CN106504251B - A detection method for cubic texture content of electronic aluminum foil based on image processing - Google Patents
A detection method for cubic texture content of electronic aluminum foil based on image processing Download PDFInfo
- Publication number
- CN106504251B CN106504251B CN201610843102.9A CN201610843102A CN106504251B CN 106504251 B CN106504251 B CN 106504251B CN 201610843102 A CN201610843102 A CN 201610843102A CN 106504251 B CN106504251 B CN 106504251B
- Authority
- CN
- China
- Prior art keywords
- aluminium foil
- picture
- pixel
- cubic texture
- spliced
- 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
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
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of electric aluminum foil cubic texture detection method of content based on picture processing, can be improved the accuracy rate of cubic texture content detection.The described method includes: reading multiple aluminium foil pictures that camera is continuously shot, and multiple aluminium foil pictures described in shooting are spliced into an aluminium foil picture；Extract the aluminium foil in spliced aluminium foil picture；Extract the non-cubic texture in spliced aluminium foil picture；The number of pixel shared by the number for counting pixel shared by the aluminium foil extracted respectively and the non-cubic texture extracted, according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture, content of the cubic texture in aluminium foil is obtained.The present invention is suitable for electric aluminum foil production technical field.
Description
Technical field
The present invention relates to electric aluminum foil production technical fields, particularly relate to a kind of electric aluminum foil cube based on picture processing
Texture detection method of content.
Background technique
The research of China's electrolytic capacitor aluminium foil and production development are later, and quality stability is poor.To adapt to production electric aluminum foil
It needs, preferably adjustment and control technical process, and meets its testing requirements, establish electric aluminum foil surface { 100 } cubic plane
The method of texture occupation rate is very necessary.
To the recrystallization texture research in the deformation texture and heat treatment process of material processing, with pole figure method, antipole
The determination methods uses such as figure method, orientation distribution function method are more universal, though these methods are accurate but complex, it is difficult to
It is applicable in industry spot；And spar method, etch pit method texture test and analyze, as a kind of easy-to-use method in high-pressure electronic aluminium
Foil is recrystallized using more generally in the measurement of cubic texture, and the application of especially spar degree determination method obtains at home
Universal, spar degree is a kind of indirect method for measuring cubic texture in aluminium foil: since the corrosion resistance of cubic texture is good, particular agent
Non-cubic texture shows brilliant white point on aluminium foil surface after etch, by artificial observation, that is, brilliant white point size, then
The number of side light cubic texture content.
In the prior art, most of production line determines cubic texture content using the method for artificial observation, only one kind half
Quantitative method needs to have certain experiences basis, can not accurately provide as a result, result accuracy rate is low.
Summary of the invention
The technical problem to be solved in the present invention is to provide a kind of electric aluminum foil cubic texture content inspections based on picture processing
Survey method determines cubic texture content to solve the method present in the prior art using artificial observation, and as a result accuracy rate is low
The problem of.
In order to solve the above technical problems, the embodiment of the present invention provides a kind of electric aluminum foil cubic texture based on picture processing
Detection method of content, comprising:
Multiple aluminium foil pictures that camera is continuously shot are read, and multiple aluminium foil pictures described in shooting are spliced into an aluminium
Foil picture；
Extract the aluminium foil in spliced aluminium foil picture；
Extract the non-cubic texture in spliced aluminium foil picture；
Picture shared by the number for counting pixel shared by the aluminium foil extracted respectively and the non-cubic texture extracted
The number of element is obtained according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture
Content of the cubic texture in aluminium foil.
Further, described to read multiple aluminium foil pictures for being continuously shot of camera, and by multiple aluminium foil figures described in shooting
Piece is spliced into an aluminium foil picture
Read multiple aluminium foil pictures that camera is continuously shot；
It is registrated based on the method for template matching every two to shooting adjacent aluminium foil pictures；
Every two adjacent aluminium foil pictures after registration are merged using the method for weighting gradual change, the institute shot
State the splicing picture of multiple aluminium foil pictures.
Further, described to carry out registration packet based on the method for template matching every two to shooting adjacent aluminium foil pictures
It includes:
Every two to shooting adjacent aluminium foil pictures, choose wherein that an aluminium foil picture is as reference picture, and another
Aluminium foil picture is as picture subject to registration；
Determine the area-of-interest in the reference picture as template；
Calculate the absolute value summation of gray value difference between the template and the picture subject to registration:
Wherein, T indicates template, the number of pixel in n expression template, u, the offset on the direction v x, y, t (x, y)
For the gray value for the pixel that coordinate in template is (x, y), (u+x, v+y) is the coordinate after (x, the y) amount of offsetting, f (u+x, v+
Y) coordinate is the gray value of the pixel of (u+x, v+y), sad in template when being template movement to picture current location subject to registration
(u, v) is a scalar value being calculated based on the gray value, the similarity of expression template and picture subject to registration, sad (u,
V) matching is completed when minimum.
Further, the method using weighting gradual change melts every two adjacent aluminium foil pictures after registration
It closes, the splicing picture of multiple the aluminium foil pictures shot includes:
Utilize formula I (x, y)=I1(x,y)(1-σ)+I2(x, y) merges every two adjacent pictures after registration,
The splicing picture of multiple the aluminium foil pictures shot；
Wherein, I1(x, y) and I2(x, y) respectively indicates the ash of each pixel of to be spliced two adjacent picture after registration
Angle value, I (x, y) indicate I1(x, y) and I2The gray value of the pixel of overlapping region in (x, y), weighting coefficient σ ∈ (0,1), works as σ
When changing to 1 by 0, picture is from I1(x, y) is transitioned into I2(x, y).
Further, the aluminium foil extracted in spliced aluminium foil picture includes:
Binary conversion treatment is carried out to spliced aluminium foil picture, so that the aluminium foil part in the spliced aluminium foil picture
White is presented, black is presented in non-aluminium foil part；
Each regard the non-adjacent point of each of aluminium foil picture after binary conversion treatment as a connected region, counts
The quantity of pixel shared by each connected region, judges whether the quantity of pixel shared by each connected region is less than preset third threshold
Value weeds out the connected region for being less than preset third threshold value if being less than preset third threshold value.
Further, described that binary conversion treatment is carried out to spliced aluminium foil picture, so that the spliced aluminium foil figure
White is presented in aluminium foil part in piece, non-aluminium foil part is presented black and includes:
Judge whether the gray value of the pixel in spliced aluminium foil picture is greater than preset first threshold；
If more than preset first threshold, then the gray value that will be greater than the pixel of preset first threshold is set as 255, institute
It states pixel and white is presented；
Otherwise, the gray value of the pixel is set as 0, black is presented in the pixel.
Further, the preset first threshold is to make inter-class variance g=ω0ω1(μ0-μ1) ^2 is when reaching maximum
The segmentation threshold of foreground and background, g are inter-class variance, ω0The ratio of the spliced aluminium foil picture of whole picture is accounted for for foreground pixel points
Example, μ0For the average gray of the pixel of prospect, ω1The ratio of the spliced aluminium foil picture of whole picture, μ are accounted for for background pixel points1
For the average gray of the pixel of background.
Further, the non-cubic texture extracted in spliced aluminium foil picture includes:
Top cap conversion process is carried out to spliced aluminium foil picture；
Binary conversion treatment is carried out to the aluminium foil picture after top cap conversion process, the brilliant white point of non-cubic texture will be represented never
Initial gross separation comes out in uniform background；
The quantity for counting pixel shared by each white connected region judges the quantity of pixel shared by each white connected region
Whether preset 4th threshold value is greater than, the white connected region that will be greater than preset 4th threshold value is weeded out, will be represented non-cubic
The brilliant white point of texture is accurately separated from background；
Wherein, the white connected region weeded out be as it is large stretch of it is reflective caused by white area.
Further, the aluminium foil picture after the conversion process to top cap carries out binary conversion treatment, will represent non-cubic knit
The initial gross separation from non-uniform background of the brilliant white point of structure part, which comes out, includes:
Whether the gray value of the pixel in aluminium foil picture after judging top cap conversion process is greater than preset second threshold；
If more than preset second threshold, then the gray value that will be greater than the pixel of preset second threshold is set as 255, institute
It states pixel and white is presented；
Otherwise, the gray value of the pixel is set as 0, black is presented in the pixel；
Wherein, the preset second threshold is determined by maximum variance between clusters.
Further, it the number for counting pixel shared by the aluminium foil extracted respectively and extracts described non-vertical
The number of pixel shared by square texture, according to pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture
Number, obtaining content of the cubic texture in aluminium foil includes:
The number for counting pixel shared by the aluminium foil extracted is S1；
The number of pixel shared by white area caused by the sheet that statistics weeds out is reflective is S2；
Statistics represents the number of pixel shared by the brilliant white point of non-cubic texture as S3；
Determine that content of the non-cubic texture in aluminium foil is S3/ (S1-S2)；
According to content of the determining non-cubic texture in spliced aluminium foil picture, cubic texture is obtained in aluminium foil
Content is 1-S3/ (S1-S2).
The advantageous effects of the above technical solutions of the present invention are as follows:
In above scheme, by reading multiple aluminium foil pictures for being continuously shot of camera, and by multiple aluminium foils described in shooting
Picture is spliced into an aluminium foil picture；Extract the aluminium foil in spliced aluminium foil picture；Extract spliced aluminium foil picture
In non-cubic texture；The number of pixel shared by the aluminium foil extracted and the non-cubic texture extracted are counted respectively
The number of shared pixel, according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture
Mesh obtains content of the cubic texture in aluminium foil.In this way, substituting eye-observation with machine vision, cubic texture accounting is carried out
Quantitative analysis can efficiently, automatically, intelligently realize the detection of cubic texture content, to be effectively reduced human input, subtract
Few detection time, and can ensure the accuracy rate of detection provides more accurately quality for the production technology of electric aluminum foil product
Information feedback.
Detailed description of the invention
Fig. 1 is the stream of the electric aluminum foil cubic texture detection method of content provided in an embodiment of the present invention based on picture processing
Journey schematic diagram；
Fig. 2 is obtained aluminium foil picture schematic diagram after splicing provided in an embodiment of the present invention；
Fig. 3 is the picture schematic diagram for the aluminium foil part that will tentatively extract after Fig. 2 binary conversion treatment；
Fig. 4 comes to obtain aluminium foil after the removal noise provided in an embodiment of the present invention using connected domain with background full distinguished
Picture schematic diagram；
Fig. 5 is the figure provided in an embodiment of the present invention for tentatively extracting non-cubic texture with binary conversion treatment using top cap transformation
Piece schematic diagram；
Fig. 6 be it is provided in an embodiment of the present invention reject large stretch of light-reflecting white region using connected region screening after finally obtain
Picture schematic diagram.
Specific embodiment
To keep the technical problem to be solved in the present invention, technical solution and advantage clearer, below in conjunction with attached drawing and tool
Body embodiment is described in detail.
The present invention determines cubic texture content for the existing method using artificial observation, and as a result accuracy rate is low asks
Topic provides a kind of electric aluminum foil cubic texture detection method of content based on picture processing.
Embodiment one
As shown in Figure 1, the electric aluminum foil cubic texture content detection side provided in an embodiment of the present invention based on picture processing
Method, comprising:
S101 reads multiple aluminium foil pictures that camera is continuously shot, and multiple aluminium foil pictures described in shooting is spliced into
One aluminium foil picture；
S102 extracts the aluminium foil in spliced aluminium foil picture；
S103 extracts the non-cubic texture in spliced aluminium foil picture；
S104 counts the number of pixel shared by the aluminium foil extracted and the non-cubic texture institute extracted respectively
The number for accounting for pixel, according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture,
Obtain content of the cubic texture in aluminium foil.
Electric aluminum foil cubic texture detection method of content based on picture processing described in the embodiment of the present invention, passes through reading
Multiple aluminium foil pictures that camera is continuously shot, and multiple aluminium foil pictures described in shooting are spliced into an aluminium foil picture；It extracts
Aluminium foil in spliced aluminium foil picture out；Extract the non-cubic texture in spliced aluminium foil picture；Statistics is extracted respectively
The number of pixel shared by the number of pixel shared by the aluminium foil out and the non-cubic texture extracted, according to the institute of statistics
The number for stating pixel shared by the number and the non-cubic texture of pixel shared by aluminium foil, obtains cubic texture containing in aluminium foil
Amount.In this way, substituting eye-observation with machine vision, quantitative analysis is carried out to cubic texture accounting, it can efficient, automatic, intelligence
The detection of cubic texture content is realized on ground, to be effectively reduced human input, reduce detection time, and can ensure detection
Accuracy rate provides more accurately quality information feedback for the production technology of electric aluminum foil product.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Step ground, described multiple aluminium foil pictures for reading camera and being continuously shot, and multiple aluminium foil pictures described in shooting are spliced into one
Aluminium foil picture includes:
Read multiple aluminium foil pictures that camera is continuously shot；
It is registrated based on the method for template matching every two to shooting adjacent aluminium foil pictures；
Every two adjacent aluminium foil pictures after registration are merged using the method for weighting gradual change, the institute shot
State the splicing picture of multiple aluminium foil pictures.
In the present embodiment, multiple aluminium foil pictures that camera is continuously shot are read, first to avoid two adjacent picture parts area
Domain overlapping impacts late detection precision, first with template matching based method to every two adjacent aluminium foils of shooting
Picture is registrated, then is merged every two adjacent aluminium foil pictures after registration with the method for weighting gradual change, final real
Multiple the aluminium foil pictures now shot it is complete seamless spliced, as a result as shown in Fig. 2, in Fig. 2,1 indicate aluminium foil, 2 indicate back
Scape, 3 indicate light-reflecting white region, and 4 indicate non-cubic texture.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Step ground, it is described based on the method for template matching every two to shooting adjacent aluminium foil pictures carry out registration include:
Every two to shooting adjacent aluminium foil pictures, choose wherein that an aluminium foil picture is as reference picture, and another
Aluminium foil picture is as picture subject to registration；
Determine the area-of-interest in the reference picture as template；
Calculate the absolute value summation of gray value difference between the template and the picture subject to registration:
Wherein, T indicates template, the number of pixel in n expression template, u, the offset on the direction v x, y, t (x, y)
For the gray value for the pixel that coordinate in template is (x, y), (u+x, v+y) is the coordinate after (x, the y) amount of offsetting, f (u+x, v+
Y) coordinate is the gray value of the pixel of (u+x, v+y), sad in template when being template movement to picture current location subject to registration
(u, v) is a scalar value being calculated based on the gray value, the similarity of expression template and picture subject to registration, sad (u,
V) the smaller expression template of value is more similar to Target Photo, and matching is completed when sad (u, v) is minimum.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Every two adjacent aluminium foil pictures after registration are merged, are shot by step ground, the method using weighting gradual change
The splicing picture of multiple aluminium foil pictures includes:
Utilize formula I (x, y)=I1(x,y)(1-σ)+I2(x, y) merges every two adjacent pictures after registration,
The splicing picture of multiple the aluminium foil pictures shot；
Wherein, I1(x, y) and I2(x, y) respectively indicates the ash of each pixel of to be spliced two adjacent picture after registration
Angle value, I (x, y) indicate I1(x, y) and I2The gray value of the pixel of overlapping region in (x, y), weighting coefficient σ ∈ (0,1), works as σ
When slowly changing to 1 by 0, picture is from I1(x, y) is slowly transitioned into I2(x, y) realizes the smooth transition between picture, to disappear
It is final to realize the complete seamless spliced of multiple the aluminium foil pictures shot in addition to the trace of splicing.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Step ground, the aluminium foil extracted in spliced aluminium foil picture include:
Binary conversion treatment is carried out to spliced aluminium foil picture, so that the aluminium foil part in the spliced aluminium foil picture
White is presented, black is presented in non-aluminium foil part；
Each regard the non-adjacent point of each of aluminium foil picture after binary conversion treatment as a connected region, counts
The quantity of pixel shared by each connected region, judges whether the quantity of pixel shared by each connected region is less than preset third threshold
Value weeds out the connected region for being less than preset third threshold value if being less than preset third threshold value.
In the present embodiment, in order to carry out binary conversion treatment to spliced aluminium foil picture, threshold value (first threshold need to be determined
Value), the aluminium foil part that the gray value of pixel in spliced aluminium foil picture is higher than the first threshold is set to 255, is presented
For white；Non- aluminium foil part by the gray value of pixel in spliced aluminium foil picture not higher than the first threshold is set to 0,
It is rendered as black.
It is described that binary conversion treatment is carried out to spliced aluminium foil picture as an alternative embodiment in the present embodiment, so that
White is presented in aluminium foil part in the spliced aluminium foil picture, non-aluminium foil part is presented black and includes:
Judge whether the gray value of the pixel in spliced aluminium foil picture is greater than preset first threshold；
If more than preset first threshold, then the gray value that will be greater than the pixel of preset first threshold is set as 255, institute
It states pixel and white is presented；
Otherwise, the gray value of the pixel is set as 0, black is presented in the pixel.
In the present embodiment, the first threshold can be used maximum variance between clusters and obtain to picture processing:
For spliced aluminium foil picture I (x, y), the segmentation threshold of prospect (i.e. target) and background is denoted as T, belongs to prospect
The number of pixel account for the ratio of the spliced aluminium foil picture of whole picture and be denoted as ω0, average gray is denoted as μ0；Background pixel point
Number account for the ratio of the spliced aluminium foil picture of whole picture and be denoted as ω1, average gray is denoted as μ1；Spliced aluminium foil picture
Overall average gray scale is denoted as μ, and inter-class variance is denoted as g；Assuming that the background of spliced aluminium foil picture is darker, and spliced aluminium foil
The size of picture is M × N, and number of pixels of the gray value of pixel less than threshold value T is denoted as N in spliced aluminium foil picture0, pixel
Number of pixels of the gray value of point more than or equal to threshold value T is denoted as N1, then have:
ω0=N0/M×N (1)
ω1=N1/M×N (2)
N0+N1=M × N (3)
ω0+ω1=1 (4)
μ=ω0*μ0+ω1*μ1 (5)
G=ω0(μ0-μ)^2+ω1(μ1-μ)^2 (6)
Formula (5) are substituted into formula (6), obtain equivalence formula:
G=ω0ω1(μ0-μ1)^2 (7)
Spliced aluminium foil picture is obtained making the maximum threshold value T of inter-class variance, as the first threshold using the method for traversal
Value, for example, threshold value T=75, carries out binary conversion treatment to spliced aluminium foil picture with this threshold value, by spliced aluminium foil picture
The aluminium foil part that the gray value of middle pixel is higher than this threshold value is set to 255, is rendered as white；By the gray value of pixel in picture
Non- aluminium foil part not higher than this threshold value is set to 0, is rendered as black, and as a result as shown in figure 3, in Fig. 3,5 indicate noise.
In the present embodiment, as another alternative embodiment, the preset first threshold is to make inter-class variance g=ω0ω1
(μ0-μ1) foreground and background of ^2 when reaching maximum segmentation threshold, g is inter-class variance, ω0Whole picture is accounted for for foreground pixel points
The ratio of spliced aluminium foil picture, μ0For the average gray of the pixel of prospect, ω1Whole picture splicing is accounted for for background pixel points
The ratio of aluminium foil picture afterwards, μ1For the average gray of the pixel of background.
In the present embodiment, because aluminium foil is out-of-flatness, the background of aluminium foil is also possible to be contaminated, and which results in spellings
After binarization, white aluminium foil part has stain generation to aluminium foil picture after connecing, and the non-aluminium foil part of black has white point generation,
In order to avoid the influence (as shown in Figure 3) of these noises, each non-adjacent point can be regarded as a connected region, counted
The quantity of pixel shared by each connected region, the area of as each connected region, judges pixel shared by each connected region
Whether quantity is less than preset third threshold value, if being less than preset third threshold value, will be less than the connection of preset third threshold value
Region weeds out, so that these noises be rejected, aluminium foil is extracted from background, as a result as shown in Figure 4.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Step ground, the non-cubic texture extracted in spliced aluminium foil picture include:
Top cap conversion process is carried out to spliced aluminium foil picture；
Binary conversion treatment is carried out to the aluminium foil picture after top cap conversion process, the brilliant white point of non-cubic texture will be represented never
Initial gross separation comes out in uniform background；
The quantity for counting pixel shared by each white connected region judges the quantity of pixel shared by each white connected region
Whether preset 4th threshold value is greater than, the white connected region that will be greater than preset 4th threshold value is weeded out, will be represented non-cubic
The brilliant white point of texture is accurately separated from background；
Wherein, the white connected region weeded out be as it is large stretch of it is reflective caused by white area.
In the present embodiment, the brilliant white point of non-cubic texture is represented as target using aluminium foil surface in spliced aluminium foil picture
In the case of, the gray value of aluminium foil other parts is very uneven.It should in this way, can first be eliminated using the method converted based on top cap
Influence: assuming that F is input picture, B is the structural element used, and G is output picture, and ο indicates that morphology opens operation, then G=F-
(F ο B), i.e., from subtracting obtained picture after morphology opens operation in original image；Maximum variance between clusters are reused to determine for two-value
The second threshold for changing operation, as inter-class variance g=ω0ω1(μ0-μ1) ^2 is when reaching maximum, threshold value T=53, as second at this time
Threshold value=53 can will represent the brilliant white point of non-cubic texture part from non-uniform background tentatively minute after binary conversion treatment
It separates out and, as a result as shown in Figure 5.
In the present embodiment, since the fold injustice of aluminium foil can cause reflective, generation large stretch of white during picture shooting
Region equally can be taken as white point to separate in binarization, extreme influence detection accuracy, so needing to reject.It is first
The quantity for first counting pixel shared by each white connected region judges whether the quantity of pixel shared by each white connected region is big
In preset 4th threshold value, the white connected region that will be greater than preset 4th threshold value is weeded out, so that Fei Li will be represented by realizing
The brilliant white point of square texture is more precisely separated from background, in addition to non-cubic texture is rendered as brilliant white point in picture
Outside, other parts are rendered as black, as a result as shown in Figure 6.
In the present embodiment, all white points or large stretch of white area can be a white connected regions, even if it
The point that a pixel number is 1, can also a white connected region at last, only the smallest white connected region, so
Afterwards using the 4th threshold value (cannot having can be regarded as that large stretch of white a little instead containing the very big white connected region of pixel number
Light region) it weeds out.
In the present embodiment, the white connected region weeded out be as it is large stretch of it is reflective caused by white area, the described 4th
The value of threshold value can be 200, be to count determining based on a large amount of pictures.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Step, the aluminium foil picture after the conversion process to top cap carries out binary conversion treatment, will represent the brilliant white of non-cubic texture part
Point initial gross separation from non-uniform background, which comes out, includes:
Whether the gray value of the pixel in aluminium foil picture after judging top cap conversion process is greater than preset second threshold；
If more than preset second threshold, then the gray value that will be greater than the pixel of preset second threshold is set as 255, institute
It states pixel and white is presented；
Otherwise, the gray value of the pixel is set as 0, black is presented in the pixel；
Wherein, the preset second threshold is determined by maximum variance between clusters.
In the specific embodiment of the aforementioned electric aluminum foil cubic texture detection method of content based on picture processing, into one
Step ground, the number for counting pixel shared by the aluminium foil extracted respectively and picture shared by the non-cubic texture extracted
The number of element is obtained according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture
Content of the cubic texture in aluminium foil include:
The number for counting pixel shared by the aluminium foil extracted is S1；
The number of pixel shared by white area caused by the sheet that statistics weeds out is reflective is S2；
Statistics represents the number of pixel shared by the brilliant white point of non-cubic texture as S3；
Determine that content of the non-cubic texture in aluminium foil is S3/ (S1-S2)；
According to content of the determining non-cubic texture in spliced aluminium foil picture, cubic texture is obtained in aluminium foil
Content is 1-S3/ (S1-S2).
In the present embodiment, the number for counting pixel shared by the aluminium foil extracted in S102 is S1=1790924 and protects
It deposits；The number of pixel shared by white area caused by the sheet that statistics weeds out is reflective is S2=115605 and saves；Count generation
The number of pixel shared by the brilliant white point of the non-cubic texture of table is S3=77849 and saves；Arithmetic is carried out to result above, really
Fixed content of the non-cubic texture in aluminium foil is S3/ (S1-S2)=4.65%；According to determining non-cubic texture spliced
Content in aluminium foil picture, finally obtaining content of the cubic texture in aluminium foil is 1-S3/ (S1-S2)=1-4.65%=
95.35%.
To sum up, it is true using maximum variance between clusters after the completion of multiple aluminium foil pictures camera being continuously shot are spliced
Determine the methods of first threshold, binary conversion treatment, connected region and extracts aluminium foil in spliced aluminium foil picture；Then pass through top
It is spliced that cap transformation, maximum variance between clusters determine that the methods of second threshold, binary conversion treatment, connected region screening extract
Non-cubic texture in aluminium foil picture；Finally, content of the non-cubic texture in aluminium foil is calculated to each result,
So as to obtain the content of cubic texture.In this way, substituting eye-observation with machine vision, phase is analyzed using picture Processing Technique
Multiple aluminium foil pictures of machine shooting, identify the brilliant white point for representing non-cubic texture on aluminium foil picture, calculate the face of brilliant white point
Product accounts for the ratio of aluminium foil area in the aluminium foil picture, that is, the content of non-cubic texture is illustrated, to obtain containing for cubic texture
Amount, can efficiently and automatically realize the detection of cubic texture content, be effectively reduced human input, reduce detection time, and protect
The accuracy rate for hindering detection, provides more accurately quality information feedback for the production technology of electric aluminum foil product.
In the present embodiment, the characteristics of machine vision is automation, objective, non-contact and high-precision.It is high precisely to abandon dependence
" original " quality testing of artificial eye mode.Because using unified standard, not by artificially generated fatigue, mood, mistake
The influence of factors such as sentence.The picture sensing photograph shot detection system equipment of high quality and the processing of high-caliber machine vision algorithm
Detection means is, it can be achieved that production truly quantifies, it is ensured that the quality of production of aluminium foil products.
The above is a preferred embodiment of the present invention, it is noted that for those skilled in the art
For, without departing from the principles of the present invention, it can also make several improvements and retouch, these improvements and modifications
It should be regarded as protection scope of the present invention.
Claims (8)
1. a kind of electric aluminum foil cubic texture detection method of content based on picture processing characterized by comprising
Multiple aluminium foil pictures that camera is continuously shot are read, and multiple aluminium foil pictures described in shooting are spliced into an aluminium foil figure
Piece；
Extract the aluminium foil in spliced aluminium foil picture；
Extract the non-cubic texture in spliced aluminium foil picture；
Pixel shared by the number for counting pixel shared by the aluminium foil extracted respectively and the non-cubic texture extracted
Number obtains cube according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture
Content of the texture in aluminium foil；
Wherein, the non-cubic texture extracted in spliced aluminium foil picture includes:
Top cap conversion process is carried out to spliced aluminium foil picture；
Binary conversion treatment is carried out to the aluminium foil picture after top cap conversion process, the brilliant white point of non-cubic texture will be represented from uneven
Background in initial gross separation come out；
The quantity for counting pixel shared by each white connected region, judge pixel shared by each white connected region quantity whether
Greater than preset 4th threshold value, the white connected region that will be greater than preset 4th threshold value is weeded out, and will represent non-cubic texture
Brilliant white point accurately separated from background；
Wherein, the white connected region weeded out be as it is large stretch of it is reflective caused by white area；
Wherein, the number for counting pixel shared by the aluminium foil extracted respectively and the non-cubic texture institute extracted
The number for accounting for pixel, according to the number of pixel shared by the number of pixel shared by the aluminium foil of statistics and the non-cubic texture,
Obtaining content of the cubic texture in aluminium foil includes:
The number for counting pixel shared by the aluminium foil extracted is S1；
The number of pixel shared by white area caused by the sheet that statistics weeds out is reflective is S2；
Statistics represents the number of pixel shared by the brilliant white point of non-cubic texture as S3；
Determine that content of the non-cubic texture in aluminium foil is S3/ (S1-S2)；
According to content of the determining non-cubic texture in spliced aluminium foil picture, content of the cubic texture in aluminium foil is obtained
For 1-S3/ (S1-S2).
2. the electric aluminum foil cubic texture detection method of content according to claim 1 based on picture processing, feature exist
In, described multiple aluminium foil pictures for reading camera and being continuously shot, and multiple aluminium foil pictures described in shooting are spliced into an aluminium
Foil picture includes:
Read multiple aluminium foil pictures that camera is continuously shot；
It is registrated based on the method for template matching every two to shooting adjacent aluminium foil pictures；
Every two adjacent aluminium foil pictures after registration are merged using the method for weighting gradual change, what is shot is described more
Open the splicing picture of aluminium foil picture.
3. the electric aluminum foil cubic texture detection method of content according to claim 2 based on picture processing, feature exist
In, it is described based on the method for template matching every two to shooting adjacent aluminium foil pictures carry out registration include:
Every two to shooting adjacent aluminium foil pictures, choose wherein an aluminium foil picture as reference picture, another aluminium foil
Picture is as picture subject to registration；
Determine the area-of-interest in the reference picture as template；
Calculate the absolute value summation of gray value difference between the template and the picture subject to registration:
Wherein, T indicates template, the number of pixel in n expression template, u, and the offset on the direction v x, y, t (x, y) is mould
Coordinate is the gray value of the pixel of (x, y) in plate, and (u+x, v+y) is the coordinate after (x, the y) amount of offsetting, and f (u+x, v+y) is
Coordinate is the gray value of the pixel of (u+x, v+y), sad (u, v) in template when template movement is to picture current location subject to registration
It is a scalar value being calculated based on the gray value, indicates the similarity of template and picture subject to registration, sad (u, v) is most
Hour completes matching.
4. the electric aluminum foil cubic texture detection method of content according to claim 2 based on picture processing, feature exist
In the method using weighting gradual change merges every two adjacent aluminium foil pictures after registration, the institute shot
The splicing picture for stating multiple aluminium foil pictures includes:
Utilize formula I (x, y)=I1(x,y)(1-σ)+I2(x, y) merges every two adjacent pictures after registration, obtains
The splicing picture of multiple aluminium foil pictures of shooting；
Wherein, I1(x, y) and I2(x, y) respectively indicates the gray value of each pixel of to be spliced two adjacent picture after registration,
I (x, y) indicates I1(x, y) and I2The gray value of the pixel of overlapping region in (x, y), weighting coefficient σ ∈ (0,1), when σ is become by 0
When changing to 1, picture is from I1(x, y) is transitioned into I2(x, y).
5. the electric aluminum foil cubic texture detection method of content according to claim 1 based on picture processing, feature exist
In the aluminium foil extracted in spliced aluminium foil picture includes:
Binary conversion treatment is carried out to spliced aluminium foil picture, so that the aluminium foil part in the spliced aluminium foil picture is presented
Black is presented in white, non-aluminium foil part；
Each regard the non-adjacent point of each of aluminium foil picture after binary conversion treatment as a connected region, statistics is each
The quantity of pixel shared by connected region, judges whether the quantity of pixel shared by each connected region is less than preset third threshold value,
If being less than preset third threshold value, the connected region for being less than preset third threshold value is weeded out.
6. the electric aluminum foil cubic texture detection method of content according to claim 5 based on picture processing, feature exist
In, it is described that binary conversion treatment is carried out to spliced aluminium foil picture, so that the aluminium foil part in the spliced aluminium foil picture
White is presented, non-aluminium foil part is presented black and includes:
Judge whether the gray value of the pixel in spliced aluminium foil picture is greater than preset first threshold；
If more than preset first threshold, then the gray value that will be greater than the pixel of preset first threshold is set as 255, the picture
White is presented in vegetarian refreshments；
Otherwise, the gray value of the pixel is set as 0, black is presented in the pixel.
7. the electric aluminum foil cubic texture detection method of content according to claim 6 based on picture processing, feature exist
In the preset first threshold is to make inter-class variance g=ω0ω1(μ0-μ1) foreground and background of ^2 when reaching maximum point
Threshold value is cut, g is inter-class variance, ω0The ratio of the spliced aluminium foil picture of whole picture, μ are accounted for for foreground pixel points0For the picture of prospect
The average gray of vegetarian refreshments, ω1The ratio of the spliced aluminium foil picture of whole picture, μ are accounted for for background pixel points1For the pixel of background
Average gray.
8. the electric aluminum foil cubic texture detection method of content according to claim 1 based on picture processing, feature exist
In the aluminium foil picture after the conversion process to top cap carries out binary conversion treatment, will represent the brilliant white point of non-cubic texture part
Initial gross separation, which comes out, from non-uniform background includes:
Whether the gray value of the pixel in aluminium foil picture after judging top cap conversion process is greater than preset second threshold；
If more than preset second threshold, then the gray value that will be greater than the pixel of preset second threshold is set as 255, the picture
White is presented in vegetarian refreshments；
Otherwise, the gray value of the pixel is set as 0, black is presented in the pixel；
Wherein, the preset second threshold is determined by maximum variance between clusters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610843102.9A CN106504251B (en) | 2016-09-22 | 2016-09-22 | A detection method for cubic texture content of electronic aluminum foil based on image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610843102.9A CN106504251B (en) | 2016-09-22 | 2016-09-22 | A detection method for cubic texture content of electronic aluminum foil based on image processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106504251A CN106504251A (en) | 2017-03-15 |
| CN106504251B true CN106504251B (en) | 2019-02-15 |
Family
ID=58290809
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610843102.9A Active CN106504251B (en) | 2016-09-22 | 2016-09-22 | A detection method for cubic texture content of electronic aluminum foil based on image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106504251B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107094683B (en) * | 2017-04-13 | 2020-11-27 | 同济大学 | An automatic feeding and water quality monitoring and control system for aquaculture |
| CN108362699B (en) * | 2018-02-13 | 2020-12-08 | 仲恺农业工程学院 | A kind of determination method of potato peeling rate |
| CN118091077B (en) * | 2024-02-29 | 2024-11-26 | 广西广投正润新材料科技有限公司 | A method for determining the cubic texture content of electronic aluminum foil |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101435769A (en) * | 2008-12-17 | 2009-05-20 | 安阳钢铁集团有限责任公司 | Method for measuring certain phase content in gold phase and measuring high-carbon steel sorbite content |
| CN102937583A (en) * | 2012-10-24 | 2013-02-20 | 浙江工业大学 | Pearl smooth-finish online automatic grading device based on monocular multi-view machine vision |
| CN103286081A (en) * | 2013-05-07 | 2013-09-11 | 浙江工业大学 | Monocular multi-perspective machine vision-based online automatic sorting device for steel ball surface defect |
| CN103456021A (en) * | 2013-09-24 | 2013-12-18 | 苏州大学 | Piece goods blemish detecting method based on morphological analysis |
| CN103760170A (en) * | 2013-12-26 | 2014-04-30 | 中国电子科技集团公司第四十一研究所 | Tobacco bale lining defect detection method based on machine vision technology |
-
2016
- 2016-09-22 CN CN201610843102.9A patent/CN106504251B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101435769A (en) * | 2008-12-17 | 2009-05-20 | 安阳钢铁集团有限责任公司 | Method for measuring certain phase content in gold phase and measuring high-carbon steel sorbite content |
| CN102937583A (en) * | 2012-10-24 | 2013-02-20 | 浙江工业大学 | Pearl smooth-finish online automatic grading device based on monocular multi-view machine vision |
| CN103286081A (en) * | 2013-05-07 | 2013-09-11 | 浙江工业大学 | Monocular multi-perspective machine vision-based online automatic sorting device for steel ball surface defect |
| CN103456021A (en) * | 2013-09-24 | 2013-12-18 | 苏州大学 | Piece goods blemish detecting method based on morphological analysis |
| CN103760170A (en) * | 2013-12-26 | 2014-04-30 | 中国电子科技集团公司第四十一研究所 | Tobacco bale lining defect detection method based on machine vision technology |
Non-Patent Citations (5)
| Title |
|---|
| Color and texture for corn seed classification by machine vision;K Kiratiratanapruk 等;《2011 International Symposium on Intelligent Signal Processing and Communications Systems(ISPACS)》;20111209;1-5 |
| 利用图像分析技术检测纸浆中胶黏物含量;闫瑛 等;《纸和造纸》;20130515;第32卷(第5期);19-22 |
| 基于机器视觉的印刷品缺陷在线检测系统关键技术研究;阚希;《中国优秀硕士学位论文全文数据库》;20140215(第02期);I140-406 |
| 基于机器视觉的晶粒度面积法自动评级;林金萱 等;《金属热处理》;20070325;第32卷(第3期);82-85 |
| 小孔腐蚀在铝电解电容器用铝箔质量评价中的应用;班朝磊 等;《表面技术》;20160320;第45卷(第3期);188-192 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106504251A (en) | 2017-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106093066B (en) | A kind of magnetic tile surface defect detection method based on improved machine vision attention mechanism | |
| CN111047655B (en) | High-definition camera cloth defect detection method based on convolutional neural network | |
| Peng et al. | Deformation feature extraction and double attention feature pyramid network for bearing surface defects detection | |
| CN105957059B (en) | Electronic component missing detection method and system | |
| CN117078608B (en) | A method for detecting highly reflective leather surface defects based on double mask guidance | |
| CN106530271B (en) | A kind of infrared image conspicuousness detection method | |
| CN115018844A (en) | Plastic film quality evaluation method based on artificial intelligence | |
| CN104197836A (en) | Vehicle lock assembly size detection method based on machine vision | |
| CN109285140A (en) | A kind of printed circuit board image registration evaluation method | |
| CN109900719A (en) | A kind of visible detection method of blade surface knife mark | |
| Liang et al. | Research on concrete cracks recognition based on dual convolutional neural network | |
| CN110443278A (en) | A method, device and equipment for detecting abnormal thickness of grid lines of solar cells | |
| CN106504251B (en) | A detection method for cubic texture content of electronic aluminum foil based on image processing | |
| You | PCB defect detection based on generative adversarial network | |
| CN119919410B (en) | Method for measuring size and appearance of O-rings | |
| CN119273992A (en) | Mobile phone screen glass defect detection method based on improved YOLOv8 | |
| CN109241948A (en) | A kind of NC cutting tool visual identity method and device | |
| KR20230140621A (en) | Concrete crack detection method and system using artificial intelligence | |
| CN106248634A (en) | Fruit exocuticle glossiness measurement apparatus and method | |
| CN109448012A (en) | A kind of method for detecting image edge and device | |
| CN108765365A (en) | A kind of rotor winding image qualification detection method | |
| CN111815600A (en) | Visual sense-based annular magnetic steel appearance defect detection method | |
| CN110660048A (en) | Leather surface defect detection algorithm based on shape characteristics | |
| Turakhia et al. | Automatic crack detection in heritage site images for image inpainting | |
| Junxiong et al. | Feature extraction of jujube fruit wrinkle based on the watershed segmentation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |