CN110084768A - The defect inspection method of LCD light guide plate based on background filtering - Google Patents
The defect inspection method of LCD light guide plate based on background filtering Download PDFInfo
- Publication number
- CN110084768A CN110084768A CN201910389155.1A CN201910389155A CN110084768A CN 110084768 A CN110084768 A CN 110084768A CN 201910389155 A CN201910389155 A CN 201910389155A CN 110084768 A CN110084768 A CN 110084768A
- Authority
- CN
- China
- Prior art keywords
- image
- light guide
- guide plate
- formula
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001914 filtration Methods 0.000 title claims abstract description 45
- 230000007547 defect Effects 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000007689 inspection Methods 0.000 title claims abstract description 13
- 238000001514 detection method Methods 0.000 claims abstract description 18
- 230000000694 effects Effects 0.000 claims abstract description 9
- 238000001228 spectrum Methods 0.000 claims description 10
- 238000012216 screening Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000010276 construction Methods 0.000 claims 1
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000012360 testing method Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 5
- 239000004973 liquid crystal related substance Substances 0.000 description 5
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
Classifications
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/149—Segmentation; Edge detection involving deformable models, e.g. active contour models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/168—Segmentation; Edge detection involving transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20056—Discrete and fast Fourier transform, [DFT, FFT]
Abstract
The present invention provides a kind of defect inspection method of LCD light guide plate based on background filtering, comprising: S110, acquisition LCD light guide plate image；S120, Steerable filter is iterated to the LCD light guide plate image；S130, frequency domain filtering is carried out using Fast Fourier Transform (FFT) method to image after the LCD light guide plate Steerable filter；S140, defect area is divided using maximum variance between clusters to image after the LCD light guide plate frequency domain filtering, obtains testing result.The defect inspection method of LCD light guide plate provided by the invention based on background filtering has preferable background texture filter effect and higher defect detection rate.
Description
Technical field
The present invention relates to detection algorithm technical field more particularly to a kind of defects of the LCD light guide plate based on background filtering
Detection method.
Background technique
Liquid crystal display-light guide plate (Liquid Crystal Display-Light Guide Plate LCD-LGP) is
The important component in backlight of LCD mould group is produced, if the round lattice structure in light guide plate surface is defective, by shadow
Brightness and uniformity that liquid crystal display is shown are rung, so the surface defect of detection light guide plate is the necessary links for producing LCD.In recent years
Come, the rapid development of China's display manufacturing industry and production line is increasingly automated, intelligentized trend is promoted based on machine
The development of surface defect automatic optics inspection (Auto Optical Inspection AOI) technology of vision, not only increases
The yield and yields of LCD, and reduce artificial vision and detect bring cost.
But existing algorithm can not well be filtered complicated background texture, and LCD light guide plate surface is led
The features such as light particle has distribution density larger, and distributed architecture lacks periodic regularity and spectrogram diverging decaying, to being based on
The detection method of frequency-domain analysis brings certain difficulty, it is therefore desirable to it is existing based on the detection method of frequency-domain analysis into
Row improves.
Summary of the invention
It is an object of the present invention to overcome the shortcomings of the prior art and provide a kind of LCD based on background filtering to lead
The defect inspection method of tabula rasa has higher recall rate compared to traditional detection method based on frequency-domain analysis.The present invention uses
Technical solution be:
A kind of detection method of surface flaw of LCD light guide plate, comprising:
S110, acquisition LCD light guide plate image；
S120, Steerable filter is iterated to the LCD light guide plate image；
S130, frequency domain filtering is carried out using Fast Fourier Transform (FFT) method to image after the LCD light guide plate Steerable filter；
S140, defect area is divided using dynamic thresholding method to image after the LCD light guide plate frequency domain filtering, is detected
As a result.
The present invention has the advantages that the present invention passes through building navigational figure, using the guiding filtering processing mode of iteration, and
Iteration stopping criterion is introduced according to structural similarity, preliminary filtering is carried out to LCD light guide plate image, to LCD light guide plate image
Background information maintains the profile of defect area to a certain extent while filtering；Filtered image is passed through into two-dimensional discrete Fu
In leaf transformation, the high-frequency noise after guiding filtering in image is further obviated, finally by dynamic threshold by target defect point
It cuts out.Experiments verify that the present invention to the complex texture of LCD light guide plate image and the background information of aperiodicity rule have compared with
Good filter effect, has higher recall rate compared to traditional detection method based on frequency-domain analysis.
Detailed description of the invention
Fig. 1 is the defect inspection method flow chart of LCD light guide plate of the invention.
Fig. 2 is the effect image of step S120 of the invention to the LCD light guide plate image after iteration guiding filtering.
Fig. 3 is the effect image in step S130 of the invention to the LCD light guide plate image after frequency domain filtering.
Fig. 4 is to obtain the reconstruct image of removal background by two-dimensional discrete Fourier inverse transformation in step S130 of the invention
Picture.
Fig. 5 is that step S140 of the invention divides image after the LCD light guide plate frequency domain filtering using dynamic thresholding method
The defect area image obtained after defect area.
Specific embodiment
Below in conjunction with attached drawing, detailed description of the preferred embodiments.It should be understood that this place is retouched
The specific embodiment stated is merely to illustrate and explain the present invention, and is not intended to restrict the invention.
It should be noted that LCD-LGP:Liquid Crystal Display-Light Guide Plate, liquid crystal display
Light guide plate；SSIM:structural similarity index measure, structural similarity index.
As one aspect of the present invention, a kind of detection method of surface flaw of LCD light guide plate is provided, wherein such as Fig. 1 institute
Show, the detection method of surface flaw of the LCD light guide plate includes:
S110, acquisition LCD light guide plate image；
S120, Steerable filter is iterated to the LCD light guide plate image；
S130, frequency domain filtering is carried out using Fast Fourier Transform (FFT) method to image after the LCD light guide plate Steerable filter；
S140, defect area is divided using dynamic thresholding method to image after the LCD light guide plate frequency domain filtering, is detected
As a result.
The defects detection algorithm of LCD light guide plate provided by the invention based on background filtering is adopted by constructing navigational figure
Iteration stopping criterion is introduced with the guiding filtering processing mode of iteration, and according to structural similarity, LCD light guide plate image is carried out
Preliminary filtering maintains the profile of defect area to a certain extent while to LCD light guide plate image background information filter；
Filtered image is passed through into two dimensional discrete Fourier transform, further obviates the high-frequency noise after guiding filtering in image, most
Target defect is split by dynamic threshold afterwards.Experiments verify that the present invention it is complicated to LCD light guide plate image and without week
The background information of phase property rule has preferable filter effect, has higher inspection compared to traditional detection method based on frequency-domain analysis
Extracting rate.
Specifically, S120, Steerable filter is iterated to the LCD light guide plate image, comprising:
Piecemeal processing is carried out to the LCD light guide plate image of acquisition；Such as it is divided into hundreds of subgraphs；
According to the corresponding initial guiding image of the length of each subgraph and wide building；
The gray scale initial value of each pixel is set as 0 in all initial guiding images；
In formula, I is guiding image, IiFor the pixel value for being oriented to image, i and k indicate the position of pixel, ωkIndicate with
The two-dimentional window that radius centered on pixel k is r, p indicate that input picture, q indicate that output image, E are so that output image
Value q at pixel iiWith value p of the input picture at pixel iiBetween the smallest cost function of gap, akAnd bkIt is logical
The coefficient that linear regression is found out is crossed, in akExpression formula in, | ω | for two-dimentional window ωkIn pixel sum,It is that p exists
Two-dimentional window ωkIn average value, μkIt is guiding image I in two-dimentional window ωkIn average value,It is guiding image I two
Tie up window ωkIn variance, ε is to prevent akExcessive regularization parameter, due to two-dimentional window ωkInside there is quantity is | ω | it is a
Pixel can cover location of pixels i, so utilizing akAnd bkAverage valueWithAchieve the effect that denoising；
Using the image after Steerable filter for the first time as the guiding image of next Steerable filter, i.e., during iteration
Continuous to update guiding image, the formula of iteration Steerable filter is as follows:
In formula, G indicates Steerable filter,It indicates in the t times iteration with qt-1It is defeated for guiding image, p
Enter the Steerable filter process of image, output image is qt, the radius of the two-dimentional window of r expression, s is smoothing factor, and regularization is taken to join
Several inverses is simultaneously rounded downwards；
Guiding filter is constructed using structural similarity index SSIM (structural similarity index measure)
The formula of the iteration stopping criterion of wave, structural similarity and iteration stopping criterion is as follows:
|SSIM(n+1)-SSIM(n)| < ζ, n=1,2,3 ..., m
In formula, μtAnd μt+1Output image q after respectively representing t moment and t+1 moment Steerable filtertAnd qt+1Mean value, σt
And σt+1Output image q after respectively representing t moment and t+1 moment Steerable filtertAnd qt+1Standard deviation, σt(t+1)Indicate t moment
With the output image q after t+1 moment Steerable filtertAnd qt+1Covariance, C1And C2For constant, enabling ζ is that iteration Steerable filter stops
The only factor, ζ are the adjustable positive number much smaller than 1, SSIM(n)Indicate n-th Steerable filter output image and n-1 Steerable filter
Export the structural similarity numerical value of image, SSIM(n+1)Indicate that (n+1)th Steerable filter output image and n-th Steerable filter are defeated
The structural similarity numerical value of image out, works as SSIM(n+1)With SSIM(n)When absolute value of the difference is much smaller than ζ, it is believed that after iteration
Steerable filter effect almost no longer changes.
Specifically, S130, to image after the LCD light guide plate Steerable filter using Fast Fourier Transform (FFT) method carry out frequency domain
Filtering, comprising:
Image is denoted as f (x, y) after Steerable filter, carries out two dimensional discrete Fourier transform to image and obtains corresponding frequency domain figure
As F (u, v), and to F (u, v) modulus, formula is as follows:
In formula, M and N indicate that the size of image, the frequency domain of image are indicated by F (u, v), and u and v are frequency variable, u=0,
1,2,...,M-1；V=0,1,2 ..., N-1, x and y are space variable, and R (u, v) and I (u, v) respectively indicate the reality of F (u, v)
Portion and imaginary part；
Power spectrum is calculated, obtained power spectrum is subjected to Gauss high-pass filtering, formula is as follows:
P (u, v)=| F (u, v) |2
PG(u, v)=G (u, v) P (u, v)
In formula, P (u, v) is power spectrum, PG(u, v) is the power spectrum after Gauss high pass wave, G (u, v)
For Gaussian filter；
Using global threshold to PGHighlight regions are split in (u, v), and formula is as follows:
TMinGrayValue< g < TMaxGrayValue
In formula, TMinGrayValueIndicate pixel gray value lower threshold used by dividing, TMaxGrayValueIndicate segmentation
Used pixel gray value upper limit threshold；G is the gray value of qualified pixel；
It is screened by size in region after global threshold is divided, wherein each pixel occupied area numerical value
Size is denoted as 1, and the condition of region screening is as follows:
TMinAera< r < TMaxAera
In formula, TMinAeraIndicate area lower threshold used by screening, TMaxAeraIndicate the area upper limit used by screening
Threshold value；
Respective mass center is calculated to all subregions filtered out, formula is as follows:
In formula, umAnd vmFor the coordinate of mass center, NrFor the number of pixel in connected domain (each sub-regions filtered out)；
Using the mass center of each sub-regions as the center of circle, justifies by radius work of R, the gray value of the pixel in border circular areas is set
It is set to zero；
Frequency domain image after removal high-frequency information is obtained into the reconstruct of removal background by two-dimensional discrete Fourier inverse transformation
Image；Formula is as follows:
Specifically, S140, to image after the LCD light guide plate frequency domain filtering using dynamic thresholding method divide defect area,
Obtain testing result, comprising:
Reconstructed image after enabling frequency domain filtering is g (s, t), carries out mean filter to reconstructed image g (s, t), formula is as follows:
In formula, CxyExpression mean filter central point is at (x, y), and mean filter window size is m and n；
Enabling β is shift term, by image f after mean filter2(x, y) is compared pixel-by-pixel with the reconstructed image after frequency domain filtering,
Fluctuation of the gray value of pixel within the scope of shift term β is considered as acceptable gray value, the fluctuation view outside shift term β range
For the gray value of target area pixel, that is, complete the detection of target defect.
Specific implementation below with reference to Fig. 2 to Fig. 5 to the detection method of surface flaw of LCD light guide plate provided by the invention
Journey is described in detail；
1, the image of product to be detected is acquired using line scan camera, as shown in Figure 2；
2, piecemeal is carried out to image, in present embodiment: block image size is 512pixels*512pixels；
3, the navigational figure that building size is 512pixels*512pixels, enabling the pixel value of each pixel is 0；
4, guiding filtering is iterated to the subgraph of piecemeal；Formula is as follows:
5, foundation iteration stopping criterion, the number of iterations of adaptive selection iteration guiding filtering process, formula are as follows:
|SSIM(n+1)-SSIM(n)| < ζ, n=1,2,3 ..., m
6, the image after iteration guiding filtering is subjected to two dimensional discrete Fourier transform, formula is as follows, as a result such as Fig. 3 institute
Show:
7, power spectrum is calculated, formula is as follows:
P (u, v)=| F (u, v) |2
8, gaussian filtering is carried out to power spectrum, formula is as follows:
PG(u, v)=G (u, v) P (u, v)
9, using global threshold to PGHighlight regions are split in (u, v), and formula is as follows:
TMinGrayValue< g < TMaxGrayValue
10, the region after dividing global threshold is screened by size, and the condition of region screening is as follows:
TMinAera< r < TMaxAera
11, respective mass center is calculated to all subregions filtered out, formula is as follows:
12, using the mass center of each sub-regions as the center of circle, justify by radius work of R, by the gray scale of the pixel in border circular areas
Value is set as zero；
13, the frequency domain image after removal high-frequency information is obtained into the reconstructed image of removal background by inverse Fourier transform,
Formula is as follows, as a result as shown in Figure 4:
14, the reconstructed image after enabling frequency domain filtering is g (s, t), carries out mean filter to image g (s, t), formula is as follows:
15, enabling β is shift term, by image f after mean filter2(x, y) and the reconstructed image after frequency domain filtering are right pixel-by-pixel
Than fluctuation of the gray value of pixel within the scope of shift term β is considered as acceptable gray value, the wave outside shift term β range
The dynamic gray value for being considered as target area pixel, that is, complete the detection of target defect, shift term of the present invention may be configured as outside can
Adjustment parameter, testing result are as shown in Figure 5.
It should be noted last that the above specific embodiment is only used to illustrate the technical scheme of the present invention and not to limit it,
Although being described the invention in detail referring to example, those skilled in the art should understand that, it can be to the present invention
Technical solution be modified or replaced equivalently, without departing from the spirit and scope of the technical solution of the present invention, should all cover
In the scope of the claims of the present invention.
Claims (4)
1. a kind of defect inspection method of the LCD light guide plate based on background filtering characterized by comprising
S110, acquisition LCD light guide plate image；
S120, Steerable filter is iterated to the LCD light guide plate image；
S130, frequency domain filtering is carried out using Fast Fourier Transform (FFT) method to image after the LCD light guide plate Steerable filter；
S140, defect area is divided using dynamic thresholding method to image after the LCD light guide plate frequency domain filtering, obtains detection knot
Fruit.
2. the defect inspection method of the LCD light guide plate as described in claim 1 based on background filtering, which is characterized in that
The S120, Steerable filter is iterated to the LCD light guide plate image, specifically included:
Piecemeal processing is carried out to the LCD light guide plate image of acquisition；
According to the corresponding initial guiding image of the length of each subgraph and wide building；
The gray scale initial value of each pixel is set as 0 in all initial guiding images；
In formula, I is guiding image, IiFor the pixel value for being oriented to image, i and k indicate the position of pixel, ωkIt indicates with pixel
The two-dimentional window that radius centered on point k is r, p indicate that input picture, q indicate that output image, E are so that output image is in picture
Value q at vegetarian refreshments iiWith value p of the input picture at pixel iiBetween the smallest cost function of gap, akAnd bkIt is to pass through line
Property return the coefficient that finds out, in akExpression formula in, | ω | for two-dimentional window ωkIn pixel sum,It is p in two dimension
Window ωkIn average value, μkIt is guiding image I in two-dimentional window ωkIn average value,It is guiding image I in two-dimentional window
Mouth ωkIn variance, ε is to prevent akExcessive regularization parameter, due to two-dimentional window ωkInside there is quantity is | ω | a picture
Vegetarian refreshments can cover location of pixels i, so utilizing akAnd bkAverage valueWithAchieve the effect that denoising；
Using the image after Steerable filter for the first time as the guiding image of next Steerable filter, i.e., during iteration constantly
Update be oriented to image, the formula of iteration Steerable filter is as follows:
In formula, G indicates Steerable filter,It indicates in the t times iteration with qt-1It is input figure for guiding image, p
The Steerable filter process of picture, output image are qt, the radius of the two-dimentional window of r expression, s is smoothing factor, takes regularization parameter
Inverse is simultaneously rounded downwards；
Using the iteration stopping criterion of structural similarity index SSIM construction Steerable filter, structural similarity and iteration stopping criterion
Formula it is as follows:
|SSIM(n+1)-SSIM(n)| < ζ, n=1,2,3 ..., m
In formula, μtAnd μt+1Output image q after respectively representing t moment and t+1 moment Steerable filtertAnd qt+1Mean value, σtWith
σt+1Output image q after respectively representing t moment and t+1 moment Steerable filtertAnd qt+1Standard deviation, σt(t+1)Indicate t moment and
Output image q after t+1 moment Steerable filtertAnd qt+1Covariance, C1And C2For constant, enable ζ for the stopping of iteration Steerable filter
The factor, ζ are the adjustable positive number much smaller than 1, SSIM(n)Indicate that n-th Steerable filter output image and n-1 Steerable filter are defeated
The structural similarity numerical value of image out, SSIM(n+1)Indicate (n+1)th Steerable filter output image and the output of n-th Steerable filter
The structural similarity numerical value of image, works as SSIM(n+1)With SSIM(n)When absolute value of the difference is much smaller than ζ, it is believed that leading after iteration
Almost no longer change to filter effect.
3. the defect inspection method of the LCD light guide plate as described in claim 1 based on background filtering, which is characterized in that
The S130, frequency domain filtering, tool are carried out using Fast Fourier Transform (FFT) method to image after the LCD light guide plate Steerable filter
Body includes:
Image is denoted as f (x, y) after Steerable filter, carries out two dimensional discrete Fourier transform to image and obtains corresponding frequency domain image F
(u, v), and to F (u, v) modulus, formula is as follows:
In formula, M and N indicate that the size of image, the frequency domain of image are indicated by F (u, v), and u and v are frequency variable, u=0,1,
2,...,M-1；V=0,1,2 ..., N-1, x and y are space variable, and R (u, v) and I (u, v) respectively indicate the real part of F (u, v)
And imaginary part；
Power spectrum is calculated, obtained power spectrum is subjected to Gauss high-pass filtering, formula is as follows:
P (u, v)=| F (u, v) |2
PG(u, v)=G (u, v) P (u, v)
In formula, P (u, v) is power spectrum, PG(u, v) is the power spectrum after Gauss high pass wave, and G (u, v) is Gauss
Filter；
Using global threshold to PGHighlight regions are split in (u, v), and formula is as follows:
TMinGrayValue< g < TMaxGrayValue
In formula, TMinGrayValueIndicate pixel gray value lower threshold used by dividing, TMaxGrayValueIndicate that segmentation is adopted
Pixel gray value upper limit threshold；G is the gray value of qualified pixel；
It is screened by size in region after global threshold is divided, wherein each pixel occupied area numerical values recited
It is denoted as 1, the condition of region screening is as follows:
TMinAera< r < TMaxAera
In formula, TMinAeraIndicate area lower threshold used by screening, TMaxAeraIndicate area upper limit threshold used by screening
Value；
Respective mass center is calculated to all subregions filtered out, formula is as follows:
In formula, umAnd vmFor the coordinate of mass center, NrIt is the number of pixel in each sub-regions filtered out for connected domain；
Using the mass center of each sub-regions as the center of circle, justifies by radius work of R, set the gray value of the pixel in border circular areas to
Zero；
Frequency domain image after removal high-frequency information is obtained into the reconstructed image of removal background by two-dimensional discrete Fourier inverse transformation；
Formula is as follows:
4. the defect inspection method of the LCD light guide plate as described in claim 1 based on background filtering, which is characterized in that
The S140, defect area is divided using dynamic thresholding method to image after the LCD light guide plate frequency domain filtering, is detected
As a result, specifically including:
Reconstructed image after enabling frequency domain filtering is g (s, t), carries out mean filter to reconstructed image g (s, t), formula is as follows:
In formula, CxyExpression mean filter central point is at (x, y), and mean filter window size is m and n；
Enabling β is shift term, by image f after mean filter2(x, y) is compared pixel-by-pixel with the reconstructed image after frequency domain filtering, pixel
Fluctuation of the gray value within the scope of shift term β be considered as acceptable gray value, the fluctuation outside shift term β range is considered as target
The gray value of area pixel point completes the detection of target defect.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910389155.1A CN110084768B (en) | 2019-05-10 | 2019-05-10 | Defect detection method of LCD light guide plate based on background filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910389155.1A CN110084768B (en) | 2019-05-10 | 2019-05-10 | Defect detection method of LCD light guide plate based on background filtering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110084768A true CN110084768A (en) | 2019-08-02 |
| CN110084768B CN110084768B (en) | 2023-06-20 |
Family
ID=67419603
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910389155.1A Active CN110084768B (en) | 2019-05-10 | 2019-05-10 | Defect detection method of LCD light guide plate based on background filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110084768B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110530883A (en) * | 2019-09-30 | 2019-12-03 | 凌云光技术集团有限责任公司 | A kind of defect inspection method |
| CN113658154A (en) * | 2021-08-24 | 2021-11-16 | 凌云光技术股份有限公司 | Image detection method and device based on frequency domain periodic texture removal |
| CN114998354A (en) * | 2022-08-08 | 2022-09-02 | 江苏多孚多智能装备科技有限公司 | Steel belt roll mark detection method based on graph filtering |
| WO2024015364A1 (en) * | 2022-07-12 | 2024-01-18 | Onto Innovation Inc. | Periodic-pattern background removal |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416740A (en) * | 2018-01-22 | 2018-08-17 | 大连大学 | A kind of iteration self-adapting median filtering algorithm for eliminating salt-pepper noise |
| CN108876768A (en) * | 2018-05-30 | 2018-11-23 | 杭州舜浩科技有限公司 | Light guide plate shadow defect inspection method |
| CN109102497A (en) * | 2018-07-13 | 2018-12-28 | 杭州舜浩科技有限公司 | High-resolution light guide plate image defect detection method |
-
2019
- 2019-05-10 CN CN201910389155.1A patent/CN110084768B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416740A (en) * | 2018-01-22 | 2018-08-17 | 大连大学 | A kind of iteration self-adapting median filtering algorithm for eliminating salt-pepper noise |
| CN108876768A (en) * | 2018-05-30 | 2018-11-23 | 杭州舜浩科技有限公司 | Light guide plate shadow defect inspection method |
| CN109102497A (en) * | 2018-07-13 | 2018-12-28 | 杭州舜浩科技有限公司 | High-resolution light guide plate image defect detection method |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110530883A (en) * | 2019-09-30 | 2019-12-03 | 凌云光技术集团有限责任公司 | A kind of defect inspection method |
| CN110530883B (en) * | 2019-09-30 | 2022-08-02 | 凌云光技术股份有限公司 | Defect detection method |
| CN113658154A (en) * | 2021-08-24 | 2021-11-16 | 凌云光技术股份有限公司 | Image detection method and device based on frequency domain periodic texture removal |
| WO2024015364A1 (en) * | 2022-07-12 | 2024-01-18 | Onto Innovation Inc. | Periodic-pattern background removal |
| CN114998354A (en) * | 2022-08-08 | 2022-09-02 | 江苏多孚多智能装备科技有限公司 | Steel belt roll mark detection method based on graph filtering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110084768B (en) | 2023-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110084768A (en) | The defect inspection method of LCD light guide plate based on background filtering | |
| CN105894036B (en) | A kind of characteristics of image template matching method applied to mobile phone screen defects detection | |
| CN108921176B (en) | Pointer instrument positioning and identifying method based on machine vision | |
| CN106846344B (en) | A kind of image segmentation optimal identification method based on the complete degree in edge | |
| CN110286124B (en) | Machine vision-based refractory brick measuring system | |
| CN109580630B (en) | Visual inspection method for defects of mechanical parts | |
| CN108346141B (en) | Method for extracting defects of single-side light-entering type light guide plate | |
| CN108416766B (en) | Double-side light-entering type light guide plate defect visual detection method | |
| CN108665458A (en) | Transparent body surface defect is extracted and recognition methods | |
| CN110207592A (en) | Building cracks measurement method, device, computer equipment and storage medium | |
| CN114066820A (en) | Fabric defect detection method based on Swin-transducer and NAS-FPN | |
| CN115100206B (en) | Printing defect identification method for textile with periodic pattern | |
| CN116030055B (en) | Quality detection method for elevator-related accessories | |
| CN110530883A (en) | A kind of defect inspection method | |
| CN112014413A (en) | Mobile phone glass cover plate window area defect detection method based on machine vision | |
| CN107895371B (en) | Textile flaw detection method based on peak coverage value and Gabor characteristics | |
| CN116664565A (en) | Hidden crack detection method and system for photovoltaic solar cell | |
| Aldemir et al. | An improved gray line profile method to inspect the warp–weft density of fabrics | |
| CN109544513A (en) | A kind of steel pipe end surface defect extraction knowledge method for distinguishing | |
| CN116485787B (en) | Method for detecting appearance defects of data line molding outer die | |
| CN114140373A (en) | Switch defect detection method based on LabVIEW deep learning | |
| George et al. | Automatic defect detection inspectacles and glass bottles based on Fuzzy C Means Clustering | |
| CN108171741B (en) | Image texture decomposition method based on self-adaptive multi-directional empirical mode decomposition | |
| CN110264434A (en) | A kind of removing rain based on single image method based on low-rank matrix completion | |
| CN113724223B (en) | YOLOv3 data set production method and system based on optical microscope |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |