CN115078383A - PVC sheet quality inspection method based on big data - Google Patents
PVC sheet quality inspection method based on big data Download PDFInfo
- Publication number
- CN115078383A CN115078383A CN202210680881.0A CN202210680881A CN115078383A CN 115078383 A CN115078383 A CN 115078383A CN 202210680881 A CN202210680881 A CN 202210680881A CN 115078383 A CN115078383 A CN 115078383A
- Authority
- CN
- China
- Prior art keywords
- pvc sheet
- pixel
- graphic information
- pixel point
- super
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000007689 inspection Methods 0.000 title claims abstract description 60
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 230000002950 deficient Effects 0.000 claims abstract description 5
- 230000005540 biological transmission Effects 0.000 claims abstract description 4
- 238000001914 filtration Methods 0.000 claims description 40
- 239000011159 matrix material Substances 0.000 claims description 23
- 239000002344 surface layer Substances 0.000 claims description 23
- 238000012545 processing Methods 0.000 claims description 15
- 230000001419 dependent effect Effects 0.000 claims description 12
- 238000004891 communication Methods 0.000 claims description 6
- 230000006872 improvement Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000005520 cutting process Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 3
- 230000008569 process Effects 0.000 description 8
- 230000007547 defect Effects 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000010410 layer Substances 0.000 description 2
- BZHJMEDXRYGGRV-UHFFFAOYSA-N Vinyl chloride Chemical compound ClC=C BZHJMEDXRYGGRV-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 229920000642 polymer Polymers 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000010526 radical polymerization reaction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
-
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- 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)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Biochemistry (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Signal Processing (AREA)
- Quality & Reliability (AREA)
- Theoretical Computer Science (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a PVC sheet quality inspection method based on big data, which utilizes the big data to carry out quality inspection on PVC sheets and improves the automation degree and accuracy, and comprises the following procedures: step S1: acquiring PVC sheet graphic information; step S2: preprocessing the graphic information to remove the noise of the graphic information; step S3: performing PVC sheet quality inspection according to the graphic information; step S4: judging whether the defective PVC sheet can be repaired; step S5: the PVC sheet after quality inspection is divided into three types of qualified quality inspection, unqualified quality inspection but repairable, unqualified quality inspection and unrepairable, and the three types are respectively transmitted to the next procedure through different transmission lines. According to the invention, the big data is used for replacing manual work to carry out quality inspection on the PVC sheet, so that the automation degree is greatly improved, the human errors are reduced, and the overall accuracy of the quality inspection is improved.
Description
Technical Field
The invention relates to the technical field of quality inspection of PVC sheets based on big data, in particular to a quality inspection method of PVC sheets based on big data.
Background
Big data is popularly said to be the integration of huge amount of data. In the internet era, legal means is adopted, various information used for capturing, managing and processing are carried out within a certain reasonable time, and then the information is integrated into a huge amount of data set, the deep meaning of the information is a data analysis mode, and finally the improvement of products and services can be helped through the integration of the huge amount of data and the data measurement, calculation and detection, so that the further development of the products and the industry is promoted.
The PVC sheet is a molded plate made of a polymer obtained by free radical polymerization of vinyl chloride monomer, and has excellent corrosion resistance, insulating property and certain mechanical strength. In the production process, the surface of some sheets can generate flaws due to process problems, accidental collision and the like, so that the quality of the surface of the PVC sheet needs to be inspected when the PVC sheet is delivered from a factory.
At present, the accuracy of the traditional quality inspection mode is not high, some procedures also need manual operation, errors exist easily, and the quality inspection efficiency is difficult to meet.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a method for quality inspection of a PVC sheet based on big data, which utilizes the big data to perform quality inspection on the PVC sheet, thereby improving the degree of automation and accuracy.
A PVC sheet quality inspection method based on big data comprises the following steps:
step S1: acquiring PVC sheet graphic information;
step S2: preprocessing the graphic information to remove the noise of the graphic information;
step S3: performing PVC sheet quality inspection according to the graphic information;
step S4: judging whether the defective PVC sheet can be repaired;
step S5: the PVC sheet after quality inspection is divided into three types of qualified quality inspection, unqualified quality inspection but repairable, unqualified quality inspection and unrepairable, and the three types are respectively transmitted to the next procedure through different transmission lines.
Further, the preprocessing method in step S2 is specifically as follows:
s210: filtering the graphic information to obtain filtered graphic information;
s220: analyzing the integral appearance graph information in the PVC sheet filtering graph according to a guiding filtering algorithm;
s230: acquiring PVC sheet fine grain pattern information according to the PVC sheet filtering pattern and the PVC sheet integral surface pattern information, and respectively performing significance processing on the PVC sheet integral surface pattern information and the PVC sheet fine grain pattern information;
s240: and obtaining the obvious graphic information of the PVC sheet according to the overall appearance graphic information of the PVC sheet and the fine grain graphic information of the PVC sheet after the significance processing.
Further, the quality inspection method in step S3 includes the following steps:
s310: carrying out superpixel cutting on the obvious graphic information of the PVC sheet to obtain a plurality of superpixel regions, analyzing the similarity of every two superpixel regions, and when the similarity is greater than a first critical value, distinguishing the corresponding every two superpixel regions into the same category;
s320: acquiring the gradient peak value of each pixel point in each category of super-pixel area;
s330: acquiring the gradient average of the super-pixel regions of each category;
s340: analyzing the category dependent index of each pixel point according to the gradient peak value of each pixel point and the gradient average number of the super pixel area where the pixel point is located;
s350: distinguishing all the pixel points in each category super-pixel region according to the category subordinate indexes of each pixel point;
s360: and obtaining the surface flaw area of the PVC sheet according to each super pixel area after the distinguishing.
Further, step S220 includes:
s221: setting a module coefficient of a loss function analysis guiding filtering algorithm, and analyzing the whole appearance graph information in the PVC sheet filtering graph according to the module coefficient;
s222: creating a pixel point characteristic data structure according to the gradient value of each pixel point in the PVC sheet filtering graph, and improving the loss function by using the pixel point characteristic data structure;
and S223, analyzing the improved module coefficient according to the improved loss function, and analyzing the overall appearance graphic information of the final PVC sheet according to the improved module coefficient.
Further, step S222 includes:
carrying out border acquisition on a PVC sheet filtering graph, obtaining PVC sheet gradient data graph information, acquiring the gradient value of each pixel point in the PVC sheet gradient data graph information, creating a pixel point characteristic data structure according to the gradient value of each pixel point, wherein the expression is as follows:
wherein θ (g) is a characteristic data structure of the pixel g, z (g) is a gradient value of the pixel g in the PVC sheet gradient data graphic information, z (b) is a gradient value of the pixel b in the PVC sheet gradient data graphic information, V is the number of the pixels in the PVC sheet gradient data graphic information, and σ is a module parameter for preventing denominator from being zero.
Further, step S223 includes:
s2231, the expression of the loss after improvement function is:
λ＝(c h G(g)+e h -M(g)) 2 ；
wherein, A (c) h ,e h ) Representing a loss function, G is a guide pattern, M represents a PVC sheet filter pattern, c h 、e h Is a partial window r for guiding graphic information to use pixel point h as center h Coefficient of module of r h The window comprises a pixel point g, the size of the window is r, theta (g) is a characteristic data structure of the g-th pixel point, and l is a regulation coefficient;
s2232, analyzing the expression of the overall appearance graphic information of the final PVC sheet as follows:
wherein D' represents the overall appearance graphic information of the final PVC sheet, c ' h 、e ' h Representing improved module coefficients, G (g) being a guide pattern, r h Is a window containing pixel points g with a size r, X is the window r h The number of pixels contained therein.
Further, step S310 includes:
s311: analyzing a gray scale co-occurrence matrix corresponding to the super pixel region according to the pixel value containing the pixel point in the super pixel region, and taking the gray scale co-occurrence matrix as a characteristic parameter matrix Y of the super pixel region o Analyzing the similarity of two super pixel regions according to the cosine similarity of each row vector in the characteristic parameter matrix of two super pixel regions, wherein the expression is as follows:
wherein the content of the first and second substances,
a characteristic parameter matrix O representing the super pixel region 1 and the super pixel region 2 1 ,O 2 Corresponding to a cosine approximation, U 1,2 Representing the approximation between superpixel region 1 and superpixel region 2, and o represents the o-th row vector in the feature parameter matrix.
Further, step S340 includes:
obtaining the gradient peak value of each pixel point in each superpixel region, obtaining the gradient average corresponding to the pixel block according to the gradient peak values of all the pixel points in each superpixel region, and analyzing the category dependent index of each pixel point in the superpixel region f:
wherein
A category dependent index, t, representing that the g-th pixel belongs to the f-th super-pixel region g The gradient peak value of the g-th pixel point is shown,
represents the mean of the gradients of the f-th superpixel region, exp is an exponential function with a natural constant e as the base.
Further, step S360 includes:
s361: when the category subordinate index of each pixel point in the obvious graphic information of the PVC sheet in the corresponding super-pixel region is greater than a preset critical value, distinguishing the pixel point into the corresponding super-pixel region;
s362: when the category subordinate index of the pixel point is lower than a preset critical value, analyzing the category subordinate indexes of the pixel point subordinate to other pixel blocks, and distinguishing the corresponding pixel point into the pixel block corresponding to the maximum index value in the category subordinate index sequence;
s363: and taking the super pixel region containing the most pixels as a normal pixel block, taking other super pixel regions as flaw pixel blocks on the surface layer of the PVC sheet, and taking the communication region corresponding to each flaw pixel block as a flaw region on the surface layer of the PVC sheet.
The beneficial results obtained by the invention are as follows:
1. according to the invention, the big data is used for replacing manual work to carry out quality inspection on the PVC sheets, so that the automation degree is greatly improved, the human errors are reduced, and the overall accuracy of the quality inspection is improved;
2. flaw quality inspection is carried out on the surface layer of the PVC sheet through graphic information data, the quality inspection process is full-automatic, no artificial contact is caused in the process, the surface layer can be prevented from being scratched, and secondary damage to the PVC sheet is reduced.
3. Quality inspection analysis is carried out based on graphic information data, obvious graphic information corresponding to the PVC sheet is obtained, differences between a flaw area and a normal area are improved, after super-pixel cutting is carried out on the obvious graphic information, subordinate categories of pixel points are further analyzed, therefore, the super-pixel areas are accurately distinguished, quality inspection efficiency is improved, and meanwhile, quality inspection accuracy of flaw conditions of the PVC sheet is greatly improved.
The foregoing description is only an overview of the technical aspects of the present invention, and in order to make the technical aspects of the present invention more clearly understood, the present invention may be implemented according to the contents of the illustration.
Drawings
FIG. 1 is a schematic view of the steps in this embodiment;
FIG. 2 is a schematic diagram illustrating the preprocessing method in step S3 according to the present embodiment;
fig. 3 is a schematic diagram of the quality inspection method in step S4 in this embodiment.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and specific embodiments, and it should be understood that the embodiments described below or technical features may be arbitrarily combined to form new embodiments without conflict. Unless defined otherwise, all technical and objective terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed names.
As shown in fig. 1-2, a method for inspecting quality of PVC sheets based on big data includes the following steps:
step S1: acquiring PVC sheet graphic information;
step S2: preprocessing the graphic information to remove the noise of the graphic information;
step S3: performing PVC sheet quality inspection according to the graphic information;
step S4: judging whether the defective PVC sheet can be repaired;
step S5: the PVC sheet after quality inspection is divided into three types of qualified quality inspection, unqualified quality inspection but repairable, unqualified quality inspection and unrepairable, and the three types are respectively transmitted to the next procedure through different transmission lines.
Further, the preprocessing method in step S2 is specifically as follows:
s210: filtering the graphic information to obtain filtered graphic information;
firstly, the method is to install a graphic information shooting device for obtaining the graphic information data of the PVC sheet surface layer, wherein the shooting area and the visual angle of a camera are processed by a worker according to specific conditions. It needs to explain that, camera shooting area can contain the PVC sheet of waiting to examine quality, so that carry out whole quality inspection to it, perhaps staff's mountable many cameras treat that quality inspection PVC sheet carries out the acquisition of graphical information, then carry out the later stage fusion processing with the graphical information data that adjacent camera obtained through graphical information post-processing software, in order to obtain the holistic graphical information data in top layer of waiting to examine quality PVC sheet, when staff adopts many cameras to treat that quality inspection PVC sheet carries out graphical information and obtains, the shooting area that need guarantee adjacent camera has the repetition region, so that carry out the overall graphical information data that the graphical information fusion obtained PVC sheet top layer in the follow-up.
After the graphical information of the surface layer of the PVC sheet is obtained, the method is mainly used for carrying out flaw quality inspection on the PVC sheet based on the graphical information data. In the process of acquiring the graphic information, considering that a large amount of noise exists in the environment and the dust on the surface layer of the PVC sheet can interfere the acquisition of the graphic information of the PVC sheet, the method performs filtering and denoising processing on the graphic information data of the PVC sheet to process the noise of the graphic information on the surface layer of the PVC sheet, and the method for denoising the graphic information comprises the following steps: and the average filtering method and the like can be used for automatically selecting a denoising method, so that the denoising processing of the pattern information on the surface layer of the PVC sheet is realized, and the denoised filtering pattern information is obtained.
S220: analyzing the integral appearance graph information in the PVC sheet filtering graph according to a guiding filtering algorithm;
the method considers the phenomena of external illumination, surface layer reflection of the PVC sheet and the like, and causes the obtained PVC sheet graphic information to have larger difference compared with the actual PVC sheet graphic information. In order to prevent the problems of reduced accuracy of the quality inspection of the PVC sheet defects caused by external phenomena, and the like, the method carries out explicit processing on the surface graphic information of the PVC sheet so as to improve the characteristic distinction of the surface defects of the PVC sheet, and the surface graphic information is used as graphic information data of the quality inspection of the PVC sheet defects so as to accurately obtain the defect area.
The graphic information data comprise overall appearance graphic information D and fine grain graphic information R, and in order to improve the significance processing result of the PVC sheet filtering graphics, the method separates the PVC sheet filtering graphics, respectively obtains the corresponding overall appearance graphic information and the corresponding fine grain graphic information, respectively performs different significance processing processes in the later stage, and improves the significance processing result of the graphic information. The method comprises the steps of firstly obtaining integral appearance graphic information data corresponding to a PVC sheet filtering graphic, and obtaining the integral appearance graphic information data of the PVC sheet.
According to the guided filtering, the graphic information data D corresponding to the whole appearance and the edge in the graphic information can be analyzed by a linear module, which is specifically expressed as:
η＝c h G(g)；
wherein G (g) is a guide graphic, D is overall appearance graphic information, c h ，e h Is a partial window r for guiding graphic information to use pixel point h as center h Coefficient of module of r h Is a window containing pixel points g, of size r. When the gradient function is taken for both sides of the derived graphic information,
it can be concluded that the partial linearity module can ensure that when there is a gradient in the graphics information G, the graphics information D will also have corresponding gradient data.
S230: acquiring PVC sheet fine grain pattern information according to the PVC sheet filtering pattern and the PVC sheet integral surface pattern information, and respectively performing significance processing on the PVC sheet integral surface pattern information and the PVC sheet fine grain pattern information;
s240: and obtaining the obvious graphic information of the PVC sheet according to the overall appearance graphic information of the PVC sheet and the fine grain graphic information of the PVC sheet after the significance processing.
Based on the PVC sheet filtering graph, analyzing fine grain graph information corresponding to the surface layer of the PVC sheet:
R＝M-D'；
wherein, R is the pattern information of the fine grain of the PVC sheet, M is the filtering pattern of the PVC sheet, and D' is the pattern information of the overall appearance of the final PVC sheet obtained after improvement.
In order to improve the accuracy of the quality inspection of the surface flaws of the PVC sheet, different significance processing methods are adopted for different graphic information data, and for the overall appearance graphic information, in order to make the inner fine and edge information more obvious, the histogram balancing method is adopted to process the graphic information so as to improve the difference between the brightest white and the darkest black in the light and shade area of the overall graphic information; for the fine grain pattern information, in order to improve the fine grain information in the fine grain pattern information, the method adopts a gamma conversion method to process the fine grain pattern information, wherein the value of a gamma conversion parameter is 0.4, so that the brightness information of darker fine grains in the fine grain pattern information is improved, and the uniformity degree of the brightness of the pattern information is improved.
Finally, the overall appearance graphic information D is processed based on the significance pq And fine grain pattern information R pq The method comprises the following steps of obtaining the final prominence-processed PVC sheet graphic information data:
M pq ＝D pq +R pq ；
in the formula, M pq (g) The obvious graphic information of the final explicitly processed PVC sheet can effectively highlight the distinguishing characteristics of pixel points in a flaw area and a normal area on the surface layer of the PVC sheet, and is favorable for carrying out accurate quality inspection on the flaw pixel points.
Further, the quality inspection method in step S3 includes the following steps:
s310: carrying out superpixel cutting on the obvious graphic information of the PVC sheet to obtain a plurality of superpixel regions, analyzing the similarity of every two superpixel regions, and when the similarity is greater than a first critical value, dividing every two corresponding superpixel regions into the same category;
for each super pixel region, the method acquires the characteristic parameters of the pixel block of the super pixel region, analyzes the gray scale co-occurrence matrix corresponding to the super pixel region according to the pixel value containing the pixel point in the super pixel region, and performs gray scale grade division on the pixel value of the pixel point in the obvious graphic information of the PVC sheet to reduce the software analysis amount, wherein the gray scale grade of the graphic information is divided into 9 grades, the dimension of the gray scale co-occurrence matrix corresponding to each super pixel region is 9 x 9, and the gray scale co-occurrence matrix is used as the characteristic parameter matrix O of the pixel block h And the characteristic parameter matrix representing the super pixel area h is used for representing the pixel distribution condition in the super pixel area.
S320: acquiring the gradient peak value of each pixel point in each category of super-pixel area;
s330: acquiring the gradient average of the super-pixel regions of each category;
s340: analyzing the category dependent index of each pixel point according to the gradient peak value of each pixel point and the gradient average number of the super pixel area where the pixel point is located;
in order to further improve the quality inspection accuracy of defective pixel points, the method analyzes each pixel block obtained after final combination.
The method for analyzing the category dependent index of each pixel point comprises the following steps:
obtaining the gradient average corresponding to the pixel block according to the gradient peak value of all the pixel points in each superpixel region, and analyzing the category dependent index of each pixel point in the superpixel region f:
wherein
A category dependent index, t, representing that the g-th pixel belongs to the f-th super-pixel region g The gradient peak value of the ith pixel point is shown,
represents the average of the gradients of the f-th super-pixel region.
S350: distinguishing all the pixel points in each category super-pixel region according to the category subordinate indexes of each pixel point;
s360: and obtaining the surface flaw area of the PVC sheet according to each super pixel area after the distinguishing.
The method for distinguishing the corresponding categories of all pixel points in the obvious graphic information of the PVC sheet to obtain the flaw area on the surface layer of the PVC sheet comprises the following steps:
when the category subordinate index of each pixel point in the obvious graphic information of the PVC sheet in the corresponding super-pixel region is greater than a preset critical value, distinguishing the pixel point into the corresponding super-pixel region;
when the category subordinate index of the existing pixel point is lower than a preset critical value, analyzing the category subordinate indexes of the pixel point subordinate to other pixel blocks, and distinguishing the corresponding pixel point into a super pixel area corresponding to the maximum index value in the category subordinate index sequence;
when the category membership index is lower than 0.7, the method considers that the category of the corresponding pixel point is undetermined and needs to be analyzed to realize accurate distinguishing.
And taking the super pixel region containing the most pixels as a normal pixel block, taking other super pixel regions as flaw pixel blocks on the surface layer of the PVC sheet, and taking the communication region corresponding to each flaw pixel block as a flaw region on the surface layer of the PVC sheet.
Under normal conditions, the flaws on the surface layer of the prior PVC sheet are various types of small-area flaws, therefore, the method takes the pixel blocks containing the most pixels as normal pixel blocks corresponding to the normal areas on the surface layer of the PVC sheet, other pixel blocks as flaw pixel blocks on the surface layer of the PVC sheet, and the communication areas corresponding to the flaw pixel blocks are marked as flaw communication areas, so that quality inspection and identification of the flaws on the surface layer of the PVC sheet are realized.
Further, step S220 includes:
s221: setting a module coefficient of a loss function analysis guiding filtering algorithm, and analyzing the whole appearance graph information in the PVC sheet filtering graph according to the module coefficient;
s222: creating a pixel point characteristic data structure (or descriptor) according to the gradient value of each pixel point in the PVC sheet filtering graph, and improving the loss function by using the pixel point characteristic data structure;
and S223, analyzing the improved module coefficient according to the improved loss function, and analyzing the overall appearance graphic information of the final PVC sheet according to the improved module coefficient.
In order to highlight the fine grain information in the graphic information, the embodiment first obtains the edge information of the PVC sheet filtering graphic through an edge quality inspection operator, the edge quality inspection operator has many edge quality inspection operators and can be selected by a worker, and the preferred method obtains the PVC sheet filtering graphic by using the sobel operator to obtain the corresponding gradient data graphic information Z, Z (g) representing the gradient value at the gradient data graphic information pixel point g.
Further, step S222 includes:
carrying out border acquisition on a PVC sheet filtering graph, obtaining PVC sheet gradient data graph information, acquiring the gradient value of each pixel point in the PVC sheet gradient data graph information, creating a pixel point characteristic data structure according to the gradient value of each pixel point, wherein the expression is as follows:
wherein θ (g) is a characteristic data structure of the pixel g, z (g) is a gradient value of the pixel g in the PVC sheet gradient data graphic information, z (b) is a gradient value of the pixel b in the PVC sheet gradient data graphic information, V is the number of the pixels in the PVC sheet gradient data graphic information, and σ is a module parameter for preventing denominator from being zero.
According to the characteristic data structure of the pixel points established by the method, when the pixel points are smooth pixel points, the corresponding characteristic data structure is less than 1, and when the pixel points are gradient change pixel points, namely edge points, the characteristic data structure installed by the method is more than 1, and different pixel points can be processed and analyzed in a targeted manner based on the characteristic information of the pixel points.
Further, step S223 includes:
s2231, the expression of the loss after improvement function is:
λ＝(c h G(g)+e h -M(g)) 2 ；
wherein, A (c) h ,e h ) Representing a loss function, G being a guide mapForm, M represents a PVC sheet filter pattern, c h 、e h Is a partial window r for guiding graphic information to use pixel point h as center h Coefficient of module of r h The window comprises a pixel point g, the size of the window is r, theta (g) is a characteristic data structure of the g-th pixel point, and l is a regulation coefficient;
adjusting the loss function of the guiding filtering through the characteristic data structures of different pixel points, further realizing the improvement of the module coefficient and obtaining the improved module coefficient c' h 、e′ h The edge appearance information in the graphic information is more highlighted, and the overall appearance graphic information of the PVC sheet is more accurately acquired.
The module coefficients are resolved by curve fitting (least squares) to obtain:
wherein epsilon h 、ψ h Respectively as guide patterns in the window r h Mean and variance of gray levels, M, of internal pixels h For inputting the gray-scale mean value of the image information in the pixel point in the window, X is the window r h The number of pixels contained therein.
S2232, analyzing the expression of the overall appearance graphic information of the final PVC sheet as follows:
wherein D ' represents the overall appearance graphic information, c ' of the final PVC sheet ' h 、e′ h Representing improved module coefficients, G (g) being a guide pattern, r h Is a window containing pixel points g with a size r, X is the window r h The number of pixels contained therein.
For promoting the significance processing result of the PVC sheet filtering graph, the embodiment improves the guided filtering step, the guided filtering step uses the same control coefficient l for all the pixel points, and the difference between the pixel points is not considered, so that in order to better highlight the appearance information in the graph information and maintain more complete edge information in the graph information, the embodiment improves the loss function of the guided filtering.
Further, step S310 includes:
s311: analyzing a gray scale co-occurrence matrix corresponding to the super pixel region according to the pixel value containing the pixel point in the super pixel region, and taking the gray scale co-occurrence matrix as a characteristic parameter matrix Y of the super pixel region o Analyzing the similarity of two super pixel regions according to the cosine similarity of each row vector in the characteristic parameter matrix of two super pixel regions, wherein the expression is as follows:
wherein the content of the first and second substances,
a characteristic parameter matrix O representing the super pixel region 1 and the super pixel region 2 1 ,O 2 Corresponding to the cosine approximation, U, of the first line vector of 1,2 Representing the approximation between superpixel region 1 and superpixel region 2, and o represents the o-th row vector in the feature parameter matrix.
The method will set a degree threshold value U for the approximation S When the similarity of any two super-pixel regions is not higher than U, the method repeats the steps until the similarity of any two super-pixel regions is not higher than U S When the classification processing of the super pixel region is stopped. Therefore, each pixel block after final classification can be obtained, and pixel point differentiation of different classes is realized.
Further, step S340 includes:
obtaining the gradient peak value of each pixel point in each superpixel region, obtaining the gradient average corresponding to the pixel block according to the gradient peak values of all the pixel points in each superpixel region, and analyzing the category dependent index of each pixel point in the superpixel region f:
wherein
A category dependent index, t, representing that the g-th pixel belongs to the f-th super-pixel region g The gradient peak value of the g-th pixel point is shown,
represents the mean of the gradients of the f-th superpixel region, exp is an exponential function with a natural constant e as the base.
Further, step S360 includes:
s361: when the category subordinate index of each pixel point in the obvious graphic information of the PVC sheet in the corresponding super-pixel region is greater than a preset critical value, distinguishing the pixel point into the corresponding super-pixel region;
s362: when the category subordinate index of the pixel point is lower than a preset critical value, analyzing the category subordinate indexes of the pixel point subordinate to other pixel blocks, and distinguishing the corresponding pixel point into the pixel block corresponding to the maximum index value in the category subordinate index sequence;
s363: and taking the super pixel region containing the most pixels as a normal pixel block, taking other super pixel regions as flaw pixel blocks on the surface layer of the PVC sheet, and taking the communication region corresponding to each flaw pixel block as a flaw region on the surface layer of the PVC sheet.
The above embodiments are only preferred embodiments of the present invention, and the scope of the present invention should not be limited thereby, and all the unrealistic transformations and substitutions made by those skilled in the art on the basis of the present invention are included in the scope of the present invention as claimed.
Claims (9)
1. A PVC sheet quality inspection method based on big data is characterized in that: the method comprises the following steps:
step S1: acquiring PVC sheet graphic information;
step S2: preprocessing the graphic information;
step S3: performing PVC sheet quality inspection according to the graphic information;
step S4: judging whether the defective PVC sheet can be repaired;
step S5: the PVC sheet after quality inspection is divided into three types of qualified quality inspection, unqualified quality inspection but repairable, unqualified quality inspection and unrepairable, and the three types are respectively transmitted to the next procedure through different transmission lines.
2. The method for inspecting the quality of the PVC sheet based on the big data as claimed in claim 1, is characterized in that: the preprocessing method in step S2 includes the following steps:
s210: filtering the graphic information to obtain filtered graphic information;
s220: analyzing the integral appearance graph information in the PVC sheet filtering graph according to a guiding filtering algorithm;
s230: acquiring PVC sheet fine grain pattern information according to the PVC sheet filtering pattern and the PVC sheet integral surface pattern information, and respectively performing significance processing on the PVC sheet integral surface pattern information and the PVC sheet fine grain pattern information;
s240: and obtaining the obvious graphic information of the PVC sheet according to the overall appearance graphic information of the PVC sheet and the fine grain graphic information of the PVC sheet after the significance processing.
3. The method for inspecting the quality of the PVC sheet based on the big data as claimed in claim 2, characterized in that: the quality inspection method in step S3 includes the following steps:
s310: carrying out superpixel cutting on the obvious graphic information of the PVC sheet to obtain a plurality of superpixel regions, analyzing the similarity of every two superpixel regions, and when the similarity is greater than a first critical value, dividing every two corresponding superpixel regions into the same category;
s320: acquiring the gradient peak value of each pixel point in each category of super-pixel area;
s330: acquiring the gradient average of the super-pixel regions of each category;
s340: analyzing the category dependent index of each pixel point according to the gradient peak value of each pixel point and the gradient average number of the super pixel area where the pixel point is located;
s350: distinguishing all the pixel points in each category super-pixel region according to the category subordinate indexes of each pixel point;
s360: and obtaining the surface flaw area of the PVC sheet according to each super pixel area after the distinguishing.
4. The method for inspecting the quality of the PVC sheet based on the big data as claimed in claim 3, is characterized in that:
step S220 includes:
s221: setting a module coefficient of a loss function analysis guiding filtering algorithm, and analyzing the whole appearance graph information in the PVC sheet filtering graph according to the module coefficient;
s222: creating a pixel point characteristic data structure according to the gradient value of each pixel point in the PVC sheet filtering graph, and improving the loss function by using the pixel point characteristic data structure;
and S223, analyzing the improved module coefficient according to the improved loss function, and analyzing the overall appearance graphic information of the final PVC sheet according to the improved module coefficient.
5. The quality inspection method of the PVC sheet based on the big data as claimed in claim 4, characterized in that:
step S222 includes:
carrying out border acquisition on a PVC sheet filtering graph, obtaining PVC sheet gradient data graph information, acquiring the gradient value of each pixel point in the PVC sheet gradient data graph information, creating a pixel point characteristic data structure according to the gradient value of each pixel point, wherein the expression is as follows:
wherein θ (g) is a characteristic data structure of the pixel g, z (g) is a gradient value of the pixel g in the PVC sheet gradient data graphic information, z (b) is a gradient value of the pixel b in the PVC sheet gradient data graphic information, V is the number of the pixels in the PVC sheet gradient data graphic information, and σ is a module parameter for preventing denominator from being zero.
6. The quality inspection method of the PVC sheet based on the big data as claimed in claim 5, characterized in that:
step S223 includes:
s2231, the expression of the loss after improvement function is:
λ＝(c h G(g)+e h -M(g)) 2 ；
wherein, A (c) h ,e h ) Representing a loss function, G is a guide pattern, M represents a PVC sheet filter pattern, c h 、e h Is a partial window r for guiding graphic information to use pixel point h as center h Coefficient of module of r h The window comprises a pixel point g, the size of the window is r, theta (g) is a characteristic data structure of the g-th pixel point, and l is a regulation coefficient;
s2232, analyzing the expression of the overall appearance graphic information of the final PVC sheet as follows:
θ＝c' h G(g)+e' h ；
wherein D ' represents the overall exterior graphic information, c ' of the final PVC sheet ' h 、e′ h Representing improved module coefficients, G (g) being a guide pattern, r h Is a window containing pixel points g with a size r, X is the window r h The number of pixels contained therein.
7. The method for inspecting the quality of the PVC sheet based on the big data as claimed in claim 6, characterized in that:
step S310 includes:
s311: analyzing a gray scale co-occurrence matrix corresponding to the super pixel region according to the pixel value containing the pixel point in the super pixel region, and taking the gray scale co-occurrence matrix as a characteristic parameter matrix Y of the super pixel region o Analyzing the similarity of two super pixel regions according to the cosine similarity of each row vector in the characteristic parameter matrix of two super pixel regions, wherein the expression is as follows:
wherein the content of the first and second substances,
a characteristic parameter matrix O representing the super pixel region 1 and the super pixel region 2 1 ,O 2 Corresponding to the cosine approximation, U, of the first line vector of 1,2 Representing the approximation between superpixel region 1 and superpixel region 2, and o represents the o-th row vector in the feature parameter matrix.
8. The method for inspecting the quality of the PVC sheet based on the big data as claimed in claim 7, is characterized in that:
step S340 includes:
obtaining the gradient peak value of each pixel point in each superpixel region, obtaining the gradient average corresponding to the pixel block according to the gradient peak values of all the pixel points in each superpixel region, and analyzing the category dependent index of each pixel point in the superpixel region f:
wherein
A category dependent index, t, representing that the g-th pixel belongs to the f-th super-pixel region g The gradient peak value of the g-th pixel point is shown,
represents the mean of the gradients of the f-th superpixel region, exp is an exponential function with a natural constant e as the base.
9. The method for inspecting the quality of the PVC sheet based on the big data as claimed in claim 8, characterized in that: step S360 includes:
s361: when the category subordinate index of each pixel point in the obvious graphic information of the PVC sheet in the corresponding super-pixel region is greater than a preset critical value, distinguishing the pixel point into the corresponding super-pixel region;
s362: when the category subordinate index of the pixel point is lower than a preset critical value, analyzing the category subordinate indexes of the pixel point subordinate to other pixel blocks, and distinguishing the corresponding pixel point into the pixel block corresponding to the maximum index value in the category subordinate index sequence;
s363: and taking the super pixel region containing the most pixels as a normal pixel block, taking other super pixel regions as flaw pixel blocks on the surface layer of the PVC sheet, and taking the communication region corresponding to each flaw pixel block as a flaw region on the surface layer of the PVC sheet.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210680881.0A CN115078383A (en) | 2022-06-15 | 2022-06-15 | PVC sheet quality inspection method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210680881.0A CN115078383A (en) | 2022-06-15 | 2022-06-15 | PVC sheet quality inspection method based on big data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115078383A true CN115078383A (en) | 2022-09-20 |
Family
ID=83253984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210680881.0A Pending CN115078383A (en) | 2022-06-15 | 2022-06-15 | PVC sheet quality inspection method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115078383A (en) |
-
2022
- 2022-06-15 CN CN202210680881.0A patent/CN115078383A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107845087B (en) | Method and system for detecting uneven brightness defect of liquid crystal panel | |
| CN110389127B (en) | System and method for identifying metal ceramic parts and detecting surface defects | |
| CN107437243B (en) | Tire impurity detection method and device based on X-ray image | |
| CN116758061A (en) | Casting surface defect detection method based on computer vision | |
| CN103792699A (en) | TFT-LCD Mura defect machine vision detecting method based on B spline surface fitting | |
| CN116630813B (en) | Highway road surface construction quality intelligent detection system | |
| CN109598681B (en) | No-reference quality evaluation method for image after repairing of symmetrical Thangka | |
| CN113177924A (en) | Industrial production line product flaw detection method | |
| CN116990323B (en) | High-precision printing plate visual detection system | |
| CN113935666B (en) | Building decoration wall tile abnormity evaluation method based on image processing | |
| CN114926407A (en) | Steel surface defect detection system based on deep learning | |
| CN116152242B (en) | Visual detection system of natural leather defect for basketball | |
| CN116310845B (en) | Intelligent monitoring system for sewage treatment | |
| CN115131354A (en) | Laboratory plastic film defect detection method based on optical means | |
| CN116091505B (en) | Automatic defect detection and classification method and system for sapphire substrate | |
| CN114972230A (en) | Method and system for detecting defects of circuit board on production line | |
| CN116883408A (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
| CN117094975A (en) | Method and device for detecting surface defects of steel and electronic equipment | |
| TW201512649A (en) | Method of chip detects inspecting, system therefor, and computer program product thereof | |
| CN113808104A (en) | Block-based metal surface defect detection method and system | |
| CN116797602A (en) | Surface defect identification method and device for industrial product detection | |
| CN115078383A (en) | PVC sheet quality inspection method based on big data | |
| CN112683902B (en) | Die-cutting product surface defect online detection method based on template | |
| CN114881952A (en) | Casting defect area detection and edge accurate segmentation method | |
| CN114486916A (en) | Mobile phone glass cover plate defect detection method based on machine vision |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220920 |