CN113608378B - Full-automatic defect detection method and system based on LCD (liquid crystal display) process - Google Patents
Full-automatic defect detection method and system based on LCD (liquid crystal display) process Download PDFInfo
- Publication number
- CN113608378B CN113608378B CN202111172203.5A CN202111172203A CN113608378B CN 113608378 B CN113608378 B CN 113608378B CN 202111172203 A CN202111172203 A CN 202111172203A CN 113608378 B CN113608378 B CN 113608378B
- Authority
- CN
- China
- Prior art keywords
- image
- gray
- liquid crystal
- integrated
- detection
- 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
Images
Classifications
-
- G—PHYSICS
- G02—OPTICS
- G02F—OPTICAL DEVICES OR ARRANGEMENTS FOR THE CONTROL OF LIGHT BY MODIFICATION OF THE OPTICAL PROPERTIES OF THE MEDIA OF THE ELEMENTS INVOLVED THEREIN; NON-LINEAR OPTICS; FREQUENCY-CHANGING OF LIGHT; OPTICAL LOGIC ELEMENTS; OPTICAL ANALOGUE/DIGITAL CONVERTERS
- G02F1/00—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics
- G02F1/01—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics for the control of the intensity, phase, polarisation or colour
- G02F1/13—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics for the control of the intensity, phase, polarisation or colour based on liquid crystals, e.g. single liquid crystal display cells
- G02F1/1306—Details
- G02F1/1309—Repairing; Testing
-
- 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
- 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
-
- 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
- G01N2021/9513—Liquid crystal panels
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Biochemistry (AREA)
- Nonlinear Science (AREA)
- Crystallography & Structural Chemistry (AREA)
- Optics & Photonics (AREA)
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Liquid Crystal (AREA)
Abstract
The invention provides a full-automatic defect detection method and system based on LCD manufacturing procedure, a conveying mechanism of an LCD production line conveys a liquid crystal panel and is positioned by a positioning mechanism, and a light source outputs detection light rays to the liquid crystal panel; adjusting the distance between a lens of the CCD camera and the liquid crystal panel to carry out focusing processing; sequentially acquiring a plurality of original images of the liquid crystal panel by the CCD camera along a preset path; carrying out deformity correction processing on each original image by an image correction module; image splicing is carried out on all the malformation corrected images by an image splicing module to obtain a global image; the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm; and the image detection device adopts a preset detection algorithm to perform image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines, and outputs a detection result. The invention can effectively improve the detection efficiency and accuracy of the LCD manufacturing process.
Description
Technical Field
The invention relates to the technical field of defect detection, in particular to a full-automatic defect detection method and system based on an LCD (liquid crystal display) process.
Background
With the continuous development of the Liquid Crystal Display (LCD) industry, the quality requirement of the LCD panel is higher and higher, and therefore, the quality inspection of the LCD panel is required in the manufacturing process of the LCD panel. The traditional detection mode is manual visual detection, but the manual detection mode cannot meet the requirement of large-scale rapid production. Meanwhile, the detection precision and the detection efficiency of a manual detection mode are not high, and the labor cost is additionally increased.
Disclosure of Invention
In order to solve at least one of the above technical problems, the present invention provides a full-automatic defect detection method and system based on an LCD process, which can improve the detection efficiency and detection accuracy of a liquid crystal panel.
The invention provides a full-automatic defect detection method based on an LCD (liquid crystal display) process, which comprises the following steps:
a plurality of LCD production lines are preset, the liquid crystal panel is conveyed to a specified position through the conveying mechanism of each LCD production line and is positioned by the positioning mechanism, and meanwhile, the light source outputs detection light rays for the liquid crystal panel;
adjusting the distance between a lens of the CCD camera and the liquid crystal panel to carry out focusing processing;
scanning the identification code of the liquid crystal panel through a scanner to obtain an identification ID of the liquid crystal panel, and enabling the CCD camera to sequentially acquire and obtain a plurality of original images of the liquid crystal panel along a preset path;
carrying out deformity correction processing on each original image by an image correction module to obtain a deformity corrected image;
acquiring all malformation correction images of the liquid crystal panel by an image splicing module, carrying out image splicing treatment to obtain a global image of the liquid crystal panel, binding an identification ID with the global image, and transmitting the binding to a detection area positioning module;
the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm, and transmits the identification ID and the area to be detected to an image detection device in an associated manner;
the image detection device adopts a preset detection algorithm to perform image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines, outputs detection results, and records the identification ID and the detection results of each liquid crystal panel into a detection database.
In this scheme, adopt predetermined detection algorithm to wait to detect the region by the liquid crystal display panel that many LCD production lines provided in step by image detection device and detect the processing, specifically include:
synchronously receiving identification IDs of liquid crystal panels and areas to be detected, which are provided by a plurality of LCD production lines, by an image detection device;
presetting an integration model, wherein the integration model comprises a plurality of composition areas, the composition areas are respectively in one-to-one correspondence with the LCD production lines, and L1 composition areas are preset;
respectively filling the areas to be detected of a plurality of liquid crystal panels synchronously obtained from a plurality of LCD production lines into corresponding composition areas, and obtaining an integrated image after the filling is finished;
carrying out binarization processing on the integrated image to obtain an integrated gray image;
and calculating the integrated gray image according to a preset defect detection algorithm to obtain defect points of the integrated gray image.
In this scheme, the integrated gray image is calculated according to a preset defect detection algorithm to obtain defect points of the integrated gray image, and the method specifically includes:
transforming the position relation of each to-be-detected area of the integrated gray level image under the limitation of the integrated model according to a random algorithm to obtain a reference gray level image;
carrying out random position relation transformation for L2 times on each to-be-detected region of the integrated gray level image to obtain L2 reference gray level images;
and carrying out difference analysis processing on the gray values of the integrated gray image and the L2 reference gray images to obtain the defect points of the integrated gray image.
In this scheme, the performing a difference analysis on the gray values of the integrated gray image and L2 reference gray images to obtain defect points of the integrated gray image specifically includes:
subtracting the gray value of the pixel point corresponding to the first reference gray image from the gray value of the first pixel point in the integrated gray image to obtain a first gray difference of the first pixel point;
calculating the absolute value of the first gray difference, judging whether the absolute value of the first gray difference is greater than a first threshold value, and if so, determining a first pixel point of the integrated gray image as a primary defect reference point;
respectively carrying out gray difference calculation on the gray value of a first pixel point in the integrated gray image and the gray values of pixel points corresponding to the residual reference gray images, comparing the absolute value of the obtained gray difference with a first threshold value, and determining whether the first pixel point can be regarded as a defect reference point or not according to the comparison result;
counting the total times of the first pixel points in the integrated gray level image which are identified as defect reference points;
judging whether the total times are greater than a second threshold value or not, and if so, marking a first pixel point in the integrated gray-scale image as a defect point;
and respectively carrying out difference ratio analysis processing on the gray values of the residual pixel points of the integrated gray image and the corresponding pixel points of the L2 reference gray images to obtain all defect points in the integrated gray image.
In this scheme, after obtaining all the defect points in the integrated gray image, the method further includes:
selecting images of a plurality of regions to be detected without the defective point marks from the integrated gray level image as inspection images, wherein L3 inspection images are preset;
each test image was separately replicated as L1 identical test images;
on the basis of each inspection image, respectively filling L1 identical inspection images into the composition areas corresponding to the integrated model to respectively form L3 integrated inspection gray-scale images;
performing difference calculation on the gray value of a certain defect point in the integrated gray image and the gray values of the pixel points corresponding to the L3 integrated check gray images one by one to obtain L3 difference values;
and judging whether the absolute value of the difference value exceeding the preset proportion is larger than a first threshold value or not according to the L3 difference values, if so, checking the defect point to be passed, otherwise, judging the defect point to be an abnormal defect point, and handing the abnormal defect point to a detector for artificial defect detection.
In this scheme, all the malformation correction images of the liquid crystal panel are obtained by the image stitching module and are subjected to image stitching processing, which specifically comprises:
acquiring two adjacent deformity correction images and respectively taking the two adjacent deformity correction images as a reference image and an image to be spliced;
detecting feature points of the reference image and the image to be spliced by an SIFT algorithm;
matching the characteristic points of the acquired reference image and the characteristic points of the image to be spliced through image registration, and calculating a projection matrix by using the matched points after registration;
judging whether the projection matrix is the optimal projection matrix, if so, performing image splicing fusion to form a fused image, otherwise, continuing to perform calculation of the projection matrix by using other matched points after registration, and stopping calculation to obtain the projection matrix which is the optimal projection matrix;
and continuing to carry out image splicing on the fused image and other adjacent malformation correction images until the splicing of all the malformation correction images is finished, thereby obtaining the global image of the liquid crystal panel.
In this scheme, the detection region positioning module extracts a region to be detected from the global image based on a template matching algorithm, and specifically includes:
providing a template image A1, and performing Fourier transform operation on the template image A1 and the global image B1 to obtain a template image Fourier transform result A2 and a global image spectrum Fourier transform result B2;
respectively calculating a magnitude spectrum A3 of the template image Fourier transform result and a magnitude spectrum B3 of the global image frequency spectrum Fourier transform result;
respectively carrying out high-pass filtering on the magnitude spectrum A3 of the template image Fourier transform result and the magnitude spectrum B3 of the global image frequency spectrum Fourier transform result, converting the magnitude spectra A3 and B3 into a logarithmic-polar coordinate space, and obtaining A4 and B4, wherein A4 is the logarithmic-polar coordinate transform result of the magnitude spectrum A3, and B4 is the logarithmic-polar coordinate transform result of the magnitude spectrum B3;
calculating the relative translation between the logarithm-polar coordinate transformation result A4 and B4 by adopting a phase correlation method, and obtaining the rotation angle and the scaling coefficient between the template image A1 and the global image B1;
transforming the template image A1 according to the rotation angle and the scaling coefficient to obtain an image A5 only with translation;
calculating the translation amount between the global image B1 and the image A5 by adopting a phase correlation method;
and positioning the position of the template image in the global image B1 according to the acquired rotation angle, the acquired scaling factor and the acquired translation amount, and taking the position as a region to be detected.
The second aspect of the present invention further provides a full-automatic defect detection system based on LCD manufacturing process, comprising: the system comprises a plurality of LCD production lines and an image detection device, wherein each LCD production line comprises a conveying mechanism, a positioning mechanism, a light source, a CCD camera, a scanner, an image correction module, an image splicing module and a detection area positioning module;
the conveying mechanism is used for conveying the liquid crystal panels of each LCD production line to a specified position;
the positioning mechanism is used for positioning the liquid crystal panel at a specified position;
the light source is used for outputting detection light rays of the liquid crystal panel;
after the focusing of the lens is finished, the CCD camera sequentially collects and acquires a plurality of original images of the liquid crystal panel along a preset path;
the scanner is used for scanning the identification code of the liquid crystal panel to obtain the identification ID of the liquid crystal panel;
the image correction module is used for carrying out deformity correction processing on each original image to obtain a deformity corrected image;
the image splicing module is used for carrying out image splicing processing on all the malformation correction images of the liquid crystal panel to obtain a global image of the liquid crystal panel;
the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm;
the image detection device performs image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines by adopting a preset detection algorithm, outputs detection results, and records the identification ID and the detection results of each liquid crystal panel into a detection database.
In this scheme, adopt predetermined detection algorithm to wait to detect the region to detect the liquid crystal display panel that many LCD production lines provided in step and carry out image fusion detection processing, specifically include:
synchronously receiving identification IDs of liquid crystal panels and areas to be detected, which are provided by a plurality of LCD production lines, by an image detection device;
presetting an integration model, wherein the integration model comprises a plurality of composition areas, the composition areas are respectively in one-to-one correspondence with the LCD production lines, and L1 composition areas are preset;
respectively filling the areas to be detected of a plurality of liquid crystal panels synchronously obtained from a plurality of LCD production lines into corresponding composition areas, and obtaining an integrated image after the filling is finished;
carrying out binarization processing on the integrated image to obtain an integrated gray image;
and calculating the integrated gray image according to a preset defect detection algorithm to obtain defect points of the integrated gray image.
In this scheme, the integrated gray image is calculated according to a preset defect detection algorithm to obtain defect points of the integrated gray image, and the method specifically includes:
transforming the position relation of each to-be-detected area of the integrated gray level image under the limitation of the integrated model according to a random algorithm to obtain a reference gray level image;
carrying out random position relation transformation for L2 times on each to-be-detected region of the integrated gray level image to obtain L2 reference gray level images;
and carrying out difference analysis processing on the gray values of the integrated gray image and the L2 reference gray images to obtain the defect points of the integrated gray image.
The full-automatic defect detection method and system based on the LCD manufacturing process can effectively improve the detection efficiency and the detection precision of the Liquid Crystal Display (LCD).
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
FIG. 1 is a flow chart of a fully automatic defect detection method based on LCD manufacturing process according to the present invention;
FIG. 2 is a schematic diagram of a fully automatic defect detection system based on LCD manufacturing process according to the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those specifically described herein, and therefore the scope of the present invention is not limited by the specific embodiments disclosed below.
FIG. 1 is a flow chart of a fully automatic defect detection method based on LCD manufacturing process according to the present invention.
As shown in fig. 1, a first aspect of the present invention provides a full-automatic defect detection method based on LCD manufacturing process, the method comprising:
s102, presetting a plurality of LCD production lines, conveying the liquid crystal panel to a specified position through a conveying mechanism of each LCD production line, positioning the liquid crystal panel by a positioning mechanism, and outputting detection light rays of the liquid crystal panel by a light source;
s104, adjusting the distance between a lens of the CCD camera and the liquid crystal panel to carry out focusing processing;
s106, scanning the identification code of the liquid crystal panel through a scanner to obtain the identification ID of the liquid crystal panel, and enabling the CCD camera to sequentially acquire and obtain a plurality of original images of the liquid crystal panel along a preset path;
s108, carrying out deformity correction processing on each original image by the image correction module to obtain a deformity corrected image;
s110, acquiring all deformity correction images of the liquid crystal panel by an image splicing module, carrying out image splicing treatment to obtain a global image of the liquid crystal panel, binding an identification ID with the global image, and transmitting the binding to a detection area positioning module;
s112, the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm, and transmits the identification ID and the area to be detected to an image detection device in a correlation manner;
s114, the image detection device performs image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines by adopting a preset detection algorithm, outputs detection results, and records the identification ID and the detection results of each liquid crystal panel into a detection database.
The conveying mechanism of the present invention may be a conveyor belt or a robot, and the positioning mechanism may be a positioning jig. The feeding and discharging of the liquid crystal panel to be detected can be realized through the conveying mechanism, automatic transportation is realized, and the labor transportation cost is saved; the invention can maintain the stable posture of the liquid crystal panel through the positioning mechanism so as to facilitate the subsequent stable optical detection process. The invention uses CCD camera to replace human eye to obtain optical image of liquid crystal panel, then forms digital image after photoelectric and analog-digital conversion treatment for facilitating follow-up defect detection.
It should be noted that, because the field of view of the CCD camera is small relative to the size of the liquid crystal panel, it is difficult to cover the entire liquid crystal panel by capturing a single image with the CCD camera, and therefore, the present invention can drive the CCD camera to capture a plurality of images along a predetermined path by the driving mechanism to cover the entire liquid crystal panel. Meanwhile, in order to facilitate the overall detection of the liquid crystal panel, the invention splices a plurality of images to obtain an overall image.
According to an embodiment of the present invention, after outputting the detection result, the method further includes:
the liquid crystal panel is moved to the corresponding storage area by the conveying mechanism according to the type of the detection result.
It is understood that the detection result may include more categories, for example, classification may be performed based on the defect levels, and if there are no defect, primary defect, secondary defect, and tertiary defect, the storage area should be four to respectively store the liquid crystal panels with the corresponding defect levels.
According to an embodiment of the present invention, the performing, by the image correction module, the deformity correction processing on each original image specifically includes:
extracting characteristic points in the original image, and solving internal and external parameters and distortion coefficients of the CCD camera according to the characteristic points;
and substituting the internal and external parameters and the distortion coefficient of the CCD camera into the distortion correction model to perform image deformity correction processing on the original image.
It will be appreciated that the resulting image appears distorted due to the distorting action of the CCD camera. In the distortion of the CCD camera, when the magnification of the center of the lens is smaller than that far away from the center of the lens, namely, the closer to the center of the lens, the more curved the light rays are, and the condition is pincushion distortion; barrel distortion is the case when the power at the center of the lens is greater than the power away from the center of the lens, i.e., the closer to the center of the lens, the straighter the light rays. The method firstly calculates the distortion parameter of the distortion and then carries out the inverse transformation of the distortion, thereby correcting the malformed image.
In the image measuring process and machine vision application, in order to determine the correlation between the three-dimensional geometric position of a certain point on the surface of a space object and the corresponding point in an image, geometric models of CCD camera imaging must be established, and the parameters of the geometric models are the parameters of the CCD camera. Under most conditions, the parameters are obtained through experiments and calculation, and the process of solving the parameters is called CCD camera calibration. It can be understood that the accuracy of the calibration result of the CCD camera and the stability of the algorithm directly affect the accuracy of the result generated by the operation of the CCD camera.
According to the embodiment of the present invention, the image stitching module obtains all the deformed corrected images of the liquid crystal panel and performs image stitching processing, and the image stitching processing specifically includes:
acquiring two adjacent deformity correction images and respectively taking the two adjacent deformity correction images as a reference image and an image to be spliced;
detecting feature points of the reference image and the image to be spliced by an SIFT algorithm;
matching the characteristic points of the acquired reference image and the characteristic points of the image to be spliced through image registration, and calculating a projection matrix by using the matched points after registration;
judging whether the projection matrix is the optimal projection matrix, if so, performing image splicing fusion to form a fused image, otherwise, continuing to perform calculation of the projection matrix by using other matched points after registration, and stopping calculation to obtain the projection matrix which is the optimal projection matrix;
and continuing to carry out image splicing on the fused image and other adjacent malformation correction images until the splicing of all the malformation correction images is finished, thereby obtaining the global image of the liquid crystal panel.
It can be understood that the CCD camera of the present invention scans and photographs the liquid crystal panel according to the "Z" path under the action of the driving mechanism, that is, each time the CCD camera is moved, the CCD camera photographs the corresponding area of the liquid crystal panel, and in order to make these images contain the entire view of the liquid crystal panel, the adjacent images have an overlapping area therebetween, the present invention performs image stitching based on the overlapping area to obtain a global image. It is understood that the adjacent images referred to in the present invention refer to images captured by the CCD camera at adjacent capturing points in the capturing order of the "Z" shaped path.
It can be understood that the SIFT (Scale-invariant feature transform) algorithm is a Scale-invariant feature transform algorithm, which is used to detect and describe local features in an image, search key points (feature points) of the image on different Scale spaces, and calculate the directions of the key points. The key points searched by the SIFT can not be changed by factors such as illumination, affine transformation, noise and the like, such as corner points, edge points, bright points of dark areas, dark points of bright areas and the like.
According to an embodiment of the present invention, after obtaining the global image of the liquid crystal panel, the method further includes:
and carrying out image inclination correction and illumination unevenness correction processing on the global image.
According to an embodiment of the present invention, the image tilt correction specifically includes:
presetting a rectangular liquid crystal panel, and performing morphological closing operation processing and morphological opening operation processing on the global image to obtain a morphological optimization image;
finding out each Harris angular point of the liquid crystal panel in the morphologically optimized image through a Harris angular point function, wherein each Harris angular point surrounds the periphery of the liquid crystal panel;
connecting each corner point by adopting a minimum external rectangle mode to detect the rectangular edge of the liquid crystal panel, determining four edge corner points according to the rectangular edge, and acquiring coordinate information of the four edge corner points;
and calculating the inclination angle of the liquid crystal panel on the global image according to the coordinate information of the four edge corner points, and rotating the global image according to the inclination angle so as to perform inclination correction.
It is understood that the posture of the liquid crystal panel in the global image may be incorrect due to the inclination of the placement angle of the liquid crystal panel, and if the inclination correction processing is not performed in time, the subsequent defect identification accuracy may be affected. It should be noted that a corner point can be simply regarded as an intersection of two edges, and a more strict definition is a feature point having two main directions in a neighborhood, that is, a gray scale changes violently in two directions.
The Harris angular point function is adopted in the invention, and the purpose is to find out the angular point on the edge of the liquid crystal panel to be detected. Due to the influence of holes inside the image, before the Harris corner detection algorithm is implemented, morphological processing needs to be performed on the image to prevent the detection of the corners inside the image. Firstly, the image is binarized, and then the image is subjected to closing operation processing and opening operation processing so as to eliminate holes in the image and prevent the edge contour of the image from being changed greatly, thereby being beneficial to further corner point detection.
According to an embodiment of the present invention, the uneven lighting correction includes:
modeling a reflection imaging process according to an imaging principle to form an illumination reflection model, wherein the illumination reflection model has an expression as follows:whereinIn order to be a global image,as the component of the illumination, there is,is the reflection component of the surface of the liquid crystal panel to be detected,representing pixel point coordinates;
respectively extracting illumination components by utilizing a Gaussian function under multiple scales, and then carrying out weighted synthesis to obtain final illumination components, wherein the calculation formula of the final illumination components is as follows:whereinWhereinIn order to be the total number of scales,is shown asThe weighting coefficients of the individual scales are,the sign of the dot product operation is represented,the expression of the function of gaussian is given,which represents a normalization constant, is shown,the function of the index is expressed in terms of,is a scale factor, and the scale factors of different scales are different;
to what is neededExpression of the illuminance reflection modelIs transformed to obtainWhereinRepresenting a logarithmic function, integrating the global imageAnd final illumination component calculated under multiple scalesSubstitution intoObtaining the reflection component of the surface of the liquid crystal panel to be measuredNamely, the global image after the uneven illumination correction processing is obtained.
According to the embodiment of the present invention, the detection region positioning module extracts a region to be detected from the global image based on a template matching algorithm, and specifically includes:
providing a template image A1, and performing Fourier transform operation on the template image A1 and the global image B1 to obtain a template image Fourier transform result A2 and a global image spectrum Fourier transform result B2;
respectively calculating a magnitude spectrum A3 of the template image Fourier transform result and a magnitude spectrum B3 of the global image frequency spectrum Fourier transform result;
respectively carrying out high-pass filtering on the magnitude spectrum A3 of the template image Fourier transform result and the magnitude spectrum B3 of the global image frequency spectrum Fourier transform result, converting the magnitude spectra A3 and B3 into a logarithmic-polar coordinate space, and obtaining A4 and B4, wherein A4 is the logarithmic-polar coordinate transform result of the magnitude spectrum A3, and B4 is the logarithmic-polar coordinate transform result of the magnitude spectrum B3;
calculating the relative translation between the logarithm-polar coordinate transformation result A4 and B4 by adopting a phase correlation method, and obtaining the rotation angle and the scaling coefficient between the template image A1 and the global image B1;
transforming the template image A1 according to the rotation angle and the scaling coefficient to obtain an image A5 only with translation;
calculating the translation amount between the global image B1 and the image A5 by adopting a phase correlation method;
and positioning the position of the template image in the global image B1 according to the acquired rotation angle, the acquired scaling factor and the acquired translation amount, and taking the position as a region to be detected.
It is understood that the phase correlation method is a method for performing phase matching in the frequency domain by using fourier transform to achieve image registration. The template image matching algorithm utilizes information in an image Fourier transform magnitude spectrum and adopts a mode of combining logarithm-polar coordinate transformation and a phase correlation algorithm to calculate a rotation angle and a scaling coefficient; and transforming the template image according to the obtained result, and calculating the translation amount so as to obtain the accurate position of the template image in the global image.
According to the embodiment of the invention, the image fusion detection processing is carried out on the liquid crystal panel to-be-detected area synchronously provided by a plurality of LCD production lines by the image detection device by adopting a preset detection algorithm, and the method specifically comprises the following steps:
synchronously receiving identification IDs of liquid crystal panels and areas to be detected, which are provided by a plurality of LCD production lines, by an image detection device;
presetting an integration model, wherein the integration model comprises a plurality of composition areas, the composition areas are respectively in one-to-one correspondence with the LCD production lines, and L1 composition areas are preset;
respectively filling the areas to be detected of a plurality of liquid crystal panels synchronously obtained from a plurality of LCD production lines into corresponding composition areas, and obtaining an integrated image after the filling is finished;
carrying out binarization processing on the integrated image to obtain an integrated gray image;
and calculating the integrated gray image according to a preset defect detection algorithm to obtain defect points of the integrated gray image.
It can be understood that in the process of manufacturing the liquid crystal panel, a plurality of production lines are usually provided, and the plurality of production lines are synchronously produced, that is, each production line almost simultaneously produces one liquid crystal panel, if each production line is provided with an image detection device and carries out defect analysis, the detection efficiency is not high, in the invention, one image detection device is used for collecting images collected by a plurality of production lines, the images are combined to form an integrated image, then the integrated image is subjected to unified image defect analysis, the detection efficiency is improved, and the detection cost is saved.
It can be understood that the region to be detected actually exists in the form of a region image to be detected; because a plurality of LCD production lines respectively synchronously produce a liquid crystal panel, namely each LCD production line can generate a region to be detected of the liquid crystal panel, the regions to be detected of the liquid crystal panels of the LCD production lines are integrated through the integration model, so that an integrated image is formed, an image detection device can conveniently perform image analysis processing on the integrated image, the detection efficiency is improved, the investment of detection calculation resources is reduced, and the detection cost is saved.
According to the embodiment of the present invention, the calculating process of the integrated gray image according to a preset defect detection algorithm to obtain the defect point of the integrated gray image specifically includes:
transforming the position relation of each to-be-detected area of the integrated gray level image under the limitation of the integrated model according to a random algorithm to obtain a reference gray level image;
carrying out random position relation transformation for L2 times on each to-be-detected region of the integrated gray level image to obtain L2 reference gray level images;
and carrying out difference analysis processing on the gray values of the integrated gray image and the L2 reference gray images to obtain the defect points of the integrated gray image.
The method and the device have the advantages that the position relation of each to-be-detected area in the integrated gray image is transformed and integrated in a random mode, so that the reference gray image is obtained, and the reference gray image is used as a defect detection template for the integrated gray image.
According to an embodiment of the present invention, performing difference analysis on the gray values of the integrated gray image and L2 reference gray images to obtain defect points of the integrated gray image includes:
subtracting the gray value of the pixel point corresponding to the first reference gray image from the gray value of the first pixel point in the integrated gray image to obtain a first gray difference of the first pixel point;
calculating the absolute value of the first gray difference, judging whether the absolute value of the first gray difference is greater than a first threshold value, and if so, determining a first pixel point of the integrated gray image as a primary defect reference point;
respectively carrying out gray difference calculation on the gray value of a first pixel point in the integrated gray image and the gray values of pixel points corresponding to the residual reference gray images, comparing the absolute value of the obtained gray difference with a first threshold value, and determining whether the first pixel point can be regarded as a defect reference point or not according to the comparison result;
counting the total times of the first pixel points in the integrated gray level image which are identified as defect reference points;
judging whether the total times are greater than a second threshold value or not, and if so, marking a first pixel point in the integrated gray-scale image as a defect point;
and respectively carrying out difference ratio analysis processing on the gray values of the residual pixel points of the integrated gray image and the corresponding pixel points of the L2 reference gray images to obtain all defect points in the integrated gray image.
It can be understood that the integrated gray image and the L2 reference gray images include pixels that are in one-to-one correspondence, and particularly, in the gray difference calculation, the calculation is performed based on the pixels at the same position. Comparing a certain pixel point of the integrated gray image with pixel points at the same positions of the L2 reference gray images for L2 times respectively, counting the total times of the pixel point which is determined as a defect reference point, if the number of the pixel points is greater than a second threshold value, the difference of the gray values of the two images at the same pixel point is larger, and at least one of the two images at the pixel point is a defect point.
According to an embodiment of the invention, after obtaining all defect points in the integrated gray scale image, the method further comprises:
selecting images of a plurality of regions to be detected without the defective point marks from the integrated gray level image as inspection images, wherein L3 inspection images are preset;
each test image was separately replicated as L1 identical test images;
on the basis of each inspection image, respectively filling L1 identical inspection images into the composition areas corresponding to the integrated model to respectively form L3 integrated inspection gray-scale images;
performing difference calculation on the gray value of a certain defect point in the integrated gray image and the gray values of the pixel points corresponding to the L3 integrated check gray images one by one to obtain L3 difference values;
and judging whether the absolute value of the difference value exceeding the preset proportion is larger than a first threshold value or not according to the L3 difference values, if so, checking the defect point to be passed, otherwise, judging the defect point to be an abnormal defect point, and handing the abnormal defect point to a detector for artificial defect detection.
It is understood that the preset ratio is preferably 1/2, 2/3, but is not limited thereto. The invention further screens out the areas to be detected without the defect point marks through the integrated gray level images marked with the defect points, and optimizes and inspects the defect detection results of the integrated gray level images through the areas to be detected without the defect point marks, thereby further improving the precision of defect detection.
According to an embodiment of the present invention, after obtaining all the defect points in the integrated gray-scale image, the method further comprises:
splitting the integrated gray image into a to-be-detected area which is restored into a plurality of marked defect points according to a plurality of composition areas;
and judging whether the marked defect points exist in the to-be-detected area of each marked defect point in advance, if not, determining the identification ID of the corresponding liquid crystal panel according to the incidence relation between the to-be-detected area and the identification ID, recording that the identification ID of the corresponding liquid crystal panel is qualified, if so, evaluating the marked defect points according to a preset evaluation mode, and judging whether the corresponding liquid crystal panel is a qualified product according to the evaluation result.
According to the specific embodiment of the present invention, the evaluating the defect point of the mark according to the preset evaluating method, and determining whether the corresponding liquid crystal panel is a qualified product according to the evaluating result specifically includes:
dividing the area to be detected of the liquid crystal panel into v areas, and presetting the weight of the influence of the defect points of different areas on the quality of the liquid crystal panelWherein;
The number of defect points of a region to be detected of a certain liquid crystal panel falling into v regions is counted as;
Counting the number of defective points in each regionRespectively corresponding influence weightMultiplying to obtain the quality influence factor of each region;
Accumulating the quality influence factors of the regionsObtaining a global quality impact factor of;
Determining a global quality impact factor ofAnd whether the third threshold value is exceeded or not, if yes, the corresponding liquid crystal panel is judged to be a qualified product, and if not, the corresponding liquid crystal panel is judged to be an unqualified product.
The method and the device can analyze the influence weight of each area on the product quality, and further calculate the comprehensive quality evaluation result of the liquid crystal panel by combining the number of the defect points of each area, thereby realizing the aim of automatically detecting the liquid crystal panel.
According to the embodiment of the present invention, after the identification ID is bound to the global image and then transmitted to the detection area positioning module, the method further includes:
the detection area positioning module extracts other areas to be detected from the global image based on other template matching algorithms, and transmits the identification ID and the other areas to be detected to an image detection device in an associated manner:
the image detection device obtains the bright part of the image based on the other images of the area to be detected and after removing the background by adopting a mask methodAnd the bright part of the image;
Respectively calculating the bright parts of the imagesAnd dark part of imageComparing the two entropy values, and selecting the bright part of the image with small entropy valueOr dark part of the imageA pixel set as a K-means clustering algorithm;
adopting a K-means clustering algorithm to carry out on the bright part of the image with small entropy valueOr dark part of the imageThe pixel set is clustered into K clusters, and each cluster is provided with a clustering center;
respectively obtaining the pixel brightness of each clustering center, sequencing K clusters based on the pixel brightness of each clustering center to form a brightness sequencing queue, and removing clusters with preset proportion after the brightness sequencing queue;
respectively calculating the convexity of the connected regions in the remaining clusters, marking the connected regions with the convexity falling into the range of a fourth threshold value as primary particles, and expressing the convexity as the ratio of the area of the connected regions to the area of the corresponding cluster regions;
respectively acquiring the gray value of each primary particle, and marking the primary particles with the gray values falling into a fifth threshold range as secondary particles;
respectively obtaining the area of each secondary particle, marking the conductive particles with the secondary particles with the areas falling into the range of a sixth threshold value, and obtaining the number and the positions of the conductive particles;
and judging whether the quantity and the position of the conductive particles respectively fall into a preset quantity threshold range and a preset position deviation threshold range, if so, determining that the conductive particle pressing detection is qualified, otherwise, determining that the conductive particle pressing detection is unqualified.
It should be noted that, after the conductive particles of the liquid crystal panel are subjected to press-fit packaging, the product appearance and the conductive particles need to be detected, and in this embodiment, the bright portion and the dark portion are obtained by using a background mask removing method, so that not only background and light field information are removed, but also other interfering impurities are removed. By simultaneously dividing the conductive particles into dark and light portions, the overlapping particle spacing can be stretched. In addition, the method can discard darker background and overlapped particle connection areas by adopting a K-means clustering algorithm, then divide the particles, further confirm the particles based on particle convexity, obtain the number and the positions of the conductive particles by multi-level and multi-dimensional screening, and finally judge whether the conductive particle pressing process is qualified according to the number and the positions of the conductive particles.
It can be understood that the K-means clustering algorithm may also be referred to as a K-means clustering algorithm, which is a clustering analysis algorithm for iterative solution.
According to a specific embodiment of the present invention, the calculating the integrated gray-scale image according to a preset defect detection algorithm specifically includes:
presetting images of a plurality of regions to be detected in the integrated gray-scale image to be arranged in an array shape, and performing criss-cross segmentation on the integrated gray-scale image according to a preset period size Z, H to form M image sets, wherein Z is a transverse segmentation size, H is a longitudinal segmentation size, and M is greater than or equal to 4;
concentrating the M imagesComparing the gray value of pixel point with that of pixel point of each image set and the gray value of the remaining image sets to calculate and obtain the first image setDefect accumulation map of individual image set,,WhereinIs the coordinates of the pixel points in a single image set, and,;is as followsEach image centralizes the pixel pointsThe gray value of (a);is as followsEach image centralizes the pixel pointsThe gray value of (a);segmenting a threshold for a preset defect;,and is and;
obtaining M defect cumulative graphs respectively by comparing gray values of M image sets, and splicing the M defect cumulative graphs to obtain an integrated gray image,WhereinTo integrate the coordinates of the pixels in the gray image, anCan be based onAnd is at presentThe position of each image set in the integrated gray level image is converted;
and aggregating the defect pixel points at the adjacent positions in the integrated gray image, communicating the adjacent defect pixel points to form defect areas, and outputting the number of the defect areas and the position of each defect area to finish defect detection.
The method can effectively improve the defect detection speed by utilizing the characteristic that the integrated gray image has periodicity, segmenting the integrated gray image according to a certain period length to obtain each image set and further carrying out periodic gray value comparison to determine whether defects exist.
As can be appreciated, the first and second,andthe conversion process specifically comprises the following steps: if it is firstThe 1 st image set, namely the image set at the upper left corner in the digital images can be directly assigned with values,. If it is firstThe image set is the 2 nd image set, and the image sets corresponding to the first row and the second column after the segmentation can be assigned with values,. In this way, the conversion of the pixel point coordinates of each image set and the integrated gray level image can be performed.
In the present invention, the image sets are divided into rectangular image sets having the same size after being divided by criss-cross division according to the horizontal division size Z and the vertical division size H. First, theWhen the gray value comparison of pixel points and pixel points is respectively carried out on the image set and the residual image set, the gray value comparison is actually carried out based on the same positions of the two image setsAnd comparing the gray values of the pixel points.
It should be noted that, when image segmentation is performed, in order to ensure detection robustness, the greater the number of segments, the better the contrast effect, and the higher the detection accuracy, but to meet the high-speed detection requirement, the calculation time needs to be reduced, and the number of segments cannot be too large.
FIG. 2 is a schematic diagram of a fully automatic defect detection system based on LCD manufacturing process according to the present invention.
As shown in fig. 2, a second aspect of the present invention further provides a full-automatic defect detection system based on LCD manufacturing process, comprising: the system comprises a plurality of LCD production lines and an image detection device, wherein each LCD production line comprises a conveying mechanism, a positioning mechanism, a light source, a CCD camera, a scanner, an image correction module, an image splicing module and a detection area positioning module;
the conveying mechanism is used for conveying the liquid crystal panels of each LCD production line to a specified position;
the positioning mechanism is used for positioning the liquid crystal panel at a specified position;
the light source is used for outputting detection light rays of the liquid crystal panel;
after the focusing of the lens is finished, the CCD camera sequentially collects and acquires a plurality of original images of the liquid crystal panel along a preset path;
the scanner is used for scanning the identification code of the liquid crystal panel to obtain the identification ID of the liquid crystal panel;
the image correction module is used for carrying out deformity correction processing on each original image to obtain a deformity corrected image;
the image splicing module is used for carrying out image splicing processing on all the malformation correction images of the liquid crystal panel to obtain a global image of the liquid crystal panel;
the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm;
the image detection device performs image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines by adopting a preset detection algorithm, outputs detection results, and records the identification ID and the detection results of each liquid crystal panel into a detection database.
According to the embodiment of the invention, the image fusion detection processing is carried out on the liquid crystal panel to-be-detected areas synchronously provided by a plurality of LCD production lines by adopting a preset detection algorithm, and the method specifically comprises the following steps:
synchronously receiving identification IDs of liquid crystal panels and areas to be detected, which are provided by a plurality of LCD production lines, by an image detection device;
presetting an integration model, wherein the integration model comprises a plurality of composition areas, the composition areas are respectively in one-to-one correspondence with the LCD production lines, and L1 composition areas are preset;
respectively filling the areas to be detected of a plurality of liquid crystal panels synchronously obtained from a plurality of LCD production lines into corresponding composition areas, and obtaining an integrated image after the filling is finished;
carrying out binarization processing on the integrated image to obtain an integrated gray image;
and calculating the integrated gray image according to a preset defect detection algorithm to obtain defect points of the integrated gray image.
According to the embodiment of the present invention, the calculating process of the integrated gray image according to a preset defect detection algorithm to obtain the defect point of the integrated gray image specifically includes:
transforming the position relation of each to-be-detected area of the integrated gray level image under the limitation of the integrated model according to a random algorithm to obtain a reference gray level image;
carrying out random position relation transformation for L2 times on each to-be-detected region of the integrated gray level image to obtain L2 reference gray level images;
and carrying out difference analysis processing on the gray values of the integrated gray image and the L2 reference gray images to obtain the defect points of the integrated gray image.
The full-automatic defect detection method and system based on the LCD manufacturing process can effectively improve the detection efficiency and the detection precision of the Liquid Crystal Display (LCD).
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units; can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, all the functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Alternatively, the integrated unit of the present invention may be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or a part contributing to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, a ROM, a RAM, a magnetic or optical disk, or various other media that can store program code.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (10)
1. A full-automatic defect detection method based on LCD manufacturing process is characterized by comprising the following steps:
a plurality of LCD production lines are preset, the liquid crystal panel is conveyed to a specified position through the conveying mechanism of each LCD production line and is positioned by the positioning mechanism, and meanwhile, the light source outputs detection light rays for the liquid crystal panel;
adjusting the distance between a lens of the CCD camera and the liquid crystal panel to carry out focusing processing;
scanning the identification code of the liquid crystal panel through a scanner to obtain the identification ID of the liquid crystal panel, and enabling the CCD camera to sequentially acquire and obtain a plurality of original images of the liquid crystal panel along a Z-shaped path;
carrying out deformity correction processing on each original image by an image correction module to obtain a deformity corrected image;
acquiring all malformation correction images of the liquid crystal panel by an image splicing module, carrying out image splicing treatment to obtain a global image of the liquid crystal panel, binding an identification ID with the global image, and transmitting the binding to a detection area positioning module;
the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm, and transmits the identification ID and the area to be detected to an image detection device in an associated manner;
the image detection device adopts a preset detection algorithm to perform image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines, outputs detection results, and records the identification ID and the detection results of each liquid crystal panel into a detection database.
2. The method according to claim 1, wherein the image detection device performs image fusion detection on the to-be-detected regions of the liquid crystal panels synchronously provided by the plurality of LCD production lines by using a preset detection algorithm, and the method specifically comprises:
synchronously receiving identification IDs of liquid crystal panels and areas to be detected, which are provided by a plurality of LCD production lines, by an image detection device;
presetting an integration model, wherein the integration model comprises a plurality of composition areas, the composition areas are respectively in one-to-one correspondence with the LCD production lines, and L1 composition areas are preset;
respectively filling the areas to be detected of a plurality of liquid crystal panels synchronously obtained from a plurality of LCD production lines into corresponding composition areas, and obtaining an integrated image after the filling is finished;
carrying out binarization processing on the integrated image to obtain an integrated gray image;
and calculating the integrated gray image according to a preset defect detection algorithm to obtain defect points of the integrated gray image.
3. The method of claim 2, wherein the step of calculating the integrated gray image according to a predetermined defect detection algorithm to obtain defect points of the integrated gray image comprises:
transforming the position relation of each to-be-detected area of the integrated gray level image under the limitation of the integrated model according to a random algorithm to obtain a reference gray level image;
carrying out random position relation transformation for L2 times on each to-be-detected region of the integrated gray level image to obtain L2 reference gray level images;
and carrying out difference analysis processing on the gray values of the integrated gray image and the L2 reference gray images to obtain the defect points of the integrated gray image.
4. The method of claim 3, wherein the integrating gray image is subjected to a gray level difference analysis with respect to L2 reference gray images to obtain defect points of the integrating gray image, and the method comprises:
subtracting the gray value of the pixel point corresponding to the first reference gray image from the gray value of the first pixel point in the integrated gray image to obtain a first gray difference of the first pixel point;
calculating the absolute value of the first gray difference, judging whether the absolute value of the first gray difference is greater than a first threshold value, and if so, determining a first pixel point of the integrated gray image as a primary defect reference point;
respectively carrying out gray difference calculation on the gray value of a first pixel point in the integrated gray image and the gray values of pixel points corresponding to the residual reference gray images, comparing the absolute value of the obtained gray difference with a first threshold value, and determining whether the first pixel point can be regarded as a defect reference point or not according to the comparison result;
counting the total times of the first pixel points in the integrated gray level image which are identified as defect reference points;
judging whether the total times are greater than a second threshold value or not, and if so, marking a first pixel point in the integrated gray-scale image as a defect point;
and respectively carrying out difference ratio analysis processing on the gray values of the residual pixel points of the integrated gray image and the corresponding pixel points of the L2 reference gray images to obtain all defect points in the integrated gray image.
5. The method of claim 4, wherein after obtaining all defect points in the integrated gray image, the method further comprises:
selecting images of a plurality of regions to be detected without the defective point marks from the integrated gray level image as inspection images, wherein L3 inspection images are preset;
each test image was separately replicated as L1 identical test images;
on the basis of each inspection image, respectively filling L1 identical inspection images into the composition areas corresponding to the integrated model to respectively form L3 integrated inspection gray-scale images;
performing difference calculation on the gray value of a certain defect point in the integrated gray image and the gray values of the pixel points corresponding to the L3 integrated check gray images one by one to obtain L3 difference values;
and judging whether the absolute value of the difference value exceeding the preset proportion is larger than a first threshold value or not according to the L3 difference values, if so, checking the defect point to be passed, otherwise, judging the defect point to be an abnormal defect point, and handing the abnormal defect point to a detector for artificial defect detection.
6. The method of claim 1, wherein the image stitching module obtains all the deformed corrected images of the liquid crystal panel and performs image stitching, and the method comprises:
acquiring two adjacent deformity correction images and respectively taking the two adjacent deformity correction images as a reference image and an image to be spliced;
detecting feature points of the reference image and the image to be spliced by an SIFT algorithm;
matching the characteristic points of the acquired reference image and the characteristic points of the image to be spliced through image registration, and calculating a projection matrix by using the matched points after registration;
judging whether the projection matrix is the optimal projection matrix, if so, performing image splicing fusion to form a fused image, otherwise, continuing to perform calculation of the projection matrix by using other matched points after registration, and stopping calculation to obtain the projection matrix which is the optimal projection matrix;
and continuing to carry out image splicing on the fused image and other adjacent malformation correction images until the splicing of all the malformation correction images is finished, thereby obtaining the global image of the liquid crystal panel.
7. The method according to claim 1, wherein the detection area locating module extracts the area to be detected from the global image based on a template matching algorithm, and specifically comprises:
providing a template image A1, and performing Fourier transform operation on the template image A1 and the global image B1 to obtain a template image Fourier transform result A2 and a global image spectrum Fourier transform result B2;
respectively calculating a magnitude spectrum A3 of the template image Fourier transform result and a magnitude spectrum B3 of the global image frequency spectrum Fourier transform result;
respectively carrying out high-pass filtering on the magnitude spectrum A3 of the template image Fourier transform result and the magnitude spectrum B3 of the global image frequency spectrum Fourier transform result, converting the magnitude spectra A3 and B3 into a logarithmic-polar coordinate space, and obtaining A4 and B4, wherein A4 is the logarithmic-polar coordinate transform result of the magnitude spectrum A3, and B4 is the logarithmic-polar coordinate transform result of the magnitude spectrum B3;
calculating the relative translation between the logarithm-polar coordinate transformation result A4 and B4 by adopting a phase correlation method, and obtaining the rotation angle and the scaling coefficient between the template image A1 and the global image B1;
transforming the template image A1 according to the rotation angle and the scaling coefficient to obtain an image A5 only with translation;
calculating the translation amount between the global image B1 and the image A5 by adopting a phase correlation method;
and positioning the position of the template image in the global image B1 according to the acquired rotation angle, the acquired scaling factor and the acquired translation amount, and taking the position as a region to be detected.
8. A full-automatic defect detection system based on LCD process is characterized by comprising: the system comprises a plurality of LCD production lines and an image detection device, wherein each LCD production line comprises a conveying mechanism, a positioning mechanism, a light source, a CCD camera, a scanner, an image correction module, an image splicing module and a detection area positioning module;
the conveying mechanism is used for conveying the liquid crystal panels of each LCD production line to a specified position;
the positioning mechanism is used for positioning the liquid crystal panel at a specified position;
the light source is used for outputting detection light rays of the liquid crystal panel;
after the focusing of the lens is finished, the CCD camera sequentially collects and acquires a plurality of original images of the liquid crystal panel along a Z-shaped path;
the scanner is used for scanning the identification code of the liquid crystal panel to obtain the identification ID of the liquid crystal panel;
the image correction module is used for carrying out deformity correction processing on each original image to obtain a deformity corrected image;
the image splicing module is used for carrying out image splicing processing on all the malformation correction images of the liquid crystal panel to obtain a global image of the liquid crystal panel;
the detection area positioning module extracts an area to be detected from the global image based on a template matching algorithm;
the image detection device performs image fusion detection processing on the liquid crystal panel to-be-detected areas synchronously provided by the plurality of LCD production lines by adopting a preset detection algorithm, outputs detection results, and records the identification ID and the detection results of each liquid crystal panel into a detection database.
9. The system of claim 8, wherein a predetermined detection algorithm is used to perform image fusion detection on the areas to be detected of the liquid crystal panels synchronously provided by the plurality of LCD production lines, and the system specifically comprises:
synchronously receiving identification IDs of liquid crystal panels and areas to be detected, which are provided by a plurality of LCD production lines, by an image detection device;
presetting an integration model, wherein the integration model comprises a plurality of composition areas, the composition areas are respectively in one-to-one correspondence with the LCD production lines, and L1 composition areas are preset;
respectively filling the areas to be detected of a plurality of liquid crystal panels synchronously obtained from a plurality of LCD production lines into corresponding composition areas, and obtaining an integrated image after the filling is finished;
carrying out binarization processing on the integrated image to obtain an integrated gray image;
and calculating the integrated gray image according to a preset defect detection algorithm to obtain defect points of the integrated gray image.
10. The system of claim 9, wherein the integrated gray image is computed according to a predetermined defect detection algorithm to obtain defect points of the integrated gray image, and the method comprises:
transforming the position relation of each to-be-detected area of the integrated gray level image under the limitation of the integrated model according to a random algorithm to obtain a reference gray level image;
carrying out random position relation transformation for L2 times on each to-be-detected region of the integrated gray level image to obtain L2 reference gray level images;
and carrying out difference analysis processing on the gray values of the integrated gray image and the L2 reference gray images to obtain the defect points of the integrated gray image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111172203.5A CN113608378B (en) | 2021-10-08 | 2021-10-08 | Full-automatic defect detection method and system based on LCD (liquid crystal display) process |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111172203.5A CN113608378B (en) | 2021-10-08 | 2021-10-08 | Full-automatic defect detection method and system based on LCD (liquid crystal display) process |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113608378A CN113608378A (en) | 2021-11-05 |
CN113608378B true CN113608378B (en) | 2021-12-10 |
Family
ID=78310840
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111172203.5A Active CN113608378B (en) | 2021-10-08 | 2021-10-08 | Full-automatic defect detection method and system based on LCD (liquid crystal display) process |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113608378B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113920437B (en) * | 2021-12-14 | 2022-04-12 | 成都考拉悠然科技有限公司 | Conductive particle identification method, system, storage medium and computer equipment |
CN115079045B (en) * | 2022-06-10 | 2023-05-02 | 郴州恒维电子股份有限公司 | Automatic detection short circuit breaking device for previous procedure image |
CN115205223B (en) * | 2022-06-22 | 2023-03-14 | 锋睿领创(珠海)科技有限公司 | Visual inspection method and device for transparent object, computer equipment and medium |
CN116309574B (en) * | 2023-05-19 | 2023-08-18 | 成都数之联科技股份有限公司 | Method, system, equipment and storage medium for detecting panel leakage process defects |
CN117541563A (en) * | 2023-11-22 | 2024-02-09 | 泸州老窖股份有限公司 | Image defect detection method, device, computer equipment and medium |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108760767B (en) * | 2018-05-31 | 2020-07-17 | 电子科技大学 | Large-size liquid crystal display defect detection method based on machine vision |
CN108961238A (en) * | 2018-07-02 | 2018-12-07 | 北京百度网讯科技有限公司 | Display screen quality determining method, device, electronic equipment and storage medium |
CN110533657A (en) * | 2019-09-04 | 2019-12-03 | 苏州精速智能科技有限公司 | A kind of liquid crystal display appearance detecting method |
CN111913312A (en) * | 2020-07-08 | 2020-11-10 | 合肥壹佰度信息科技有限公司 | Device and method for inspecting liquid crystal panel of module production line |
-
2021
- 2021-10-08 CN CN202111172203.5A patent/CN113608378B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN113608378A (en) | 2021-11-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113608378B (en) | Full-automatic defect detection method and system based on LCD (liquid crystal display) process | |
CN113570605B (en) | Defect detection method and system based on liquid crystal display panel | |
CN111815630B (en) | Defect detection method and device for LCD screen | |
CN111325713B (en) | Neural network-based wood defect detection method, system and storage medium | |
CN108764257B (en) | Multi-view pointer instrument identification method | |
CN108355981B (en) | Battery connector quality detection method based on machine vision | |
US11132786B2 (en) | Board defect filtering method based on defect list and circuit layout image and device thereof and computer-readable recording medium | |
US11544916B2 (en) | Automated gauge reading and related systems, methods, and devices | |
CN109856156A (en) | A kind of display panel tiny flaw determination method and device based on AOI | |
CN113240626A (en) | Neural network-based method for detecting and classifying concave-convex flaws of glass cover plate | |
CN116559183B (en) | Method and system for improving defect judging efficiency | |
CN106340044A (en) | Camera external parameter automatic calibration method and calibration device | |
TWI669519B (en) | Board defect filtering method and device thereof and computer-readabel recording medium | |
CN111122590B (en) | Ceramic surface defect detection device and detection method | |
CN112686920A (en) | Visual measurement method and system for geometric dimension parameters of circular part | |
CN116228780B (en) | Silicon wafer defect detection method and system based on computer vision | |
Chen et al. | Evaluating fabric pilling with light-projected image analysis | |
CN115830004A (en) | Surface defect detection method, device, computer equipment and storage medium | |
CN114612418A (en) | Method, device and system for detecting surface defects of mouse shell and electronic equipment | |
CN115457541A (en) | Jewelry quality identification method and device based on image recognition | |
CN111353992A (en) | Agricultural product defect detection method and system based on textural features | |
CN112329845B (en) | Method and device for changing paper money, terminal equipment and computer readable storage medium | |
CN116580026B (en) | Automatic optical detection method, equipment and storage medium for appearance defects of precision parts | |
CN116309561B (en) | Surface defect identification method based on anti-creeping insulating material | |
CN116129365B (en) | Method and system for detecting particle materials on conveying equipment |
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 |