprises the following steps:
and calculating the cosine similarity mean value of each column vector and all other column vectors in the dictionary matrix of the glass fiber board as a correlation value, and obtaining a defect heat map corresponding to each defect characteristic image according to the product of the defect degree of each defect characteristic image and the correlation value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211009352.4A CN115082483B (en) | 2022-08-23 | 2022-08-23 | Glass fiber board surface defect identification method based on optical camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211009352.4A CN115082483B (en) | 2022-08-23 | 2022-08-23 | Glass fiber board surface defect identification method based on optical camera |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115082483A true CN115082483A (en) | 2022-09-20 |
| CN115082483B CN115082483B (en) | 2022-12-02 |
Family
ID=83244414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211009352.4A Active CN115082483B (en) | 2022-08-23 | 2022-08-23 | Glass fiber board surface defect identification method based on optical camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115082483B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115311284A (en) * | 2022-10-12 | 2022-11-08 | 南通嘉鹏家居用品有限公司 | Production quality detection method for glass fiber mat |
| CN115330766A (en) * | 2022-10-12 | 2022-11-11 | 南通新诚电子有限公司 | Method for identifying formation quality defects of capacitor anode aluminum foil |
| CN117152444A (en) * | 2023-10-30 | 2023-12-01 | 山东泰普锂业科技有限公司 | Equipment data acquisition method and system for lithium battery industry |
| CN117218117A (en) * | 2023-11-07 | 2023-12-12 | 常熟市东宇绝缘复合材料有限公司 | Glass fiber yarn detection method and system |
-
2022
- 2022-08-23 CN CN202211009352.4A patent/CN115082483B/en active Active
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115311284A (en) * | 2022-10-12 | 2022-11-08 | 南通嘉鹏家居用品有限公司 | Production quality detection method for glass fiber mat |
| CN115330766A (en) * | 2022-10-12 | 2022-11-11 | 南通新诚电子有限公司 | Method for identifying formation quality defects of capacitor anode aluminum foil |
| CN115330766B (en) * | 2022-10-12 | 2023-04-11 | 南通新诚电子有限公司 | Method for identifying formation quality defects of capacitor anode aluminum foil |
| CN117152444A (en) * | 2023-10-30 | 2023-12-01 | 山东泰普锂业科技有限公司 | Equipment data acquisition method and system for lithium battery industry |
| CN117152444B (en) * | 2023-10-30 | 2024-01-26 | 山东泰普锂业科技有限公司 | Equipment data acquisition method and system for lithium battery industry |
| CN117218117A (en) * | 2023-11-07 | 2023-12-12 | 常熟市东宇绝缘复合材料有限公司 | Glass fiber yarn detection method and system |
| CN117218117B (en) * | 2023-11-07 | 2024-01-26 | 常熟市东宇绝缘复合材料有限公司 | Glass fiber yarn detection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115082483B (en) | 2022-12-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115082483B (en) | Glass fiber board surface defect identification method based on optical camera | |
| CN111062282B (en) | Substation pointer instrument identification method based on improved YOLOV3 model | |
| WO2023083059A1 (en) | Road surface defect detection method and apparatus, and electronic device and readable storage medium | |
| CN113160192A (en) | Visual sense-based snow pressing vehicle appearance defect detection method and device under complex background | |
| CN108932518B (en) | Shoe print image feature extraction and retrieval method based on visual bag-of-words model | |
| CN110781882A (en) | License plate positioning and identifying method based on YOLO model | |
| CN109145964B (en) | Method and system for realizing image color clustering | |
| CN112307919B (en) | Improved YOLOv 3-based digital information area identification method in document image | |
| US20200285917A1 (en) | Image classification method, computer device and medium | |
| CN113888536B (en) | Printed matter double image detection method and system based on computer vision | |
| CN114511718B (en) | Intelligent management method and system for materials for building construction | |
| CN110852327A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN115527072A (en) | Chip surface defect detection method based on sparse space perception and meta-learning | |
| CN111274964B (en) | Detection method for analyzing water surface pollutants based on visual saliency of unmanned aerial vehicle | |
| CN116311201A (en) | Substation equipment state identification method and system based on image identification technology | |
| CN115082781A (en) | Ship image detection method and device and storage medium | |
| CN113421223B (en) | Industrial product surface defect detection method based on deep learning and Gaussian mixture | |
| CN113344047A (en) | Platen state identification method based on improved K-means algorithm | |
| CN113223037A (en) | Unsupervised semantic segmentation method and unsupervised semantic segmentation system for large-scale data | |
| CN114065798A (en) | Visual identification method and device based on machine identification | |
| CN116993985A (en) | Method for realizing Zero-Shot automatic cutting of safety belt based on CLIP | |
| CN115272778A (en) | Recyclable garbage classification method and system based on RPA and computer vision | |
| CN112884022B (en) | Unsupervised depth characterization learning method and system based on image translation | |
| CN115100406A (en) | Weight information entropy fuzzy C-means clustering method based on superpixel processing | |
| CN113177602A (en) | Image classification method and device, electronic equipment and storage medium |
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 |