CN114820625A - Automobile top block defect detection method - Google Patents
Automobile top block defect detection method Download PDFInfo
- Publication number
- CN114820625A CN114820625A CN202210754325.3A CN202210754325A CN114820625A CN 114820625 A CN114820625 A CN 114820625A CN 202210754325 A CN202210754325 A CN 202210754325A CN 114820625 A CN114820625 A CN 114820625A
- Authority
- CN
- China
- Prior art keywords
- threshold
- processed
- defect
- obtaining
- region
- 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
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- 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
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/772—Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of artificial intelligence, and provides a method for detecting defects of an automobile top block, which is a method based on artificial intelligence in the production field and comprises the following steps: acquiring a grey-scale image of the car roof leather; obtaining a white spot pixel point set; obtaining a candidate defect area; obtaining a binary image after each threshold value is segmented; extracting a region to be processed; obtaining a dictionary matrix of each region to be processed; obtaining the change rate of the comprehensive dictionary of each region to be processed; resulting in the final defect area. The invention can divide accurate defect areas and has high detection precision.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a method for detecting defects of an automobile top block.
Background
Along with the development of social economy and the improvement of the living standard of people, the usage amount of automobiles is also improved. The quality of the automobile roof in the automobile is an important part of the appearance of the automobile, and the defects of the automobile roof directly influence the sale of the automobile. Meanwhile, the automobile roof is easy to scratch when being installed, so that the quality of the roof area needs to be detected before the automobile is on the market, and texture structures exist in the leather area of the automobile roof and can interfere with the identification of the scratch defect.
The traditional defect identification method comprises threshold segmentation and edge detection. However, the color difference between the car roof leather and the scratch defect is small, and the accurate defect area is difficult to divide by directly utilizing the threshold value division. Meanwhile, the car roof leather has a unique texture structure, the texture can interfere with defect identification, and it is difficult to respectively determine which areas are the defects and which areas are the leather textures by using an edge detection method. Meanwhile, certain texture intervals exist among scratch defects, and accurate defect regions are difficult to segment by directly utilizing the area information of the connected region. The invention provides a method for detecting defects of an automobile top block, which is a detection method based on artificial intelligence in the production field, and is characterized in that the characteristics that scratch defects have direction consistency characteristics and the continuity of the scratch defects is increased under different thresholds are considered, an initial threshold is firstly utilized to segment out a white spot region, then cluster analysis is carried out based on the direction consistency characteristics to obtain a candidate defect region, and then a dictionary change rate different from the threshold is combined to obtain a defect probability, so that an accurate defect region is segmented out.
Disclosure of Invention
The invention provides a method for detecting defects of an automobile top block, which aims to solve the problem of low detection precision in the prior art.
The invention discloses a method for detecting defects of an automobile top block, which adopts the following technical scheme:
acquiring a grey-scale image of the car roof leather;
performing threshold segmentation on the grey-scale image of the car roof leather through a first segmentation threshold of pixel points to obtain a white spot pixel point set which is larger than the segmentation threshold;
obtaining a principal component direction of the white spot pixel point set by adopting a PCA algorithm, obtaining an included angle of each white spot pixel point by utilizing a gradient direction angle and a principal component direction of each white spot pixel point in the white spot pixel point set, and clustering all white spot pixel points through the included angle to obtain a candidate defect region;
respectively carrying out threshold segmentation on the grey-scale images of the car roof leather through different thresholds to obtain binary images after each threshold segmentation;
extracting a region corresponding to the candidate defect region in the binarized image after each threshold segmentation as a region to be processed;
obtaining a dictionary matrix of each to-be-processed area under the corresponding binary image through the pixel information of each to-be-processed area in the corresponding binary image;
obtaining the comprehensive dictionary change rate of each to-be-processed area through the dictionary matrix of each to-be-processed area under the binary image after all threshold segmentation;
and obtaining the defect probability of each to-be-processed area through the average value of the included angles of all pixel points in the candidate defect area corresponding to each to-be-processed area and the change rate of the comprehensive dictionary of the to-be-processed area, and judging whether each to-be-processed area is a defect area according to the defect probability to obtain a final defect area.
Further, the method for detecting the defect of the top block of the automobile comprises the following steps of:
obtaining a gray level histogram corresponding to the image through the gray level image of the car roof leather;
acquiring gray values corresponding to the maximum peak value and the second peak value in the gray histogram;
and averaging the gray values corresponding to the maximum peak value and the second peak value to be used as a first segmentation threshold of the pixel point.
Further, the method for detecting the defect of the automobile top block comprises the following steps:
clustering all white spot pixel points through included angle angles to obtain different white spot pixel point categories;
and selecting the white spot pixel point category of which the average included angle of the pixel points is smaller than the included angle threshold value from all the white spot pixel point categories as candidate defect regions.
Further, the method for detecting the defect of the automobile top block, which is used for obtaining the binary image after each threshold value is segmented, comprises the following steps:
threshold segmentation is carried out on the grey-level image of the car roof leather through different thresholds, and a pixel point set which is larger than the threshold and smaller than the threshold under each threshold are obtained;
and setting the pixel points which are larger than the threshold value under each threshold value to be 1, and setting the pixel points which are smaller than the threshold value to be 0, so as to obtain the binary image after each threshold value is segmented.
Further, the method for detecting the defect of the automobile top block is to obtain the change rate of the comprehensive dictionary of each region to be processed, and comprises the following steps:
obtaining the change rate of the dictionary matrix of each to-be-processed area under the adjacent threshold value by the dictionary matrix of each to-be-processed area under the binary image after the adjacent threshold value is divided;
and obtaining the comprehensive dictionary change rate of each to-be-processed area according to the change rate of the dictionary matrix of each to-be-processed area under all pairwise adjacent threshold values.
Further, in the method for detecting defects of the top block of the automobile, the expression of the change rate of the dictionary matrix of each to-be-processed area under the adjacent threshold values in pairs is as follows:
in the formula:representing the change rate of the dictionary matrix of the ith to-be-processed area under the jth threshold value and the dictionary matrix under the adjacent threshold value,a dictionary matrix representing the ith area to be processed under the jth threshold,representing the dictionary matrix of the ith to-be-processed area under the j-1 st threshold,representing the norm of the matrix.
Further, in the method for detecting the defect of the automobile top block, the expression of the defect probability of the region to be processed is as follows:
in the formula:indicates the defect probability of the ith area to be processed,represents the integrated dictionary change rate of the ith region to be processed,and representing the average value of included angles of all pixel points in the candidate defect region corresponding to the ith region to be processed.
The invention has the beneficial effects that: the invention relates to a method based on artificial intelligence in the production field, which is characterized in that an initial threshold value is utilized to segment out a white spot region, then clustering analysis is carried out based on direction consistency characteristics to obtain a candidate defect region, and then the change rate of a dictionary different from the threshold value is combined to obtain the defect probability, so that the accurate defect region is segmented out. Compared with the prior art, the method can be used for segmenting accurate defect areas, is high in detection precision, simple in detection method and labor-saving.
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 flow chart of an embodiment of a method for detecting defects of an automobile top block according to 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.
Example 1
An embodiment of a method for detecting defects of an automobile top block of the present invention, as shown in fig. 1, includes:
101. acquiring a grey-scale image of the car roof leather;
the embodiment processes the collected car roof leather image, and then cuts off the scratch defect on the car roof leather.
The scratch defect is similar to the original color of the car roof leather, so that the defect area is hardly accurately segmented by utilizing a threshold value, the car roof leather has original shading texture, the edge boundary of the scratch defect is not clear enough, the internal defect in the middle of the scratch defect has discontinuous texture, the scratch defect which is segmented can have a plurality of communicated areas, and the defect area is also hardly accurately positioned by directly utilizing the area information.
Firstly, arranging a camera, collecting a car roof leather image, and analyzing the image to obtain an accurate defect area.
The operation is to fix the camera on a device, drive the camera to extend into the interior of the automobile carriage by the device, and acquire the leather image on the roof of the carriage.
Further converting the leather image from RGB image to gray imageAnd subsequently, analyzing through the gray level image to obtain an accurate defect position.
At this point, a camera is erected to collect an image, and the image is subjected to graying processing to obtain a grayscale image.
102. Performing threshold segmentation on the grey-scale image of the car roof leather through a first segmentation threshold of pixel points to obtain a white spot pixel point set which is larger than the segmentation threshold;
because the defect area is an area with larger gray scale, a pixel set with larger gray scale value is divided by a segmentation threshold value, and meanwhile, because the scratch defect has a direction consistency characteristic, the candidate defect area is obtained by carrying out cluster analysis based on the characteristic, and because the scratch defect area shows larger characteristic change compared with the common leather texture spot along with the change of the threshold value, the defect probability of each candidate defect needs to be calculated by utilizing the change characteristic of each candidate area dictionary. The specific operation is as follows:
firstly, a region with a larger gray value is segmented by using an initial threshold, specifically: counting the frequency of the gray value of each pixel point of the gray image to generate a gray histogram, solving the gray value corresponding to the maximum peak value and the gray value corresponding to the second peak value in the gray histogram, taking the average value of the two gray values as a segmentation threshold, and screening out a pixel point set which is larger than the threshold, wherein the pixel in the pixel point set is a white spot pixel point.
103. Obtaining a principal component direction of the white spot pixel point set by adopting a PCA algorithm, obtaining an included angle of each white spot pixel point by utilizing a gradient direction angle and a principal component direction of each white spot pixel point in the white spot pixel point set, and clustering all white spot pixel points through the included angle to obtain a candidate defect region;
PCA principal component analysis is carried out on each exudate pixel based on the texture direction characteristics, and the principal component direction with the maximum feature value of the exudate pixel is obtained and becomes the first principal component direction. A direction angle of the first principal component direction is acquired, the direction angle being obtained based on the horizontal direction being the 0-degree direction. The direction angle is the main texture direction of the white spot pixel in the image. Since the white spot pixels have both defective pixels and leather texture pixels, the leather texture is an irregular polygon, the texture has no obvious directionality, and the scratch-off defective pixels have obvious directional characteristics, the main texture direction obtained here mainly reflects the main texture direction of the defective pixels.
And obtaining the gradient direction angle of each white spot pixel point, and calculating the included angle between the gradient direction angle of each white spot pixel point and the direction angle of the first principal component direction. And carrying out density clustering on each white spot pixel point based on the included angle to obtain a plurality of category sets. The probability that the class set with the smaller included angle is a defect is higher, so that a candidate defect area is screened out based on the probability, and the specific method is as follows:
and acquiring the average included angle of the pixel point sets of all categories, and acquiring the category set with the included angle smaller than 30 degrees.
And taking the area surrounded by the boundary pixel points as a candidate defect area through the boundary pixel points of each category set.
Clustering white spot pixel points according to the included angle to obtain different categories, then calculating the average value of the included angle of all the pixel points in each category, and selecting the cluster with the smaller average value of the included angle as a candidate defect region.
104. Respectively carrying out threshold segmentation on the grey-scale images of the car roof leather through different thresholds to obtain binary images after each threshold segmentation;
each candidate defect region is obtained in the above manner, however, there may be non-defect regions in these regions, so it is necessary to analyze the possibility that each region belongs to a defect through the feature change of each defect region under different thresholds.
Since scratch defects may exhibit different characteristics at different thresholds, for example discrete points at high thresholds and progressive formation of connected directional scratch areas at low thresholds. And the white spots of the non-defective region do not have directional connected regions under different thresholds. The scratch-off defect area thus exhibits a greater rate of change in characteristics at different thresholds than the non-defect area.
Respectively utilize,,,,,,,The original gray level image is subjected to threshold segmentation by the threshold value, pixel sets larger than the threshold value and pixel sets smaller than the threshold value under different threshold values are segmented, binarization processing is carried out on each image (the pixel set larger than the threshold value is set to be 1, and the pixel set smaller than the threshold value is set to be 0, wherein the pixel set of the center 1 is the white spot pixel 2 obtained under the threshold value), and binary images under different threshold value segmentation are obtained. Here, theIs the initial threshold in the previous step.
The above steps are to segment the gray-scale image by using different thresholds respectively, for example, usingSegmenting the gray level image to obtain a value greater thanSum of pixel points less thanThen, the image is subjected to binarization processing, and the value of the pixel point set is larger than that of the pixel point setThe pixel value of the pixel point of (1) is less thanSetting the pixel value of the pixel point to 0 to obtain a corresponding binary image, namely the threshold valueAnd (4) corresponding to the binary image.
105. Extracting a region corresponding to the candidate defect region in the binarized image after each threshold segmentation as a region to be processed;
and finding out pixel points of the pixel points of each candidate defect region at corresponding positions in the binary image corresponding to different thresholds to obtain candidate defect regions of the binary image corresponding to different thresholds.
106. Obtaining a dictionary matrix of each to-be-processed area under the corresponding binary image through the pixel information of each to-be-processed area in the corresponding binary image;
acquiring pixel information of each candidate defect region in each threshold value binary image, taking the pixel information of the binary image under each threshold value in each candidate defect region as input, training to obtain a dictionary matrix and a sparse matrix, wherein the dictionary matrix is obtained through a dictionary learning algorithm, and the method is a known technology. By the method, dictionary matrixes of all candidate defect regions under different threshold values are obtained.
107. Obtaining the comprehensive dictionary change rate of each to-be-processed area through the dictionary matrix of each to-be-processed area under the binary image after all threshold segmentation;
analyzing the change rate of dictionary matrixes of the same candidate defect region under different threshold values, wherein the higher the change rate is, the higher the probability that the candidate defect region belongs to the defect is, and the calculation mode of the change rate of the dictionary is as follows:
analyzing based on a single candidate defect area, acquiring dictionary matrixes under different thresholds, and calculating the change rate of every two dictionary matrixes, wherein the expression is as follows:
in the formula:representing the change rate of the dictionary matrix of the ith candidate defect region under the jth threshold value and the dictionary matrix under the adjacent threshold value,a dictionary matrix representing the ith defect candidate area under the jth threshold,a dictionary matrix representing the ith defect candidate area under the j-1 th threshold valueJ-1, the adjacent threshold value,representing the norm of the matrix.
Analogy to this method to obtain the dictionary matrix change rate under two adjacent threshold values.
Averaging the dictionary change rates of the ith candidate defect area and the adjacent threshold values to obtain the comprehensive dictionary change rate of the ith candidate defect area。
108. And obtaining the defect probability of each to-be-processed area through the average value of the included angles of all pixel points in the candidate defect area corresponding to each to-be-processed area and the change rate of the comprehensive dictionary of the to-be-processed area, and judging whether each to-be-processed area is a defect area according to the defect probability to obtain a final defect area.
The salient features of the scratch defects include directional consistency and dictionary presenting large-variation features along with threshold variation, so that the defect probability of each candidate defect area can be calculated by utilizing the directional consistency and the dictionary variation rate, specifically:
in the formula:indicating the defect probability of the ith candidate defect region,the change rate of the comprehensive dictionary representing the ith candidate defect area is used for describing the change condition of the characteristics of the candidate defect area,and representing the average value of included angle angles of all pixel points in the ith candidate defect region, and reflecting the direction consistency characteristic of the region through the value.
And obtaining candidate defect regions through direction consistency, and then obtaining the defect probability of each candidate defect region by combining the dictionary change rates under different thresholds.
Screening out the defect probability greater than the threshold valueThe candidate defect area is judged to be the accurate defect area.
The invention relates to a method based on artificial intelligence in the production field, which is characterized in that an initial threshold value is utilized to segment out a white spot region, then clustering analysis is carried out based on direction consistency characteristics to obtain a candidate defect region, and then the change rate of a dictionary different from the threshold value is combined to obtain the defect probability, so that the accurate defect region is segmented out. Compared with the prior art, the method can be used for segmenting accurate defect areas, is high in detection precision, simple in detection method and labor-saving.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (7)
1. A method for detecting defects of an automobile top block is characterized by comprising the following steps:
acquiring a grey-scale image of the car roof leather;
performing threshold segmentation on the grey-scale image of the car roof leather through a first segmentation threshold of pixel points to obtain a white spot pixel point set which is larger than the segmentation threshold;
obtaining a principal component direction of the white spot pixel point set by adopting a PCA algorithm, obtaining an included angle of each white spot pixel point by utilizing a gradient direction angle and a principal component direction of each white spot pixel point in the white spot pixel point set, and clustering all white spot pixel points through the included angle to obtain a candidate defect region;
respectively carrying out threshold segmentation on the grey-scale images of the car roof leather through different thresholds to obtain binary images after each threshold segmentation;
extracting a region corresponding to the candidate defect region in the binarized image after each threshold segmentation as a region to be processed;
obtaining a dictionary matrix of each to-be-processed area under the corresponding binary image through the pixel information of each to-be-processed area in the corresponding binary image;
obtaining the comprehensive dictionary change rate of each region to be processed through the dictionary matrix of each region to be processed under the binarized image after all threshold segmentation;
and obtaining the defect probability of each to-be-processed area through the average value of the included angle angles of all the pixel points in the candidate defect area corresponding to each to-be-processed area and the change rate of the comprehensive dictionary of the to-be-processed area, and judging whether each to-be-processed area is a defect area according to the defect probability to obtain a final defect area.
2. The method for detecting the defect of the automobile top block according to claim 1, wherein the method for obtaining the first segmentation threshold of the pixel point comprises the following steps:
obtaining a gray level histogram corresponding to the image through the gray level image of the car roof leather;
acquiring gray values corresponding to the maximum peak value and the second peak value in the gray histogram;
and averaging the gray values corresponding to the maximum peak value and the second peak value to be used as a first segmentation threshold of the pixel point.
3. The method for detecting the defect of the automobile top block according to claim 1, wherein the method for obtaining the candidate defect area comprises the following steps:
clustering all white spot pixel points through included angle angles to obtain different white spot pixel point categories;
and selecting the white spot pixel point category of which the average included angle of the pixel points is smaller than an included angle threshold value from all the white spot pixel point categories as candidate defect regions.
4. The method for detecting the defect of the automobile top block according to claim 1, wherein the method for obtaining the binarized image after each threshold segmentation comprises the following steps:
threshold segmentation is carried out on the grey-level image of the car roof leather through different thresholds, and a pixel point set which is larger than the threshold and smaller than the threshold under each threshold are obtained;
and setting the pixel points which are larger than the threshold value under each threshold value to be 1, and setting the pixel points which are smaller than the threshold value to be 0, so as to obtain the binary image after each threshold value is segmented.
5. The method for detecting the defects of the automobile top block according to claim 1, wherein the method for obtaining the change rate of the comprehensive dictionary of each region to be processed comprises the following steps:
obtaining the change rate of the dictionary matrix of each to-be-processed area under the adjacent threshold value by the dictionary matrix of each to-be-processed area under the binary image after the adjacent threshold value is divided;
and obtaining the comprehensive dictionary change rate of each to-be-processed area according to the change rate of the dictionary matrix of each to-be-processed area under all pairwise adjacent threshold values.
6. The method for detecting the defects of the automobile top block according to claim 5, wherein the expression of the change rate of the dictionary matrix of each to-be-processed area under the adjacent threshold values in pairs is as follows:
in the formula:representing the change rate of the dictionary matrix of the ith to-be-processed area under the jth threshold value and the dictionary matrix under the adjacent threshold value,a dictionary matrix representing the ith area to be processed under the jth threshold,representing the dictionary matrix of the ith to-be-processed area under the j-1 st threshold,representing the norm of the matrix.
7. The method for detecting the defect of the automobile ceiling block according to claim 1, wherein the defect probability expression of the region to be processed is as follows:
in the formula:indicates the defect probability of the ith area to be processed,represents the integrated dictionary change rate of the ith region to be processed,and representing the average value of included angles of all pixel points in the candidate defect region corresponding to the ith region to be processed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210754325.3A CN114820625B (en) | 2022-06-30 | 2022-06-30 | Automobile top block defect detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210754325.3A CN114820625B (en) | 2022-06-30 | 2022-06-30 | Automobile top block defect detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114820625A true CN114820625A (en) | 2022-07-29 |
CN114820625B CN114820625B (en) | 2022-09-09 |
Family
ID=82523253
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210754325.3A Active CN114820625B (en) | 2022-06-30 | 2022-06-30 | Automobile top block defect detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114820625B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115049713A (en) * | 2022-08-11 | 2022-09-13 | 武汉中导光电设备有限公司 | Image registration method, device, equipment and readable storage medium |
CN115082488A (en) * | 2022-08-23 | 2022-09-20 | 南通浩盛汽车科技有限公司 | Surface feather mark control method and device in automobile part galvanizing process |
CN115082683A (en) * | 2022-08-22 | 2022-09-20 | 南通三信塑胶装备科技股份有限公司 | Injection molding defect detection method based on image processing |
CN116152242A (en) * | 2023-04-18 | 2023-05-23 | 济南市莱芜区综合检验检测中心 | Visual detection system of natural leather defect for basketball |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504674A (en) * | 2014-10-15 | 2015-04-08 | 西北工业大学 | Space debris star extraction and positioning method |
CN109454006A (en) * | 2018-10-08 | 2019-03-12 | 杭州慧知连科技有限公司 | Chemical fiber wire ingot is stumbled the device and its detection stage division of a defect on-line checking and classification |
CN114581475A (en) * | 2022-03-02 | 2022-06-03 | 西南石油大学 | Laser stripe segmentation method based on multi-scale saliency features |
-
2022
- 2022-06-30 CN CN202210754325.3A patent/CN114820625B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504674A (en) * | 2014-10-15 | 2015-04-08 | 西北工业大学 | Space debris star extraction and positioning method |
CN109454006A (en) * | 2018-10-08 | 2019-03-12 | 杭州慧知连科技有限公司 | Chemical fiber wire ingot is stumbled the device and its detection stage division of a defect on-line checking and classification |
CN114581475A (en) * | 2022-03-02 | 2022-06-03 | 西南石油大学 | Laser stripe segmentation method based on multi-scale saliency features |
Non-Patent Citations (1)
Title |
---|
郭慧等: "基于灰度共生矩阵和分层聚类的刨花板表面图像缺陷提取方法", 《林业科学》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115049713A (en) * | 2022-08-11 | 2022-09-13 | 武汉中导光电设备有限公司 | Image registration method, device, equipment and readable storage medium |
CN115049713B (en) * | 2022-08-11 | 2022-11-25 | 武汉中导光电设备有限公司 | Image registration method, device, equipment and readable storage medium |
CN115082683A (en) * | 2022-08-22 | 2022-09-20 | 南通三信塑胶装备科技股份有限公司 | Injection molding defect detection method based on image processing |
CN115082488A (en) * | 2022-08-23 | 2022-09-20 | 南通浩盛汽车科技有限公司 | Surface feather mark control method and device in automobile part galvanizing process |
CN115082488B (en) * | 2022-08-23 | 2022-11-11 | 南通浩盛汽车科技有限公司 | Surface feather mark control method and device in automobile part galvanizing process |
CN116152242A (en) * | 2023-04-18 | 2023-05-23 | 济南市莱芜区综合检验检测中心 | Visual detection system of natural leather defect for basketball |
CN116152242B (en) * | 2023-04-18 | 2023-07-18 | 济南市莱芜区综合检验检测中心 | Visual detection system of natural leather defect for basketball |
Also Published As
Publication number | Publication date |
---|---|
CN114820625B (en) | 2022-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114820625B (en) | Automobile top block defect detection method | |
CN115018828B (en) | Defect detection method for electronic component | |
CN115082683B (en) | Injection molding defect detection method based on image processing | |
CN115294140B (en) | Hardware part defect detection method and system | |
CN108961235A (en) | A kind of disordered insulator recognition methods based on YOLOv3 network and particle filter algorithm | |
CN111539330B (en) | Transformer substation digital display instrument identification method based on double-SVM multi-classifier | |
CN115147414A (en) | Method for detecting surface breakdown defect of bipolar power transistor | |
CN116645367B (en) | Steel plate cutting quality detection method for high-end manufacturing | |
CN114972356B (en) | Plastic product surface defect detection and identification method and system | |
CN115100199B (en) | Method for detecting wafer low-texture defects | |
CN108734131B (en) | Method for detecting symmetry of traffic sign in image | |
CN110648330B (en) | Defect detection method for camera glass | |
CN116152242B (en) | Visual detection system of natural leather defect for basketball | |
CN113256624A (en) | Continuous casting round billet defect detection method and device, electronic equipment and readable storage medium | |
CN114359288B (en) | Medical image cerebral aneurysm detection and positioning method based on artificial intelligence | |
CN116883408B (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
CN113870202A (en) | Far-end chip defect detection system based on deep learning technology | |
CN117635609B (en) | Visual inspection method for production quality of plastic products | |
CN102214290A (en) | License plate positioning method and license plate positioning template training method | |
CN118096739B (en) | Metal forging surface defect detection method based on image processing | |
CN117437228B (en) | Photovoltaic steel structure component fracture detection method based on image characteristics | |
CN117078678B (en) | Waste silicon wafer shape detection method based on image recognition | |
CN110349119A (en) | Pavement disease detection method and device based on edge detection neural network | |
CN115471494A (en) | Wo citrus quality inspection method, device, equipment and storage medium based on image processing | |
CN114677428A (en) | Power transmission line icing thickness detection method based on unmanned aerial vehicle image processing |
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 |