CN117314899B - Carbon fiber plate quality detection method based on image characteristics - Google Patents
Carbon fiber plate quality detection method based on image characteristics Download PDFInfo
- Publication number
- CN117314899B CN117314899B CN202311595923.1A CN202311595923A CN117314899B CN 117314899 B CN117314899 B CN 117314899B CN 202311595923 A CN202311595923 A CN 202311595923A CN 117314899 B CN117314899 B CN 117314899B
- Authority
- CN
- China
- Prior art keywords
- carbon fiber
- image
- scratch
- pixel
- points
- 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
- 229920000049 Carbon (fiber) Polymers 0.000 title claims abstract description 303
- 239000004917 carbon fiber Substances 0.000 title claims abstract description 303
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 title claims abstract description 303
- 238000001514 detection method Methods 0.000 title claims abstract description 17
- 238000000034 method Methods 0.000 claims abstract description 45
- 230000001186 cumulative effect Effects 0.000 claims abstract description 36
- 230000009036 growth inhibition Effects 0.000 claims abstract description 29
- 239000000835 fiber Substances 0.000 claims description 29
- 206010053759 Growth retardation Diseases 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 6
- 231100000001 growth retardation Toxicity 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 4
- 230000000903 blocking effect Effects 0.000 claims description 2
- 230000007547 defect Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 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
-
- 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
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Geometry (AREA)
- Treatment Of Fiber Materials (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to a carbon fiber plate quality detection method based on image characteristics, which comprises the following steps: obtaining a carbon fiber region of the segmented image according to the carbon fiber plate image, and obtaining scratch judgment parameters of each pixel point in the carbon fiber region of the segmented image; acquiring gray stack parameters of each pixel point in each carbon fiber region of the segmented image, and clustering the gray stack parameters to obtain each cluster, thereby obtaining a cumulative value of each pixel point in each carbon fiber region of the segmented image; acquiring a growth inhibition parameter of each pixel point in each carbon fiber region of the segmented image, performing region growth on each carbon fiber region of the segmented image, and acquiring a scratch point of each carbon fiber region of the segmented image, thereby acquiring each scratch point in the carbon fiber plate image; the authenticity of each scratch point in the carbon fiber board image is obtained, so that the real scratch points and scratch areas are obtained, and the scratch areas identified by the method are more accurate.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a carbon fiber plate quality detection method based on image characteristics.
Background
In the process of producing the carbon fiber plate in a factory, the carbon fiber plate can generate scratch defects when being acted by external force, and the scratch defects on the carbon fiber plate can influence the quality and the performance of the carbon fiber plate, so that after the carbon fiber plate is produced, the defect detection is carried out on the carbon fiber plate, defective plates can be found and removed early, and the quality of products is ensured to meet the standard requirements.
Since the texture features of the carbon fiber image are too regular, a plurality of grids with distinct intervals are formed by two gray values, and scratches are continuous and possibly continuously distributed in the grids, the calculated amount is too large when the existing region growing algorithm is used for detection, the growing rule is difficult to determine, and the detection result is easy to generate false detection.
Disclosure of Invention
In order to solve the problems, the invention provides a carbon fiber plate quality detection method based on image characteristics.
The carbon fiber plate quality detection method based on image features adopts the following technical scheme:
one embodiment of the invention provides a carbon fiber plate quality detection method based on image characteristics, which comprises the following steps:
collecting carbon fiber plate images;
dividing the carbon fiber plate image to obtain a plurality of divided images; acquiring each carbon fiber region of each segmented image;
obtaining scratch judgment parameters of each pixel point in each carbon fiber region of each divided image according to gray value differences among the pixel points in each carbon fiber region of each divided image;
acquiring gray stack parameters of each pixel point in each carbon fiber region of each segmented image according to the gray average value of the pixel point in a sliding window taking each pixel point as the center and the number of the pixel points with gray values different from 0 in the sliding window in each carbon fiber region of each segmented image; clustering gray stack parameters of each pixel point in each carbon fiber region of each segmented image to obtain a cluster of each carbon fiber region of each segmented image; acquiring a cumulative value of each pixel point in each carbon fiber region of each segmented image according to a difference value between gray stack parameters of the pixel points in a cluster of each carbon fiber region of each segmented image, wherein the cumulative value represents an evaluation value when the pixel point is used as a scratch point;
acquiring a growth inhibition parameter of each pixel point in each carbon fiber region of each segmented image according to the cumulative value of each pixel point in each carbon fiber region of each segmented image and the scratch judgment parameter; according to the growth inhibition parameters of each pixel point in each carbon fiber region of each segmented image, carrying out region growth on each carbon fiber region of each segmented image, and obtaining scratch points of each carbon fiber region of each segmented image;
taking the scratch points of the carbon fiber areas of all the divided images as each scratch point in the carbon fiber plate image; obtaining the authenticity of each scratch point in the carbon fiber board image according to the growth blocking parameters of the scratch points and the distance between the scratch points, and obtaining the real scratch points in the carbon fiber board image according to the authenticity of each scratch point in the carbon fiber board image;
and carrying out connected domain analysis on the real scratch points in the carbon fiber plate image, and taking the obtained connected domain as a scratch area.
Preferably, the segmenting the carbon fiber plate image to obtain a plurality of segmented images includes the following specific steps:
obtaining an optimal threshold value of a carbon fiber plate image by using an Ojin method, keeping the gray value of a pixel point larger than the optimal threshold value in the carbon fiber plate image unchanged, and setting the gray of the pixel point smaller than the optimal threshold value to be 0 to obtain a segmentation image; and setting the gray value of the pixel point larger than the optimal threshold value in the carbon fiber plate image to be 0, and obtaining another segmentation image with the gray value of the pixel point smaller than the optimal threshold value unchanged.
Preferably, the step of acquiring each carbon fiber region of each segmented image includes the following specific steps:
and carrying out four-connected domain analysis on each divided image to obtain each connected domain in the two divided images, and marking the connected domain with gray value not being 0 in each divided image as a carbon fiber region.
Preferably, the step of obtaining the scratch evaluation parameter of each pixel point in each carbon fiber region of each segmented image includes the following specific steps:
marking any carbon fiber area of any segmented image as a current carbon fiber area;
in the method, in the process of the invention,represents the +.>Gray values of the individual pixels; />Represents the division of the present carbon fiber region by +.>No. except for the individual pixels>Gray values of the individual pixels; />Representing the number of pixel points in the current carbon fiber area; />Represents the +.>Gray scale difference degree between each pixel point and other pixel points;
in the method, in the process of the invention,represents the +.>Scratch judging parameters of the pixel points; />Representing the average value of the difference degrees of all pixel points in the current carbon fiber region; />Representing absolute value symbols.
Preferably, the step of obtaining the gray stack parameter of each pixel point in each carbon fiber region of each divided image according to the gray average value of the pixel point in the sliding window with each pixel point as the center and the number of the pixel points with gray values different from 0 in the sliding window includes the following specific steps:
acquiring the area of the current carbon fiber area asAcquiring side length of the current carbon fiber region>，/>Representing the whole symbol down, if +.>Odd number, the sliding window of the current carbon fiber region is +.>If->Even, the sliding window of the current carbon fiber region is +.>Obtaining a sliding window of each carbon fiber area in each segmented image;
traversing each pixel point in the current carbon fiber region according to the sliding window of the current carbon fiber region;
in the method, in the process of the invention,represents the +.>Gray stack parameters for each pixel; />Representing the current carbon fiber region with +.>The first +.in the sliding window with the pixel as the center>Gray values of the individual pixels; />Representing the current carbon fiber region with +.>The number of pixels in the sliding window with the pixels as the center; />Representing the current carbon fiber region with +.>The number of pixels with gray values other than 0 in the sliding window with the center of each pixel.
Preferably, the step of obtaining the cluster of each carbon fiber region of each segmented image includes the following specific steps:
traversing each pixel point in the current carbon fiber region by using a sliding window of the current carbon fiber region, counting the number of pixel points with gray values not being 0 in the sliding window taking each pixel point as the center in the current carbon fiber region, constructing a number set, recording the same data in the number set as a category, and classifying the categories in the number setThe other number is used as the clustering number and is reusedClustering the gray stack parameters of each pixel point in the current carbon fiber region by a clustering algorithm to obtain a cluster of the current fiber region;
and acquiring clusters of the carbon fiber areas of each segmented image.
Preferably, the acquiring the cumulative value of each pixel point in each carbon fiber region of each segmented image includes the following specific steps:
in the method, in the process of the invention,represents the +.>Cumulative values of the individual pixel points; />Represents the +.>Gray stack parameters for each pixel; />Represents the +.>Dividing the clustering cluster to which each pixel point belongs by the +.>No. except for the individual pixels>Gray stack parameters for each pixel; />Represents the +.>The number of the pixel points in the cluster to which the pixel points belong; />Representing absolute value symbols;
and obtaining the cumulative value of each pixel point in each carbon fiber region of each segmented image.
Preferably, the step of obtaining the growth inhibition parameter of each pixel point in each carbon fiber region of each segmented image includes the following specific steps:
wherein,represents the +.>Growth retardation parameters of the individual pixels; />Represents the +.>Cumulative values of the individual pixel points; />Represents the +.>The cumulative value average value of other pixel points in the cluster to which the pixel points belong is +.>Represents the +.>Scratch judging parameters of the pixel points;
and obtaining the growth inhibition parameter of each pixel point in each carbon fiber region of each segmented image.
Preferably, the step of obtaining the scratch points of the respective carbon fiber regions of each segmented image includes the following specific steps:
and presetting a growth threshold, acquiring a pixel point with the maximum gray value in the current fiber region as a seed point of the current fiber region for region growth, if the growth inhibition parameter of any pixel point in the current fiber region is larger than the growth threshold, growing the pixel point, otherwise, not growing, and marking the pixel point which is not grown in the current fiber region as a scratch point of the current fiber region, thereby acquiring the scratch point of each carbon fiber region in each segmented image.
Preferably, the step of obtaining the authenticity of each scratch point in the carbon fiber board image according to the growth inhibition parameter of the scratch point and the distance between the scratch points, and obtaining the authenticity of each scratch point in the carbon fiber board image according to the authenticity of each scratch point in the carbon fiber board image comprises the following specific steps:
in the method, in the process of the invention,representing the +.>Authenticity of the individual scratch points; />Representing the +.>Growth retardation parameters for the individual scratch points; />Representing the sum of growth inhibition parameters of all scratch points in the carbon fiber plate image; />Representing the +.>The average value of the distances between each scratch point and all other scratch points in the carbon fiber board image; obtaining the distance between every two combined scratch points in the carbon fiber image, and obtaining the average value of the distances, namely +.>；/>Representing the longest distance between every two combined scratch points in the carbon fiber image; />Represents an exponential function based on a natural constant;
presetting an authenticity thresholdThe authenticity of any scratch point in the carbon fiber plate image is larger than the authenticity threshold +.>When the scratch points are real scratch points, all real scratch points in the carbon fiber image are obtained.
The technical scheme of the invention has the beneficial effects that: according to the invention, the carbon fiber image is split according to the two gray value characteristics of the carbon fiber image, so that each carbon fiber region of the two divided images is obtained for subsequent analysis, the influence of the two gray value characteristics of the carbon fiber image on a region growing algorithm is avoided, then the scratch judging parameter of each pixel point in each carbon fiber region of each divided image is obtained, the gray stack parameter of each pixel point in each carbon fiber region of each divided image is obtained, and further the gray stack parameter of each pixel point in each carbon fiber region of each divided image is clustered to obtain the cumulative value of each pixel point in each carbon fiber region of each divided image; according to the scratch judgment parameters and the cumulative values of each pixel point in each carbon fiber region, the growth inhibition parameters of each pixel point in each carbon fiber region are obtained, then the non-scratch points in each carbon fiber region of each segmented image are subjected to region growth, the non-grown scratch points are the scratch points of each carbon fiber region of each segmented image, namely, each scratch point in the carbon fiber plate image, the authenticity of each scratch point in the carbon fiber plate image is finally obtained, the real scratch points in the carbon fiber plate image are further obtained, the scratch points obtained by setting the growth rules of the region growth algorithm are judged, the real scratch points are obtained, and the obtained scratch regions are more accurate.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of steps of a carbon fiber plate quality detection method based on image features of the present invention;
FIG. 2 is a carbon fiber sheet image;
fig. 3 is a schematic diagram of a carbon fiber region and a background region in a segmented image.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of the specific implementation, structure, characteristics and effects of the carbon fiber plate quality detection method based on image characteristics according to the invention with reference to the attached drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the carbon fiber plate quality detection method based on image characteristics provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a method for detecting quality of a carbon fiber plate based on image features according to an embodiment of the invention is shown, the method includes the following steps:
s001, collecting carbon fiber plate images.
In the embodiment of the invention, a camera is used for shooting and producing the carbon fiber plate, the carbon fiber plate is subjected to gray-scale treatment, and an image obtained through the gray-scale treatment is recorded as a carbon fiber plate image.
S002, dividing the carbon fiber plate image into two divided images according to gray features of the carbon fiber plate image, and acquiring each carbon fiber region of each divided image.
It should be noted that, the carbon fiber plate image has a certain texture characteristic and is composed of two grids which are mainly gray values and are distributed in a crossing way, see fig. 2, and because the two main gray values of the carbon fiber image are greatly different, the carbon fiber plate image can be divided into two images according to the gray value characteristic of the carbon fiber plate, so that each divided image can be analyzed independently.
In the embodiment of the invention, the optimal threshold value of the carbon fiber board image is obtained by using the Ojin method, and it is to be noted that the threshold value is obtained by using the Ojin method in the prior art, in the embodiment of the invention, the gray value of the pixel point in the carbon fiber board image is compared with the optimal threshold value, the gray value of the pixel point larger than the optimal threshold value is unchanged, the gray value of the pixel point smaller than the optimal threshold value is set as 0, a segmented image is obtained, the gray value of the pixel point larger than the optimal threshold value is set as 0, and the gray value of the pixel point smaller than the optimal threshold value is unchanged, so that another segmented image is obtained.
It should be noted that, the two obtained divided images are formed by respective spaced lattices, and the two obtained divided images may not be complete according to the optimal threshold, so in the embodiment of the present invention, four connected domain analysis is performed on each divided image to obtain each connected domain in each divided image, the connected domain with a gray value other than 0 in each divided image is marked as a carbon fiber region, the connected domain with a gray value of 0 in the two divided images is marked as a background region, and the middle, upper left, upper right, lower left and lower right regions are carbon fiber regions, and the other regions are background regions, as shown in fig. 3.
S003, scratch judgment parameters of each pixel point in each carbon fiber region of each segmented image are obtained.
It should be noted that, given that the carbon fiber image is a plurality of grids with distinct intervals formed by two gray values, and scratches are continuous and may be continuously distributed in a plurality of grids, so that when the existing region growing algorithm is used for detecting, the growth rule is difficult to determine, and the detection result is easy to generate a false detection phenomenon.
In the embodiment of the invention, any carbon fiber region of any segmented image is recorded as a current carbon fiber region, and scratch judgment parameters of each pixel point in the current carbon fiber region are obtained:
in the method, in the process of the invention,represents the +.>Gray values of the individual pixels; />Represents the division of the present carbon fiber region by +.>No. except for the individual pixels>Gray values of the individual pixels; />Representing the number of pixel points in the current carbon fiber area; />Represents the +.>Gray level difference degree of each pixel point and other pixel points, < >>When the value of (2) is larger, the first +.>The more likely a pixel point is a scratch point.
In the method, in the process of the invention,represents the +.>Scratch of individual pixel pointsJudging parameters; />Represents the +.>The degree of difference of the individual pixel points; />Representing the average value of the difference degrees of all pixel points in the current carbon fiber region; />Representing absolute value symbols; if the current carbon fiber region is +.>When the gray level difference degree between each pixel point and other pixel points is larger than the gray level difference degree between all pixel points and other pixel points in the current carbon fiber region, the method indicates the (th) in the current carbon fiber region>The more likely the pixel is the pixel of the scratch area, so the +.>The larger the scratch evaluation parameter of each pixel point.
And similarly, acquiring scratch judgment parameters of each pixel point in each carbon fiber region of each segmented image.
Thus, scratch judgment parameters of each pixel point in each carbon fiber region of each divided image are obtained.
S004, acquiring gray stack parameters of each pixel point in each carbon fiber region of each segmented image, clustering the gray stack parameters to obtain each cluster of each carbon fiber region of each segmented image, and acquiring a cumulative value of each pixel point in each carbon fiber region of each segmented image according to each cluster of each carbon fiber region of each segmented image.
It should be noted that, the gray value of the pixel point in the scratch area is lower than the gray value of the normal pixel point in the carbon fiber area, so in the embodiment of the invention, the sliding window is set to traverse each pixel point in the carbon fiber area, when the gray average value of all the pixel points in the sliding window corresponding to each pixel point is lower, the pixel point is likely to be the scratch point, namely the pixel point in the scratch area, so that the sliding window of each carbon fiber area in each divided image is firstly obtained, and then the gray stack parameter of each pixel point of each carbon fiber area in each divided image is obtained according to the sliding window of each carbon fiber area in each divided image.
In the embodiment of the invention, any carbon fiber region of any segmented image is recorded as a current carbon fiber region, and a sliding window of the current carbon fiber region is obtained: acquiring the area of the current carbon fiber area asAcquiring side length of the current carbon fiber region>，/>Represents a downward rounding symbol, if +.>Odd number, the sliding window of the current carbon fiber region is +.>If->Even, the sliding window of the current carbon fiber region is +.>And similarly, acquiring a sliding window of each carbon fiber area in each segmented image.
In the embodiment of the invention, any carbon fiber region of any segmented image is recorded as a current carbon fiber region, each pixel point in the current carbon fiber region is traversed according to a sliding window of the current carbon fiber region, and gray stack parameters of each pixel point in the current carbon fiber region are obtained:
in the method, in the process of the invention,represents the +.>Gray stack parameters for each pixel; />Representing the current carbon fiber region with +.>The first +.in the sliding window with the pixel as the center>Gray values of the individual pixels; />Representing the current carbon fiber region with +.>The number of pixels in the sliding window with the pixels as the center; />Representing the current carbon fiber region with +.>The number of pixels with gray values different from 0 in the sliding window with the pixel points as the center; if the current carbon fiber region is at +.>All images in a sliding window with a single pixel as the centerWhen the gray value of the pixel is smaller, the gray stack parameter is smaller, which indicates the +.>The more likely a pixel point is a scratch point; and similarly, acquiring the gray stack parameters of each pixel point in each carbon fiber region in each divided image.
It should be noted that, when the gray stack parameter of each pixel point in the carbon fiber area is obtained, the smaller the gray stack parameter is, the more likely the pixel point is the pixel point of the scratch area, when the current carbon fiber area is traversed according to the sliding window of the current carbon fiber area and the gray stack parameter of the pixel point is calculated, the number of non-0 pixel points participating in calculation in the sliding window of each pixel point is different, so the obtained gray stack parameter has different values, therefore, the pixel point which is likely to be the scratch point is obtained according to the difference of the gray stack parameters of all the pixel points in the carbon fiber area is inaccurate, therefore, in the embodiment of the invention, the gray stack parameters of all the pixel points in the carbon fiber area need to be clustered, and the cumulative value of each pixel point in the carbon fiber area is obtained according to the difference between the gray stack parameters of the pixel points in the clustering type.
In the embodiment of the invention, the clustering number of the current carbon fiber area is obtained: recording any carbon fiber region of any segmented image as a current carbon fiber region, traversing each pixel point in the current carbon fiber region by using a sliding window of the current carbon fiber region, counting the number of pixel points with gray values not being 0 in the sliding window taking each pixel point as the center in the current carbon fiber region, constructing a number set, recording the same data in the number set as a category, taking the category number in the number set as a clustering number, and then usingClustering the gray stack parameters of each pixel point in the current carbon fiber region by a clustering algorithm to obtain a cluster of the current fiber region; and similarly, obtaining cluster clusters of each carbon fiber area in each segmented image.
For example: assume that the number set isAnd 5, 6 and 7 are respectively one category, and the category number of the quantity set is 3.
In the embodiment of the invention, the cumulative value of each pixel point in the current fiber region is acquired:
in the method, in the process of the invention,represents the +.>Cumulative values of the individual pixel points; />Represents the +.>Gray stack parameters for each pixel; />Represents the +.>Dividing the clustering cluster to which each pixel point belongs by the +.>No. except for the individual pixels>Gray stack parameters for each pixel; />Represents the +.>The number of the pixel points in the cluster to which the pixel points belong; />Representing absolute value symbols; if the current carbon fiber region is +.>The larger the gray stack parameter of each pixel point is than the sum of the differences of the gray stack parameters of other pixel points in the cluster to which the gray stack parameter belongs, namely +.>The larger the cumulative value of each pixel point is, the cumulative value represents the evaluation value when the pixel point is used as the scratch point, and when the cumulative value is larger, the description of the +.>The more likely that a pixel is a pixel where a scratch area is located; and similarly, obtaining the cumulative value of each pixel point in each carbon fiber area in each segmented image.
The gray stack parameters of each pixel point in each carbon fiber region of each segmented image are acquired, and are clustered to obtain each cluster of each carbon fiber region of each segmented image, and the cumulative value of each pixel point in each carbon fiber region of each segmented image is acquired according to each cluster of each carbon fiber region of each segmented image.
S005, obtaining growth inhibition parameters of each pixel point in each carbon fiber region of each segmented image according to the cumulative value of each pixel point in each carbon fiber region of each segmented image and scratch judgment parameters, and further obtaining the scratch points of each fiber region in each segmented image.
It should be noted that, when the difference between the cumulative value of any pixel point and the cumulative value of other pixel points in the cluster to which the pixel point belongs is larger, the pixel point is more likely to be a scratch point, so that by searching for an abnormal value of the cumulative value of each pixel point in each cluster, the pixel point generating the abnormal value has a high probability of being a scratch point, then the scratch judgment parameter of each pixel point is combined to obtain the growth inhibition parameter of each pixel point, and the larger the value is, the more likely the pixel point is to be a scratch point, and the less likely the pixel point is to be grown subsequently.
In the embodiment of the invention, the growth inhibition parameter of each pixel point in the current fiber area is obtained:
wherein,represents the +.>Growth retardation parameters of the individual pixels; />Represents the +.>Cumulative values of the individual pixel points; />Represents the +.>The cumulative value average of other pixels in the cluster to which the pixel belongs is equal to the +.>When the cumulative value of each pixel point is far greater than the cumulative value average value of other pixel points of the cluster to which the pixel point belongs, then the +.>The more likely a pixel is a pixel of a scratch area, < >>Represents the +.>Scratch evaluation parameter of each pixel, when +.>The greater the value of (2), the +.>The more likely a pixel is a pixel of a scratch area, < >>The greater the value, the description of the +.>The larger the growth inhibition parameter of each pixel point is, the less likely the pixel point is to be grown when the pixel point in the current fiber area is grown in the area, and the growth inhibition parameter of each pixel point in the current fiber area is obtained.
It should be noted that after the growth inhibition parameter of each pixel point in the current fiber area is obtained, the pixel points in the current fiber area need to be subjected to area growth, when the pixel points with larger growth inhibition are encountered during growth, the pixel points in the scratch area may not grow, and because the gray value of the normal pixel points in the carbon fiber area of each divided image is larger than the gray value of the pixel points in the scratch area, the seed points in the current fiber area can be obtained according to the characteristics, and the current fiber area needs to be subjected to area growth according to the growth inhibition parameter of each pixel point in the current fiber area, so that the scratch points in the current fiber area, namely, the pixel points which are not grown, are obtained.
In the embodiment of the invention, the growth threshold is presetObtaining a pixel point with the maximum gray value in the current fiber area as a seed point of the current fiber area to perform area growth, and if the growth inhibition parameter of any pixel point in the current fiber areaGrowing the pixel points, otherwise, marking the pixel points which are not grown in the current fiber area as scratch points of the current fiber area, and setting a growth threshold value +_ in the embodiment of the invention>In other embodiments, the practitioner can set +.>Is a value of (2). Similarly, scratch points of the respective fiber regions in each segmented image are obtained.
Thus, the scratch points of the respective fiber regions in each segmented image are obtained.
S006, obtaining each scratch point in the carbon fiber plate image according to the scratch points of each fiber region in each segmented image, obtaining the authenticity of each scratch point in the carbon fiber plate image, and obtaining the real scratch point according to the authenticity of each scratch point in the carbon fiber plate image.
It should be noted that it is known that the scratch points of each carbon fiber region in each divided image are obtained, the scratch points of each carbon fiber region in each divided image are returned to the carbon fiber image to be analyzed, and each scratch point in the carbon fiber image may be a noise point, so that it is necessary to distinguish the obtained scratch points, it is known that the scratch points should be continuous, that is, if the distance from any one scratch point to other scratch points is short, the authenticity of the scratch point is greater, and when the growth inhibition parameter of the scratch point is greater, the authenticity of the scratch point is greater, so in the embodiment of the present invention, the authenticity of each scratch point in the carbon fiber image is obtained according to the distance between the scratch points in the carbon fiber image and the growth inhibition parameter of the scratch point.
In the embodiment of the invention, the authenticity of each scratch point in the carbon fiber board image is obtained:
wherein,representing the +.>Authenticity of the individual scratch points; />Representing the +.>Growth retardation parameters for the individual scratch points; />Representing the sum of growth inhibition parameters of all scratch points in the carbon fiber sheet image, when the carbon fiber sheet image is +>The larger the growth inhibition parameter of each scratch point is, the larger the authenticity of the scratch point is, and the more likely the scratch point is the true scratch point; />Representing the +.>The average value of the distances between each scratch point and all other scratch points in the carbon fiber board image; obtaining the distance between every two combined scratch points in the carbon fiber image, obtaining the average value of the distances, and recording as；/>Representing the longest distance between every two combined scratch points in the carbon fiber image; />Represents an exponential function based on a natural constant; when the carbon fiber plate is imaged +.>When the average value of the distances between each scratch point and all other scratch points in the carbon fiber board image is larger than the average value of the distances between all the scratch points combined in the carbon fiber board image, the distances between each scratch point and other scratch points are far and discontinuous, and the authenticity of each scratch point is low, and the scratch points are less likely to be real scratch points.
Similarly, obtaining the authenticity of all scratch points in the carbon fiber board image, and presetting an authenticity threshold valueWhen the authenticity of any one of the scratch points in the carbon fiber sheet image is larger than the authenticity threshold +.>When the pixel point is a real scratch point, otherwise, the pixel point is not the real scratch point; similarly, all real scratch points in the carbon fiber image are acquired, and in the embodiment of the invention, an authenticity threshold value is preset +.>In other embodiments, the practitioner can preset +_ according to the specific implementation>Is a value of (2).
And obtaining each scratch point in the carbon fiber board image according to the scratch point of each fiber area in each segmented image, obtaining the authenticity of each scratch point in the carbon fiber board image, and obtaining the authenticity of each scratch point in the carbon fiber board image according to the authenticity of each scratch point in the carbon fiber board image.
S007, acquiring a scratch area according to the real scratch points in the carbon fiber plate image.
In the embodiment of the invention, connected domain analysis is carried out on the real scratch points in the acquired carbon fiber plate image, and the obtained connected domain area is marked as a scratch area.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (9)
1. The carbon fiber plate quality detection method based on the image characteristics is characterized by comprising the following steps of:
collecting carbon fiber plate images;
dividing the carbon fiber plate image to obtain a plurality of divided images; acquiring each carbon fiber region of each segmented image;
obtaining scratch judgment parameters of each pixel point in each carbon fiber region of each divided image according to gray value differences among the pixel points in each carbon fiber region of each divided image;
acquiring gray stack parameters of each pixel point in each carbon fiber region of each segmented image according to the gray average value of the pixel point in a sliding window taking each pixel point as the center and the number of the pixel points with gray values different from 0 in the sliding window in each carbon fiber region of each segmented image; clustering gray stack parameters of each pixel point in each carbon fiber region of each segmented image to obtain a cluster of each carbon fiber region of each segmented image; acquiring a cumulative value of each pixel point in each carbon fiber region of each segmented image according to a difference value between gray stack parameters of the pixel points in a cluster of each carbon fiber region of each segmented image, wherein the cumulative value represents an evaluation value when the pixel point is used as a scratch point;
acquiring a growth inhibition parameter of each pixel point in each carbon fiber region of each segmented image according to the cumulative value of each pixel point in each carbon fiber region of each segmented image and the scratch judgment parameter; according to the growth inhibition parameters of each pixel point in each carbon fiber region of each segmented image, carrying out region growth on each carbon fiber region of each segmented image, and obtaining scratch points of each carbon fiber region of each segmented image;
taking the scratch points of the carbon fiber areas of all the divided images as each scratch point in the carbon fiber plate image; obtaining the authenticity of each scratch point in the carbon fiber board image according to the growth blocking parameters of the scratch points and the distance between the scratch points, and obtaining the real scratch points in the carbon fiber board image according to the authenticity of each scratch point in the carbon fiber board image;
carrying out connected domain analysis on real scratch points in the carbon fiber board image, and taking the obtained connected domain as a scratch area;
the step of obtaining scratch judgment parameters of each pixel point in each carbon fiber region of each segmented image comprises the following specific steps:
marking any carbon fiber area of any segmented image as a current carbon fiber area;
in the method, in the process of the invention,represents the +.>Gray values of the individual pixels; />Represents the division of the present carbon fiber region by +.>No. except for the individual pixels>Gray values of the individual pixels; />Representing the current carbonThe number of pixel points in the fiber area; />Represents the +.>Gray scale difference degree between each pixel point and other pixel points;
in the method, in the process of the invention,represents the +.>Scratch judging parameters of the pixel points; />Representing the average value of the difference degrees of all pixel points in the current carbon fiber region; />Representing absolute value symbols.
2. The method for detecting the quality of the carbon fiber plate based on the image characteristics according to claim 1, wherein the steps of segmenting the carbon fiber plate image to obtain a plurality of segmented images include the following specific steps:
obtaining an optimal threshold value of a carbon fiber plate image by using an Ojin method, keeping the gray value of a pixel point larger than the optimal threshold value in the carbon fiber plate image unchanged, and setting the gray of the pixel point smaller than the optimal threshold value to be 0 to obtain a segmentation image; and setting the gray value of the pixel point larger than the optimal threshold value in the carbon fiber plate image to be 0, and obtaining another segmentation image with the gray value of the pixel point smaller than the optimal threshold value unchanged.
3. The method for detecting the quality of the carbon fiber plate based on the image features as set forth in claim 1, wherein the step of acquiring the respective carbon fiber regions of each divided image comprises the following specific steps:
and carrying out four-connected domain analysis on each divided image to obtain each connected domain in the two divided images, and marking the connected domain with gray value not being 0 in each divided image as a carbon fiber region.
4. The method for detecting quality of carbon fiber plates based on image features according to claim 1, wherein the step of obtaining the gray stack parameter of each pixel in each carbon fiber region of each divided image according to the gray average value of the pixel in the sliding window centered on each pixel and the number of pixels with gray values different from 0 in the sliding window, comprises the following specific steps:
recording any carbon fiber region of any segmented image as a current carbon fiber region, and recording the area of the obtained current carbon fiber region asAcquiring side length of the current carbon fiber region>，/>Representing the whole symbol down, if +.>Odd number, the sliding window of the current carbon fiber region is +.>If->Even, then the current carbonThe sliding window of the fibre area is +.>Obtaining a sliding window of each carbon fiber area in each segmented image;
traversing each pixel point in the current carbon fiber region according to the sliding window of the current carbon fiber region;
in the method, in the process of the invention,represents the +.>Gray stack parameters for each pixel; />Representing the current carbon fiber region with +.>The first +.in the sliding window with the pixel as the center>Gray values of the individual pixels; />Representing the current carbon fiber region to the firstThe number of pixels in the sliding window with the pixels as the center; />Representing the current carbon fiber region with +.>The number of pixels with gray values other than 0 in the sliding window with the center of each pixel.
5. The method for detecting quality of carbon fiber sheet based on image features as defined in claim 4, wherein the step of obtaining clusters of the respective carbon fiber regions of each segmented image comprises the specific steps of:
recording any carbon fiber region of any segmented image as a current carbon fiber region, traversing each pixel point in the current carbon fiber region by using a sliding window of the current carbon fiber region, counting the number of pixel points with gray values not being 0 in the sliding window taking each pixel point as the center in the current carbon fiber region, constructing a number set, recording the same data in the number set as a category, taking the category number in the number set as a clustering number, and then usingClustering the gray stack parameters of each pixel point in the current carbon fiber region by a clustering algorithm to obtain a cluster of the current carbon fiber region;
and acquiring clusters of the carbon fiber areas of each segmented image.
6. The method for detecting the quality of the carbon fiber plate based on the image features as set forth in claim 1, wherein the step of obtaining the cumulative value of each pixel point in each carbon fiber region of each divided image comprises the following specific steps:
in the method, in the process of the invention,represents the +.>Cumulative values of the individual pixel points; />Represents the +.>Gray stack parameters for each pixel; />Represents the +.>Dividing the clustering cluster to which each pixel point belongs by the +.>No. except for the individual pixels>Gray stack parameters for each pixel; />Represents the +.>The number of the pixel points in the cluster to which the pixel points belong; />Representing absolute value symbols;
and obtaining the cumulative value of each pixel point in each carbon fiber region of each segmented image.
7. The method for detecting quality of carbon fiber sheet based on image features as defined in claim 1, wherein the step of obtaining the growth inhibition parameter of each pixel point in each carbon fiber region of each divided image comprises the following specific steps:
marking any carbon fiber area of any segmented image as a current carbon fiber area;
wherein,represents the +.>Growth retardation parameters of the individual pixels; />Represents the +.>Cumulative values of the individual pixel points; />Represents the +.>The cumulative value average value of other pixel points in the cluster to which the pixel points belong is +.>Represents the +.>Scratch judging parameters of the pixel points;
and obtaining the growth inhibition parameter of each pixel point in each carbon fiber region of each segmented image.
8. The method for detecting quality of carbon fiber sheet based on image features as defined in claim 1, wherein the step of obtaining the scratch points of the respective carbon fiber regions of each divided image comprises the following specific steps:
and marking any carbon fiber region of any segmented image as a current carbon fiber region, presetting a growth threshold, acquiring a pixel point with the maximum gray value in the current carbon fiber region as a seed point of the current carbon fiber region for region growth, if the growth inhibition parameter of any pixel point in the current carbon fiber region is greater than the growth threshold, growing the pixel point, otherwise, marking the pixel point which is not grown in the current carbon fiber region as a scratch point of the current carbon fiber region, and acquiring the scratch point of each carbon fiber region in each segmented image.
9. The method for detecting quality of carbon fiber plate based on image features of claim 1, wherein the steps of obtaining authenticity of each scratch point in the carbon fiber plate image according to growth inhibition parameters of the scratch points and distances between the scratch points, and obtaining the authenticity of each scratch point in the carbon fiber plate image according to the authenticity of each scratch point in the carbon fiber plate image, comprise the following specific steps:
in the method, in the process of the invention,representing the +.>Authenticity of the individual scratch points; />Representing the +.>Growth retardation parameters for the individual scratch points; />Representing the sum of growth inhibition parameters of all scratch points in the carbon fiber plate image; />Representing the +.>The average value of the distances between each scratch point and all other scratch points in the carbon fiber board image; obtaining the distance between every two pairs of combined scratch points in the carbon fiber image, and obtaining the average value of the distance, and marking the average value as +.>；/>Representing the longest distance between every two combined scratch points in the carbon fiber image; />Represents an exponential function based on a natural constant;
presetting an authenticity thresholdThe authenticity of any scratch point in the carbon fiber plate image is larger than the authenticity threshold +.>When the scratch points are real scratch points, all real scratch points in the carbon fiber image are obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311595923.1A CN117314899B (en) | 2023-11-28 | 2023-11-28 | Carbon fiber plate quality detection method based on image characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311595923.1A CN117314899B (en) | 2023-11-28 | 2023-11-28 | Carbon fiber plate quality detection method based on image characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117314899A CN117314899A (en) | 2023-12-29 |
| CN117314899B true CN117314899B (en) | 2024-03-08 |
Family
ID=89255564
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311595923.1A Active CN117314899B (en) | 2023-11-28 | 2023-11-28 | Carbon fiber plate quality detection method based on image characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117314899B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108074235A (en) * | 2017-12-26 | 2018-05-25 | 太原理工大学 | Carbon fiber surface defect degree method of estimation based on region growing algorithm |
| CN112233133A (en) * | 2020-10-29 | 2021-01-15 | 上海电力大学 | Power plant high-temperature pipeline defect detection and segmentation method based on OTSU and region growing method |
| CN113724241A (en) * | 2021-09-09 | 2021-11-30 | 常州市宏发纵横新材料科技股份有限公司 | Broken filament detection method and device for carbon fiber warp-knitted fabric and storage medium |
| CN114445387A (en) * | 2022-01-29 | 2022-05-06 | 泗阳富艺木业股份有限公司 | Fiberboard quality classification method based on machine vision |
| CN115082431A (en) * | 2022-07-20 | 2022-09-20 | 惠州威尔高电子有限公司 | PCB surface defect detection method |
| CN115511907A (en) * | 2022-11-24 | 2022-12-23 | 深圳市晶台股份有限公司 | Scratch detection method for LED screen |
| CN116310360A (en) * | 2023-05-18 | 2023-06-23 | 实德电气集团有限公司 | Reactor surface defect detection method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8331669B2 (en) * | 2009-08-25 | 2012-12-11 | Siemens Aktiengesellschaft | Method and system for interactive segmentation using texture and intensity cues |
-
2023
- 2023-11-28 CN CN202311595923.1A patent/CN117314899B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108074235A (en) * | 2017-12-26 | 2018-05-25 | 太原理工大学 | Carbon fiber surface defect degree method of estimation based on region growing algorithm |
| CN112233133A (en) * | 2020-10-29 | 2021-01-15 | 上海电力大学 | Power plant high-temperature pipeline defect detection and segmentation method based on OTSU and region growing method |
| CN113724241A (en) * | 2021-09-09 | 2021-11-30 | 常州市宏发纵横新材料科技股份有限公司 | Broken filament detection method and device for carbon fiber warp-knitted fabric and storage medium |
| CN114445387A (en) * | 2022-01-29 | 2022-05-06 | 泗阳富艺木业股份有限公司 | Fiberboard quality classification method based on machine vision |
| CN115082431A (en) * | 2022-07-20 | 2022-09-20 | 惠州威尔高电子有限公司 | PCB surface defect detection method |
| CN115511907A (en) * | 2022-11-24 | 2022-12-23 | 深圳市晶台股份有限公司 | Scratch detection method for LED screen |
| CN116310360A (en) * | 2023-05-18 | 2023-06-23 | 实德电气集团有限公司 | Reactor surface defect detection method |
Non-Patent Citations (2)
| Title |
|---|
| 基于二维OTSU 选取种子点的区域生长图像分割;吴海滨;《大气与环境光学学报》;第8卷(第6期);448-453 * |
| 基于热图重构区域生长算法的碳纤维增强复合材料脱粘缺陷检测;冯琪智;《无损检测》;第39卷(第9期);29-34 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117314899A (en) | 2023-12-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115082467B (en) | Building material welding surface defect detection method based on computer vision | |
| CN116205919B (en) | Hardware part production quality detection method and system based on artificial intelligence | |
| CN115830022B (en) | Filter screen defect detection method based on machine vision | |
| CN116030052B (en) | Etching quality detection method for lamination process of computer display panel | |
| CN115311267B (en) | Method for detecting abnormity of check fabric | |
| CN117392469B (en) | Perovskite battery surface coating detection method and system based on machine vision | |
| CN116309599B (en) | Water quality visual monitoring method based on sewage pretreatment | |
| CN117635609B (en) | Visual inspection method for production quality of plastic products | |
| CN115311277A (en) | Pit defect identification method for stainless steel product | |
| CN116843692B (en) | Regenerated active carbon state detection method based on artificial intelligence | |
| CN117197140A (en) | Irregular metal buckle forming detection method based on machine vision | |
| CN116703251A (en) | Rubber ring production quality detection method based on artificial intelligence | |
| CN117437228B (en) | Photovoltaic steel structure component fracture detection method based on image characteristics | |
| CN116900412B (en) | Steel plate equidistant automatic cutting regulation and control method based on visual processing | |
| CN116681696B (en) | Mold quality monitoring method for automatic production equipment | |
| CN117314899B (en) | Carbon fiber plate quality detection method based on image characteristics | |
| CN116205923B (en) | Nondestructive testing method for internal defects of automobile hub based on X-RAY | |
| CN115063603B (en) | Wood annual ring line accurate extraction and restoration method based on edge information | |
| CN113160166B (en) | Medical image data mining working method through convolutional neural network model | |
| CN114693652A (en) | Fabric defect detection method based on Gaussian mixture model | |
| CN117455911B (en) | Production test method and system for network cable crystal head | |
| CN117557785B (en) | Image processing-based long-distance fishing boat plate recognition method | |
| CN117173168B (en) | OLED display screen quality detection method based on artificial intelligence | |
| CN117689637B (en) | Visual detection method for processing quality of new energy battery hardware | |
| CN116958134B (en) | Plastic film extrusion quality evaluation method based on 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 |