CN117115171A - Slight bright point defect detection method applied to subway LCD display screen - Google Patents
Slight bright point defect detection method applied to subway LCD display screen Download PDFInfo
- Publication number
- CN117115171A CN117115171A CN202311385368.XA CN202311385368A CN117115171A CN 117115171 A CN117115171 A CN 117115171A CN 202311385368 A CN202311385368 A CN 202311385368A CN 117115171 A CN117115171 A CN 117115171A
- Authority
- CN
- China
- Prior art keywords
- picture
- detected
- detection
- lcd display
- axis
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 172
- 230000007547 defect Effects 0.000 title claims abstract description 148
- 230000000007 visual effect Effects 0.000 claims abstract description 25
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 230000011218 segmentation Effects 0.000 claims description 31
- 238000000034 method Methods 0.000 claims description 23
- 238000004806 packaging method and process Methods 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 7
- 230000001502 supplementing effect Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000009826 distribution Methods 0.000 claims description 5
- 230000007613 environmental effect Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 5
- 239000004973 liquid crystal related substance Substances 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 5
- 238000003860 storage Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 238000005520 cutting process Methods 0.000 claims description 4
- 238000013135 deep learning Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 8
- 238000012545 processing Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 241001270131 Agaricus moelleri Species 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30121—CRT, LCD or plasma display
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Testing Of Optical Devices Or Fibers (AREA)
Abstract
The application relates to a light bright point defect detection method applied to a subway LCD display screen, which comprises the following steps: acquiring a screen image data set of an LCD display screen to be detected in a closed environment; performing conventional defect detection through a preset AI visual detection model based on a normal screen image dataset, and outputting a defect detection result; preprocessing an image picture in a darkroom screen image data set of the LCD display after conventional defect detection is qualified to obtain a picture set to be detected; and selecting a detection direction for moving and dividing the preprocessed picture set to be detected, and extracting and determining a slight bright point defect based on the average gray scale and the average pixel brightness of each divided image. The application has the effect of accurately and efficiently extracting the slight bright point defect.
Description
Technical Field
The application relates to the technical field of LCD defect detection, in particular to a light bright point defect detection method applied to a subway LCD display screen.
Background
Along with the development of technology, artificial intelligence is widely applied to people's life, and technologies such as data linkage and infrared induction are already applied to people's quick trip. The LED subway guide display screen is a typical case of data linkage. The system has the core that the data of various traffic guiding marks and gate linkage are realized through a centralized control system, the content to be displayed is realized through the LED display screen, and the system has high technical content, is very convenient for integrated management of a subway system, is convenient to install, and can be properly debugged and modified according to different use environments.
In the LCD display production process, various display defects, such as bright point defects, occur due to the influence of factors such as process and materials: star, inner dirt, abnormal color point, and some slight bright point defects. The slight bright point defect has the characteristics of low brightness, small area, random uncertainty which can only appear on an individual display picture, and the like, so that manual detection and CCD detection are extremely difficult, and missed detection is easy to cause, and the product quality is reduced.
The conventional detection method for the surface defect of the LCD is such as a detection method for the light bright point defect of an LCD display screen with the publication number of CN116137031A, which is to divide an image, select an ROI (region of interest) based on an ROI framework after an Otsu algorithm is adopted, and then detect whether the region in an LCD picture has the defect in the ROI. However, the above method may obtain a better detection effect in view of conventional avoidance of defects and bright point defects, but when the ROI architecture is adopted to select the picture image in view of slight bright point defects, the phenomenon of region missed selection is easily caused by the reasons of low brightness and small area of the slight bright point defects, so that the accuracy of the detection result is lower.
Disclosure of Invention
In order to solve the problems of missed detection, false detection and low detection precision caused by low brightness and small area of the light bright spot defect when the existing LCD defect detection technology faces the light bright spot defect on the LCD, the application provides a light bright spot defect detection method applied to a subway LCD display screen.
In a first aspect, the present application provides a light bright point defect detection method applied to a subway LCD display screen, which adopts the following technical scheme:
a light bright point defect detection method applied to a subway LCD display screen comprises the following steps:
acquiring a screen image data set of an LCD display screen to be detected in a closed environment, wherein the screen image data set comprises a normal screen image data set and a darkroom screen image data set;
performing conventional defect detection through a preset AI visual detection model based on a normal screen image dataset, and outputting a defect detection result;
preprocessing an image picture in a darkroom screen image data set of the LCD display after conventional defect detection is qualified to obtain a picture set to be detected;
and selecting a detection direction for moving and dividing the preprocessed picture set to be detected, and extracting and determining a slight bright point defect based on the average gray scale and the average pixel brightness of each divided image.
Preferably, the acquiring the screen image data set of the LCD display screen to be detected in the closed environment specifically includes the following steps:
performing light supplementing on the LCD display of the image to be acquired through a preset light supplementing light source in a closed environment to form a normal use environment of the display;
switching screen pictures of the LCD display into black pictures, red pictures, green pictures and blue pictures in sequence in a normal use environment of the display, and collecting four-color pictures of a screen area of the LCD display through a CCD camera and packaging to generate a normal screen image data set;
extinguishing the closed environment light source to form a darkroom environment, sequentially switching the screen picture of the LCD display into a black picture, a red picture, a green picture and a blue picture, and collecting four-color pictures of the screen area of the LCD display through a CCD camera and packaging to generate a darkroom screen image data set;
and packaging the normal screen image data set and the darkroom screen image data set to generate a screen image data set.
Preferably, the conventional defect detection is performed through a preset AI visual detection model based on the normal screen image dataset, and the defect detection result is output, which specifically includes the following steps:
inputting a four-color picture in a normal screen image data set into a preset AI visual detection model, wherein the AI visual detection model is obtained by deep learning training of a machine learning model based on a convolutional neural network through standard four-color picture sample data of an LCD (liquid crystal display);
the AI visual detection model carries out pretreatment on a black picture, a red picture, a green picture and a blue picture in a normal screen image data set and carries out rough searching on defects in the four-color picture to obtain an ROI (region of interest) of the four-color picture;
and the AI visual detection model carries out conventional defect detection on the ROI area of the four-color picture and outputs a defect detection result.
Preferably, the preprocessing of the image frames in the darkroom screen image dataset to obtain the frame set to be detected after the conventional defect detection of the LCD display is qualified specifically comprises the following steps:
after the conventional defect detection of the LCD display is qualified, cutting black pictures, green pictures, blue pictures and red pictures in a darkroom screen image data set, eliminating the frames of the LCD display and the environmental background area in the four-color pictures, and only keeping the four-color pictures of the LCD screen area;
filtering the cut four-color picture, and sequentially removing texture backgrounds in the black picture, the green picture, the blue picture and the red picture;
sharpening the four-color picture after the filtering treatment, enhancing the sharpness and the definition of possible slight bright point defects in the black picture, the green picture, the blue picture and the red picture, and obtaining a picture set to be detected.
Preferably, the moving segmentation is performed on the preprocessed to-be-detected picture set in the detection direction, and the light bright point defect is determined based on the average gray scale and the average pixel brightness of each segmented image, which specifically includes the following steps:
constructing a detection coordinate system for black pictures in a picture set to be detected, and determining X-axis and Y-axis detection directions;
obtaining the optimal dividing size and the moving amount of the LCD display to be detected, moving and dividing the LCD display to obtain a plurality of areas to be detected along the X-axis and the Y-axis from the origin of a detection coordinate system of a black picture, calculating the gray difference value of the plurality of areas to be detected, extracting and determining a slight bright point defect;
sequentially constructing a detection coordinate system for a green picture, a blue picture and a red picture in the detection picture, and determining the detection directions of an X axis and a Y axis;
moving and dividing the original point of each detection coordinate system along the X axis and the Y axis to obtain a plurality of areas to be detected corresponding to the green picture, the blue picture and the red picture, calculating pixel brightness difference values of the areas to be detected, extracting and determining slight bright point defects;
and overlapping the slight bright point defect marks of the black picture, the green picture, the blue picture and the red picture, and drawing to generate a slight bright point defect distribution map.
Preferably, the obtaining the optimal dividing size and the optimal moving amount of the LCD display to be detected specifically includes: and acquiring a light bright spot flaw size threshold of the LCD to be detected, and calculating an optimal segmentation size and a moving amount based on a preset optimal segmentation size calculation formula, wherein the moving amount is one third of the optimal segmentation size.
Preferably, the optimal segmentation size calculation formula specifically includes:
；
wherein M is the optimal dividing size of the LCD display to be detected, N is the light bright spot flaw size threshold of the LCD display to be detected, S is the number of pixels in the X-axis direction of four pictures in the face to be detected, C is the length in the X-axis direction of the four pictures in the face to be detected, and the units of N, C are all mm.
Preferably, the moving and dividing the black frame from the origin of the detection coordinate system along the two detection directions of the X axis and the Y axis to obtain a plurality of areas to be detected, and calculating the gray difference values of the plurality of areas to be detected to extract and determine the slight bright point defect specifically includes the following steps:
extracting square areas to be detected from the origin of a detection coordinate system of the black picture according to the optimal segmentation size, and capturing and obtaining a plurality of areas to be detected based on movement amount along the X axis and the Y axis in two detection directions;
dividing the image in each region to be detected along the X-axis and Y-axis directions based on the movement amount to generate region nine-grid, and measuring the average gray value of each grid in the region nine-grid;
and calculating the maximum average gray level and the minimum average gray level difference value in the region nine grids in each region to be detected, and judging that the detection region has a slight bright point defect if the difference value is larger than a preset gray level difference value threshold value.
Preferably, the moving and dividing the detecting coordinate system origin along the two detecting directions of the X axis and the Y axis to obtain a plurality of to-be-detected areas corresponding to the green picture, the blue picture and the red picture, calculating the pixel brightness difference value of each to-be-detected area, extracting and determining the slight bright point defect specifically includes the following steps:
extracting square areas to be detected from the origins of detection coordinate systems of the green picture, the blue picture and the red picture according to the optimal segmentation size, and moving and intercepting the square areas to be detected along the X axis and the Y axis based on the movement amounts to obtain a plurality of areas to be detected;
dividing the image in each region to be detected along the X-axis and Y-axis directions based on the movement amount to generate a region nine-grid, and measuring the average pixel brightness of each grid in the region nine-grid;
and calculating the difference value between the maximum average pixel brightness and the minimum average pixel brightness in the region nine grids in each region to be detected, and judging that the detection region has slight bright point defects if the difference value is larger than a preset pixel brightness difference threshold value.
In a second aspect, the present application provides a computer readable storage medium, which adopts the following technical scheme:
a computer readable storage medium storing a computer program capable of being loaded by a processor and performing any one of the methods described above.
In summary, the present application includes at least one of the following beneficial technical effects:
1. the method comprises the steps of firstly, realizing high-efficiency and accurate detection of conventional defects of an LCD screen through an AI visual detection model and a normal screen image dataset, realizing preliminary screening of the LCD, preprocessing an image picture of the LCD darkroom screen image dataset which is qualified in conventional defect detection, determining the dividing size of a detection area based on a size threshold for detecting light bright point defects and flaws of the preprocessed image picture, further moving and dividing to form a plurality of areas to be detected, determining whether the light bright point defects exist in the areas to be detected based on average gray scales and average pixel brightness of the areas to be detected, and further realizing the effect of accurately and efficiently extracting the light bright point defects;
2. the frame of the LCD display and the environmental background area in the four-color picture are removed through pretreatment, only the four-color picture of the LCD screen area is reserved, so that the subsequent selection of the segmentation detection area is facilitated, the sharpness and the definition of the slight bright point defect are enhanced while the texture background of the four-color picture is removed through a pretreatment means such as sharpening combination, and the Gao Qingwei bright point defect detection precision is facilitated;
3. the detection optimization of the light bright spot defects is realized by extracting the to-be-detected area with a small moving amount and a small range in the face of the light bright spot defects, the high speed can be realized while the interference influence is reduced through segmentation processing, and the high precision detection of the light bright spot defects can be realized by comparing the average pixel brightness of each grid in the area nine grids with the surrounding grids based on the RGB picture bright spot distinguishing rule.
Drawings
Fig. 1 is a flowchart of a method for detecting a slight bright point defect applied to a LCD panel of a subway according to an embodiment of the present application;
FIG. 2 is a flow chart of a method for acquiring a set of screen image data in an embodiment of the application;
FIG. 3 is a flow chart of a method of conventional defect detection in an embodiment of the present application;
FIG. 4 is a flowchart of a method for preprocessing a frame in an embodiment of the present application;
FIG. 5 is a flowchart of a method for extracting a slight bright spot defect by moving and dividing a picture to be detected according to an embodiment of the present application;
FIG. 6 is a schematic diagram of a frame segmentation to be detected in an embodiment of the present application;
FIG. 7 is a schematic diagram of a position relationship between a next to-be-detected area and a previous to-be-detected area in an X-axis direction during a moving and dividing process of a to-be-detected frame according to an embodiment of the present application;
FIG. 8 is a flowchart of a method for extracting a slight bright spot defect by moving and dividing a black frame to be detected according to an embodiment of the present application;
fig. 9 is a flowchart of a method for extracting a slight bright spot defect by moving and dividing a red, green and blue picture to be detected in an embodiment of the application.
Detailed Description
The present application is described in further detail below with reference to fig. 1-9.
The screen bright point generally refers to an unrepairable single color point appearing on a liquid crystal display, a point which still presents R, G, B (red, green, blue) in the case of a black screen, and a white point in only one of R or G or B in the case of switching to a red, green, blue display mode, while other color points exist in both other modes.
The embodiment of the application discloses a light bright point defect detection method applied to a subway LCD display screen. Referring to fig. 1, a light bright point defect detection method applied to a subway LCD display screen includes the steps of:
s1, acquiring an image: acquiring a screen image data set of an LCD display screen to be detected in a closed environment, wherein the screen image data set comprises a normal screen image data set and a darkroom screen image data set;
s2, conventional defect detection: performing conventional defect detection through a preset AI visual detection model based on a normal screen image dataset, and outputting a defect detection result; the AI visual detection model is obtained by deep learning training of a machine learning model based on a convolutional neural network through standard four-color picture sample data of an LCD (liquid crystal display); the specific training steps of the machine learning model to be described are not described in detail in the prior art;
s3, image preprocessing: preprocessing an image picture in a darkroom screen image data set of the LCD display after conventional defect detection is qualified to obtain a picture set to be detected;
s4, extracting and determining a slight bright point defect by moving the segmented image: and selecting a detection direction for moving and dividing the preprocessed picture set to be detected, and extracting and determining a slight bright point defect based on the average gray scale and the average pixel brightness of each divided image. The method comprises the steps of firstly realizing high-efficiency and accurate detection of conventional defects of an LCD screen through an AI visual detection model and a normal screen image dataset, realizing preliminary screening of the LCD, preprocessing an image picture of the LCD display darkroom screen image dataset qualified in conventional defect detection, determining the dividing size of a detection area based on a size threshold for detecting light bright point defects and flaws of the preprocessed image picture, further moving and dividing to form a plurality of areas to be detected, determining whether the light bright point defects exist in the areas to be detected based on average gray scales and average pixel brightness of the areas to be detected, and further realizing the effect of accurately and efficiently extracting the light bright point defects.
Referring to fig. 2, the acquiring the screen image data set of the LCD display screen to be detected in the enclosed environment specifically includes the following steps:
a1, supplementing light to form a normal use environment of the display: performing light supplementing on the LCD display of the image to be acquired through a preset light supplementing light source in a closed environment to form a normal use environment of the display;
a2, collecting four-color pictures in a normal environment: switching screen pictures of the LCD display into black pictures, red pictures, green pictures and blue pictures in sequence in a normal use environment of the display, and collecting four-color pictures of a screen area of the LCD display through a CCD camera and packaging to generate a normal screen image data set;
a3, collecting four-color pictures in a darkroom environment: extinguishing the closed environment light source to form a darkroom environment, sequentially switching the screen picture of the LCD display into a black picture, a red picture, a green picture and a blue picture, and collecting four-color pictures of the screen area of the LCD display through a CCD camera and packaging to generate a darkroom screen image data set;
a4, packaging to generate a screen image data set: and packaging the normal screen image data set and the darkroom screen image data set to generate a screen image data set. Based on the luminous principle of the LCD display and the type of bright spots, four-color pictures of the screen area of the LCD display are collected under the normal environment and the darkroom environment respectively, so that the AI visual detection model is helpful for rapidly extracting and identifying the image characteristics of the four-color pictures in the normal screen image dataset, and is helpful for accurately and efficiently detecting the conventional defects of the display; four-color pictures are collected in a darkroom environment, so that the detection and selection of light bright spot defects possibly existing in various pictures of the LCD are facilitated, the phenomenon of missing detection and false detection caused by low brightness and small area of the light bright spot defects is avoided, and the effect of effectively improving the detection precision is achieved.
Referring to fig. 3, the conventional defect detection is performed through a preset AI visual detection model based on the normal screen image dataset, and the defect detection result is output, specifically including the following steps:
b1, picture input: inputting a four-color picture in a normal screen image data set into a preset AI visual detection model, wherein the AI visual detection model is obtained by deep learning training of a machine learning model based on a convolutional neural network through standard four-color picture sample data of an LCD (liquid crystal display);
b2, preprocessing the picture: the AI visual detection model carries out pretreatment on a black picture, a red picture, a green picture and a blue picture in a normal screen image data set and carries out rough searching on defects in the four-color picture to obtain an ROI (region of interest) of the four-color picture;
the preprocessing can filter the screen image, generate a differential image based on the filtered image, and perform rough searching by utilizing the differential image to obtain an ROI (region of interest) with possible defects of the four-color picture;
b3, performing conventional defect detection on the ROI: and the AI visual detection model carries out conventional defect detection on the ROI area of the four-color picture and outputs a defect detection result. Four-color pictures in a normal screen image dataset are preprocessed through an AI visual detection model, then ROI regions possibly with defects are selected based on ROI architecture rules, and finally fine feature collection and defect inspection are carried out on the obtained ROI regions, so that the defect detection workload of the AI visual detection model can be greatly reduced, the data processing load is reduced, and the conventional defect detection precision and detection efficiency are improved.
Referring to fig. 4, after the conventional defect detection of the LCD display is qualified, preprocessing an image picture in a darkroom screen image dataset to obtain a picture set to be detected specifically includes the following steps:
c1, cutting a picture: after the conventional defect detection of the LCD display is qualified, cutting black pictures, green pictures, blue pictures and red pictures in a darkroom screen image data set, eliminating the frames of the LCD display and the environmental background area in the four-color pictures, and only keeping the four-color pictures of the LCD screen area;
and C2, picture filtering: filtering the cut four-color picture, and sequentially removing texture backgrounds in the black picture, the green picture, the blue picture and the red picture;
and C3, sharpening a picture: sharpening the four-color picture after the filtering treatment, enhancing the sharpness and the definition of possible slight bright point defects in the black picture, the green picture, the blue picture and the red picture, and obtaining a picture set to be detected. The frame of the LCD display and the environmental background area in the four-color picture are removed through pretreatment, only the four-color picture of the LCD screen area is reserved, the subsequent selection of the segmentation detection area is facilitated, the sharpness and the definition of the slight bright point defect are enhanced while the texture background of the four-color picture is removed through a pretreatment means such as sharpening combination, and the Gao Qingwei bright point defect detection precision is facilitated.
Referring to fig. 5-7, the moving segmentation is performed on the preprocessed to-be-detected image set in a detection direction, and the light bright point defect is determined based on the average gray scale and the average pixel brightness extraction of each segmented image, which specifically includes the following steps:
d1, constructing a detection coordinate system for the black picture: constructing a detection coordinate system for black pictures in a picture set to be detected, and determining X-axis and Y-axis detection directions;
d2, carrying out mobile segmentation and extraction on the black picture to determine a slight bright point defect: obtaining the optimal dividing size and the moving amount of the LCD display to be detected, moving and dividing the LCD display to obtain a plurality of areas to be detected along the X-axis and the Y-axis from the origin of a detection coordinate system of a black picture, calculating the gray difference value of the plurality of areas to be detected, extracting and determining a slight bright point defect;
d3, sequentially constructing a detection coordinate system for the green picture, the blue picture and the red picture: sequentially constructing a detection coordinate system for a green picture, a blue picture and a red picture in the detection picture, and determining the detection directions of an X axis and a Y axis;
d4, moving and dividing the green picture, the blue picture and the red picture to obtain and determine slight bright point defects: moving and dividing the original point of each detection coordinate system along the X axis and the Y axis to obtain a plurality of areas to be detected corresponding to the green picture, the blue picture and the red picture, calculating pixel brightness difference values of the areas to be detected, extracting and determining slight bright point defects;
d5, drawing a defect distribution diagram: and overlapping the slight bright point defect marks of the black picture, the green picture, the blue picture and the red picture, and drawing to generate a slight bright point defect distribution map. In the process of constructing the detection coordinate system for the four-color pictures in sequence, one corner of the picture coincides with the origin of the detection coordinate system. The detection coordinate system is sequentially constructed for the four-color picture, the to-be-detected area is extracted and segmented efficiently by means of single movement of the latest segmentation size and the movement amount in the X-axis detection direction and the Y-axis detection direction, whether light bright spot flaws exist in the to-be-detected area of the black picture or not is determined based on the gray level difference value, whether light bright spot flaws exist in the to-be-detected area of other three-color pictures or not is determined based on the pixel brightness difference value, and then the light bright spot flaws in the picture are accurately positioned based on the detection area with the light bright spot flaws, and a light bright spot defect distribution map containing all the light bright spot flaws of the LCD can be drawn by marking and overlapping the light spot flaws existing in the four-color picture, so that the effect of effectively improving the detection precision is achieved.
The obtaining the optimal dividing size and the moving amount of the LCD display to be detected specifically includes: and acquiring a light bright spot flaw size threshold of the LCD to be detected, and calculating an optimal segmentation size and a moving amount based on a preset optimal segmentation size calculation formula, wherein the moving amount is one third of the optimal segmentation size. The optimal segmentation size calculation formula specifically comprises:
；
wherein M is the optimal dividing size of the LCD display to be detected, N is the light bright spot flaw size threshold of the LCD display to be detected, S is the number of pixels in the X-axis direction of four pictures in the face to be detected, C is the length in the X-axis direction of the four pictures in the face to be detected, and the units of N, C are all mm. Based on a light bright spot flaw size threshold to be detected, namely a light bright spot flaw minimum detection size, the optimal segmentation size is calculated and determined by combining the number of picture pixels and the picture size, and basically, in order to achieve that the segmentation size is basically the same as the size of a detection object, the detection sensitivity and the processing time are considered, the picture detection time is effectively shortened while the light bright spot flaw detection precision is improved, and the effects of passing through the detection precision and the detection efficiency are achieved.
Referring to fig. 8, the steps of moving and dividing the black frame from the origin of the detection coordinate system along the two detection directions of the X axis and the Y axis to obtain a plurality of areas to be detected, calculating the gray difference values of the plurality of areas to be detected, extracting and determining the slight bright point defect specifically include the following steps:
e1, dividing a plurality of areas to be detected on the black picture based on movement amount: extracting square areas to be detected from the origin of a detection coordinate system of the black picture according to the optimal segmentation size, and capturing and obtaining a plurality of areas to be detected based on movement amount along the X axis and the Y axis in two detection directions;
e2, generating a region nine-square grid and measuring the average gray value of each grid in the region nine-square grid: dividing the image in each region to be detected along the X-axis and Y-axis directions based on the movement amount to generate region nine-grid, and measuring the average gray value of each grid in the region nine-grid;
e3, determining slight bright point defects based on the average gray level difference value of the region nine grids in each region to be detected: and calculating the maximum average gray level and the minimum average gray level difference value in the region nine grids in each region to be detected, and judging that the detection region has a slight bright point defect if the difference value is larger than a preset gray level difference value threshold value. And determining a picture extraction frame for the black picture based on the optimal segmentation size, further moving the picture extraction frame by one moving amount every time along two detection directions of an X axis and a Y axis from the origin of a detection coordinate system of the black picture, obtaining a plurality of areas to be detected, and judging whether slight bright point defects exist in each area to be detected based on the difference value between the maximum average gray level and the minimum average gray level in a nine grid of the area in each area to be detected. The detection optimization of the light bright spot defects is realized by extracting the to-be-detected area with a small moving amount and a small range in the face of the light bright spot defects, the high speed can be realized while the interference influence is reduced through segmentation processing, and the high-precision detection of the light bright spot defects can be realized by comparing each grid in the area nine grids with the surrounding Gong Gejin line average gray values based on the black picture bright spot distinguishing rule.
Referring to fig. 9, the steps of calculating the pixel brightness difference value of each to-be-detected area to extract and determine the slight bright point defect specifically include the following steps:
f1, dividing a plurality of areas to be detected on the basis of movement amount movement on a green picture, a blue picture and a red picture in sequence: extracting square areas to be detected from the origins of detection coordinate systems of the green picture, the blue picture and the red picture according to the optimal segmentation size, and moving and intercepting the square areas to be detected along the X axis and the Y axis based on the movement amounts to obtain a plurality of areas to be detected;
f2, generating a region nine-square grid and measuring the average pixel brightness of each grid in the region nine-square grid: dividing the image in each region to be detected along the X-axis and Y-axis directions based on the movement amount to generate a region nine-grid, and measuring the average pixel brightness of each grid in the region nine-grid;
f3, determining a slight bright point defect based on the average pixel brightness difference value: and calculating the difference value between the maximum average pixel brightness and the minimum average pixel brightness in the region nine grids in each region to be detected, and judging that the detection region has slight bright point defects if the difference value is larger than a preset pixel brightness difference threshold value. The detection optimization of the light bright spot defects is realized by extracting the to-be-detected area with a small moving amount and a small range in the face of the light bright spot defects, the high speed can be realized while the interference influence is reduced through segmentation processing, and the high precision detection of the light bright spot defects can be realized by comparing the average pixel brightness of each grid in the area nine grids with the surrounding grids based on the RGB picture bright spot distinguishing rule.
The embodiment of the present application also discloses a computer readable storage medium storing a computer program capable of being loaded by a processor and executing the method as described above, the computer readable storage medium for example comprising: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the scope of the present application. It will be apparent that the described embodiments are merely some, but not all, embodiments of the application. Based on these embodiments, all other embodiments that may be obtained by one of ordinary skill in the art without inventive effort are within the scope of the application. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art may still combine, add or delete features of the embodiments of the present application or make other adjustments according to circumstances without any conflict, so as to obtain different technical solutions without substantially departing from the spirit of the present application, which also falls within the scope of the present application.
Claims (10)
1. The light bright point defect detection method applied to the subway LCD display screen is characterized by comprising the following steps of:
acquiring a screen image data set of an LCD display screen to be detected in a closed environment, wherein the screen image data set comprises a normal screen image data set and a darkroom screen image data set;
performing conventional defect detection through a preset AI visual detection model based on a normal screen image dataset, and outputting a defect detection result;
preprocessing an image picture in a darkroom screen image data set of the LCD display after conventional defect detection is qualified to obtain a picture set to be detected;
and selecting a detection direction for moving and dividing the preprocessed picture set to be detected, and extracting and determining a slight bright point defect based on the average gray scale and the average pixel brightness of each divided image.
2. The method for detecting light bright point defects applied to an LCD display screen of a subway according to claim 1, wherein the step of acquiring the screen image data set of the LCD display screen to be detected in the closed environment comprises the following steps:
performing light supplementing on the LCD display of the image to be acquired through a preset light supplementing light source in a closed environment to form a normal use environment of the display;
switching screen pictures of the LCD display into black pictures, red pictures, green pictures and blue pictures in sequence in a normal use environment of the display, and collecting four-color pictures of a screen area of the LCD display through a CCD camera and packaging to generate a normal screen image data set;
extinguishing the closed environment light source to form a darkroom environment, sequentially switching the screen picture of the LCD display into a black picture, a red picture, a green picture and a blue picture, and collecting four-color pictures of the screen area of the LCD display through a CCD camera and packaging to generate a darkroom screen image data set;
and packaging the normal screen image data set and the darkroom screen image data set to generate a screen image data set.
3. The method for detecting light bright point defects applied to a subway LCD panel according to claim 2, wherein the normal screen image dataset-based conventional defect detection is performed by a preset AI visual detection model, and the defect detection result is outputted, specifically comprising the steps of:
inputting a four-color picture in a normal screen image data set into a preset AI visual detection model, wherein the AI visual detection model is obtained by deep learning training of a machine learning model based on a convolutional neural network through standard four-color picture sample data of an LCD (liquid crystal display);
the AI visual detection model carries out pretreatment on a black picture, a red picture, a green picture and a blue picture in a normal screen image data set and carries out rough searching on defects in the four-color picture to obtain an ROI (region of interest) of the four-color picture;
and the AI visual detection model carries out conventional defect detection on the ROI area of the four-color picture and outputs a defect detection result.
4. The method for detecting the light bright point defect applied to the subway LCD display screen according to claim 2, wherein the step of preprocessing the image frames in the darkroom screen image data set to obtain the frame set to be detected after the conventional defect detection of the LCD display is qualified is specifically performed comprises the following steps:
after the conventional defect detection of the LCD display is qualified, cutting black pictures, green pictures, blue pictures and red pictures in a darkroom screen image data set, eliminating the frames of the LCD display and the environmental background area in the four-color pictures, and only keeping the four-color pictures of the LCD screen area;
filtering the cut four-color picture, and sequentially removing texture backgrounds in the black picture, the green picture, the blue picture and the red picture;
sharpening the four-color picture after the filtering treatment, enhancing the sharpness and the definition of possible slight bright point defects in the black picture, the green picture, the blue picture and the red picture, and obtaining a picture set to be detected.
5. The method for detecting light bright point defects applied to a metro LCD display according to claim 2, wherein the moving segmentation is performed on the detection direction selected by the preprocessed set of pictures to be detected, and the light bright point defects are determined based on the average gray level and the average pixel brightness extraction of each segmented image, which specifically comprises the following steps:
constructing a detection coordinate system for black pictures in a picture set to be detected, and determining X-axis and Y-axis detection directions;
obtaining the optimal dividing size and the moving amount of the LCD display to be detected, moving and dividing the LCD display to obtain a plurality of areas to be detected along the X-axis and the Y-axis from the origin of a detection coordinate system of a black picture, calculating the gray difference value of the plurality of areas to be detected, extracting and determining a slight bright point defect;
sequentially constructing a detection coordinate system for a green picture, a blue picture and a red picture in the detection picture, and determining the detection directions of an X axis and a Y axis;
moving and dividing the original point of each detection coordinate system along the X axis and the Y axis to obtain a plurality of areas to be detected corresponding to the green picture, the blue picture and the red picture, calculating pixel brightness difference values of the areas to be detected, extracting and determining slight bright point defects;
and overlapping the slight bright point defect marks of the black picture, the green picture, the blue picture and the red picture, and drawing to generate a slight bright point defect distribution map.
6. The method for detecting a slight bright point defect applied to a subway LCD panel according to claim 5, wherein the obtaining the optimal dividing size and moving amount of the LCD panel to be detected comprises: and acquiring a light bright spot flaw size threshold of the LCD to be detected, and calculating an optimal segmentation size and a moving amount based on a preset optimal segmentation size calculation formula, wherein the moving amount is one third of the optimal segmentation size.
7. The method for detecting light spot defect applied to LCD panel of subway as claimed in claim 6, wherein the calculation formula of optimal dividing size is specifically as follows:
；
wherein M is the optimal dividing size of the LCD display to be detected, N is the light bright spot flaw size threshold of the LCD display to be detected, S is the number of pixels in the X-axis direction of four pictures in the face to be detected, C is the length in the X-axis direction of the four pictures in the face to be detected, and the units of N, C are all mm.
8. The method for detecting light bright point defects applied to a subway LCD display screen according to claim 6, wherein the steps of moving and dividing the subway LCD display screen from the origin of a detection coordinate system of a black picture along two detection directions of an X axis and a Y axis to obtain a plurality of areas to be detected, calculating gray scale difference values of the plurality of areas to be detected, extracting and determining the light bright point defects comprise the following steps:
extracting square areas to be detected from the origin of a detection coordinate system of the black picture according to the optimal segmentation size, and capturing and obtaining a plurality of areas to be detected based on movement amount along the X axis and the Y axis in two detection directions;
dividing the image in each region to be detected along the X-axis and Y-axis directions based on the movement amount to generate region nine-grid, and measuring the average gray value of each grid in the region nine-grid;
and calculating the maximum average gray level and the minimum average gray level difference value in the region nine grids in each region to be detected, and judging that the detection region has a slight bright point defect if the difference value is larger than a preset gray level difference value threshold value.
9. The method for detecting light spot defects on a LCD panel of a subway as claimed in claim 6, wherein the steps of moving and dividing the detected coordinate system origin along two detection directions of X-axis and Y-axis to obtain a plurality of areas to be detected corresponding to green, blue and red images, calculating the pixel brightness difference value of each area to be detected, extracting and determining the light spot defects comprise the following steps:
extracting square areas to be detected from the origins of detection coordinate systems of the green picture, the blue picture and the red picture according to the optimal segmentation size, and moving and intercepting the square areas to be detected along the X axis and the Y axis based on the movement amounts to obtain a plurality of areas to be detected;
dividing the image in each region to be detected along the X-axis and Y-axis directions based on the movement amount to generate a region nine-grid, and measuring the average pixel brightness of each grid in the region nine-grid;
and calculating the difference value between the maximum average pixel brightness and the minimum average pixel brightness in the region nine grids in each region to be detected, and judging that the detection region has slight bright point defects if the difference value is larger than a preset pixel brightness difference threshold value.
10. A computer-readable storage medium, characterized by: a computer program capable of being loaded by a processor and executing a slight bright point defect detection method applied to a metro LCD display as claimed in any one of claims 1 to 9 is stored.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311385368.XA CN117115171B (en) | 2023-10-25 | 2023-10-25 | Slight bright point defect detection method applied to subway LCD display screen |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311385368.XA CN117115171B (en) | 2023-10-25 | 2023-10-25 | Slight bright point defect detection method applied to subway LCD display screen |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117115171A true CN117115171A (en) | 2023-11-24 |
| CN117115171B CN117115171B (en) | 2024-01-26 |
Family
ID=88800640
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311385368.XA Active CN117115171B (en) | 2023-10-25 | 2023-10-25 | Slight bright point defect detection method applied to subway LCD display screen |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117115171B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117130186A (en) * | 2023-10-25 | 2023-11-28 | 深圳市晶鸿电子有限公司 | LCD display screen flaw defect intelligent detection method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103674975A (en) * | 2013-09-09 | 2014-03-26 | 中华人民共和国四川出入境检验检疫局 | Visual LCD (Liquid Crystal Display) Mura defect machine inspection method based on Gabor filtering and C-V (Chan-Vese) model segmentation |
| CN104977304A (en) * | 2015-06-26 | 2015-10-14 | 清华大学 | Device and method for detecting defects of spatial light modulator with sub-pixel structure |
| CN111524101A (en) * | 2020-04-10 | 2020-08-11 | 苏州赛腾精密电子股份有限公司 | Electronic screen defect detection method based on machine vision technology |
| CN114693656A (en) * | 2022-04-06 | 2022-07-01 | 厦门大学 | Method for detecting display defects of LED display screen and method for calibrating camera and optical filter |
| CN115131336A (en) * | 2022-07-22 | 2022-09-30 | 电子科技大学 | Dark line defect detection method under four-color background picture of display screen |
| CN116879313A (en) * | 2023-06-09 | 2023-10-13 | 重庆交通大学 | Real-time structural fracture determination method based on radar super fusion |
-
2023
- 2023-10-25 CN CN202311385368.XA patent/CN117115171B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103674975A (en) * | 2013-09-09 | 2014-03-26 | 中华人民共和国四川出入境检验检疫局 | Visual LCD (Liquid Crystal Display) Mura defect machine inspection method based on Gabor filtering and C-V (Chan-Vese) model segmentation |
| CN104977304A (en) * | 2015-06-26 | 2015-10-14 | 清华大学 | Device and method for detecting defects of spatial light modulator with sub-pixel structure |
| CN111524101A (en) * | 2020-04-10 | 2020-08-11 | 苏州赛腾精密电子股份有限公司 | Electronic screen defect detection method based on machine vision technology |
| CN114693656A (en) * | 2022-04-06 | 2022-07-01 | 厦门大学 | Method for detecting display defects of LED display screen and method for calibrating camera and optical filter |
| CN115131336A (en) * | 2022-07-22 | 2022-09-30 | 电子科技大学 | Dark line defect detection method under four-color background picture of display screen |
| CN116879313A (en) * | 2023-06-09 | 2023-10-13 | 重庆交通大学 | Real-time structural fracture determination method based on radar super fusion |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117130186A (en) * | 2023-10-25 | 2023-11-28 | 深圳市晶鸿电子有限公司 | LCD display screen flaw defect intelligent detection method |
| CN117130186B (en) * | 2023-10-25 | 2024-01-05 | 深圳市晶鸿电子有限公司 | LCD display screen flaw defect intelligent detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117115171B (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021143343A1 (en) | Method and device for testing product quality | |
| CN114140679B (en) | Defect fusion method, device, recognition system and storage medium | |
| KR20230124713A (en) | Fault detection methods, devices and systems | |
| CN111047655B (en) | High-definition camera cloth defect detection method based on convolutional neural network | |
| CN117115171B (en) | Slight bright point defect detection method applied to subway LCD display screen | |
| CN111209876A (en) | Oil leakage defect detection method and system | |
| CN115100221A (en) | Glass defect segmentation method | |
| CN113034464B (en) | Visual real-time detection method for defects of liquid crystal display under multiple backgrounds | |
| CN110599453A (en) | Panel defect detection method and device based on image fusion and equipment terminal | |
| CN113658131A (en) | Tour type ring spinning broken yarn detection method based on machine vision | |
| CN112465746A (en) | Method for detecting small defects in radiographic film | |
| CN114119591A (en) | Display screen picture quality detection method | |
| CN112308854A (en) | Automatic detection method and system for chip surface flaws and electronic equipment | |
| CN109064439B (en) | Partition-based single-side light-entering type light guide plate shadow defect extraction method | |
| CN115937101A (en) | Quality detection method, device, equipment and storage medium | |
| CN115266732A (en) | Carbon fiber tow defect detection method based on machine vision | |
| CN114428110A (en) | Method and system for detecting defects of fluorescent magnetic powder inspection image of bearing ring | |
| CN116091503B (en) | Method, device, equipment and medium for discriminating panel foreign matter defects | |
| CN112488986A (en) | Cloth surface flaw identification method, device and system based on Yolo convolutional neural network | |
| CN116228659A (en) | Visual detection method for oil leakage of EMS trolley | |
| CN115546141A (en) | Small sample Mini LED defect detection method and system based on multi-dimensional measurement | |
| CN114926429A (en) | Method, device and equipment for detecting trace length and readable storage medium | |
| JP2001028059A (en) | Method and device for color unevenness inspection | |
| JPH03175343A (en) | Method for extracting flaw by inspection appearance | |
| CN112967224A (en) | Electronic circuit board detection system, method and medium based on artificial intelligence |
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 |