Automatic detection system and method for surface defects of OLED display screen based on machine vision

Abstract

The present invention relates to the technical field of image analysis, and discloses an automatic detection system and method for surface defects of an OLED display screen based on machine vision. The system comprises a display screen surface image data processing module, a display screen surface defect detection module, and a display screen surface defect information feedback module. An intelligent search algorithm is used to compare image number values ​​according to OLED display screen surface partition defect type parameters and standard OLED display screen surface feature partition image coordinate parameters, and the OLED display screen surface partition defect coordinate parameters are accurately searched out to achieve accurate positioning of the OLED display screen surface defect position. The OLED display screen surface partition defect type parameters and the OLED display screen surface partition defect coordinate parameters are matched according to image numbers and coordinate values, accurately mapped to the standard OLED display screen surface feature image coordinate combination parameters, and displayed through a display device for feedback, so as to achieve visual and intuitive feedback of the OLED display screen surface defect type and defect position.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and in particular to a system and method for automatically detecting surface defects of an OLED display screen based on machine vision. Background Art

[0002] OLED, or organic light-emitting diode, is a new product on mobile phone OLED and is known as a "dream display". OLED display technology is different from traditional LCD display. It does not require a backlight and uses a very thin organic material coating and a glass substrate or a flexible organic substrate. When current passes through, these organic materials will emit light. Moreover, OLED display screens can be made lighter and thinner, with a larger viewing angle and can significantly save power consumption. OLED is also known as the third generation display technology. OLED is not only lighter and thinner, has low energy consumption, high brightness, good luminous efficiency, can display pure black, but can also be bent, such as today's curved screen TVs and mobile phones. The OLED display principle is essentially different from LCD. It is mainly driven by an electric field, and organic semiconductor materials and luminescent materials emit light through overcarrier injection and recombination. In essence, it is to use the ITO glass transparent electrode as the device anode and the metal electrode as the cathode. Through the power drive, electrons are transferred from the cathode to the electron transport layer, and holes are injected from the anode into the hole transport layer, and then migrate to the luminescent layer. When the two meet, excitons are generated, which excite the luminescent molecules and generate light sources after radiation. Simply put, an OLED screen is composed of hundreds of millions of "small light bulbs". The preparation process of OLED display technology has very high technical requirements. It is generally divided into pre-process and post-process. Among them, the pre-process is mainly based on photolithography and evaporation technology; the post-process is mainly based on packaging and cutting technology. Although advanced OLED technology is mastered by Samsung and LG manufacturers, many manufacturers in my country are also constantly strengthening their research on OLED screens, including Huaxing Optoelectronics, BOE, Tianma Technology, etc., which are constantly increasing their investment in OLED, and OLED products are also applied to products. Although there is a big gap compared with international giants, it has also reached a usable level. Compared with traditional LCD technology, OLED display technology has obvious advantages. The thickness of OLED screen can be controlled within 1mm, while the thickness of LCD screen is usually around 3mm, and it is lighter. The liquid structure of OLED screen can ensure the anti-aging performance of the screen, and has a wide viewing angle that LCD does not have, which can realize the same screen in a super large range without distortion. The response time is one thousandth of that of LCD screen. In addition, OLED screens are resistant to low temperatures and can display content normally in an environment of -40°C. They have higher luminous efficiency, low energy consumption, and are environmentally friendly. They can be made into curved screens. During the production and manufacturing stage of OLED displays, surface defects may occur due to raw materials, processing procedures, processing workshop environment, and processing personnel operations. Existing surface defect detection for OLED displays usually performs defect detection on the entire OLED screen, with large detection errors. It is impossible to perform defect zoning and positioning detection on the surface of the OLED display, nor is it possible to identify and feedback the type of surface defects on the OLED display, which reduces the accuracy of surface defect detection for OLED displays.

[0003] The Chinese invention patent application with publication number CN116503382A discloses a method and system for detecting scratch defects on a display screen, which obtains the segmentation threshold of a defect image; segments and preprocesses the defect image; obtains defect feature parameters; and detects scratch defects based on the defect feature parameters. The process of obtaining the segmentation threshold of the defect image is as follows: A) Calculate the gradient of the defect image: adopt the four-neighborhood method, take the current pixel point as the center, calculate the grayscale difference between the four points of the top, bottom, left, and right and the current pixel point, and take the one with the largest grayscale difference as the gradient of the current pixel point. If the difference of the one with the largest grayscale difference is less than 0, the gradient of the pixel point is 0; B) Calculate the image segmentation threshold: according to the gradient of the defect image, find the position coordinates of the pixel point when the gradient is the maximum value. If the maximum gradient exceeds the set segmentation threshold, 0.8 times the maximum gradient value is used as the image segmentation threshold. The above technical scheme cannot perform defect partition positioning detection on the display screen surface, nor can it identify and feedback the defect type on the display screen surface, and cannot achieve accurate and efficient defect detection on the display screen surface. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In order to solve the problem that the existing OLED display surface defect detection usually performs defect detection on the entire OLED screen, resulting in large detection errors, and cannot perform defect zoning and positioning detection on the surface of the OLED display, nor can it identify and feedback the surface defect types of the OLED display, which reduces the detection accuracy of the OLED display surface defects, the above purposes of accurately collecting OLED display surface image parameters online, accurately zoning and positioning the OLED display surface image, and accurately identifying and feedback the surface defect types and positions of the OLED display are achieved.

[0006] (II) Technical solution

[0007] The present invention is implemented by the following technical solution: an automatic detection method for surface defects of an OLED display screen, the method comprising the following steps:

[0008] S1, collecting surface feature image data of OLED display screen;

[0009] S2, performing image data preprocessing on the surface feature image data of the OLED display screen, and generating standard surface feature image data of the OLED display screen;

[0010] S3, performing image plane grid division processing on the surface feature image data of the standard OLED display screen to construct the surface feature partition image data of the standard OLED display screen;

[0011] S4, establishing a plane rectangular coordinate system according to the surface feature image data of the standard OLED display screen, measuring the plane coordinates of the surface feature partition image data of the standard OLED display screen in the plane rectangular coordinate system, and constructing the surface feature partition image coordinate data of the standard OLED display screen;

[0012] S5. When the standard OLED display screen surface feature partition image coordinate data is constructed, a data recognition algorithm is used to match the standard OLED display screen surface feature partition image data with the OLED display screen surface defect feature image data, and the OLED display screen surface partition defect type data is analyzed and constructed and outputted;

[0013] S6, comparing the numerical values ​​of the OLED display screen surface feature partition image numbers according to the OLED display screen surface feature partition defect type data and the standard OLED display screen surface feature partition image coordinate data, searching for the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface feature partition defect type data and identifying and generating the OLED display screen surface feature partition defect coordinate data;

[0014] S7, matching the standard OLED display screen surface feature partition image coordinate data with the standard OLED display screen surface feature image data according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data; obtaining the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data according to the OLED display screen surface feature partition image number and coordinate value matching, mapping them to the standard OLED display screen surface feature image coordinate combination data, and displaying them through a display device.

[0015] Preferably, the operation steps of collecting the surface feature image data of the OLED display screen are as follows:

[0016] S11, collecting a characteristic image of the surface of the OLED display screen online through a shooting lens and generating OLED display screen surface characteristic image data A, wherein the OLED display screen surface characteristic image data represents characteristic image data of the complete OLED display screen surface.

[0017] The present invention accurately collects surface characteristic image parameters of the OLED display screen online through a shooting lens, thereby achieving the effect of providing reliable data support for accurate detection of surface defects of the OLED display screen.

[0018] Preferably, the steps of performing image data preprocessing on the surface feature image data of the OLED display screen and generating the surface feature image data of the standard OLED display screen are as follows:

[0019] S21, using an adaptive filtering method to perform image data noise reduction preprocessing on the surface feature image data A of the OLED display screen, and generating standard surface feature image data A′ of the OLED display screen.

[0020] The present invention adopts an adaptive filtering method to perform image data noise reduction preprocessing on the collected surface characteristic image parameters of an OLED display screen, so as to achieve the effect of improving the quality of the collected surface characteristic image data of the OLED display screen.

[0021] Preferably, the steps of performing image plane grid division processing on the surface feature image data of the standard OLED display screen to construct the surface feature partition image data of the standard OLED display screen are as follows:

[0022] S31, the surface feature image data A′ of the standard OLED display screen is processed by using a square with a side length a to perform image plane grid division processing, and a standard OLED display screen surface feature partition image data set B is constructed. m ,,b μ ), m = 1, 2, 3,, μ; where b m represents the mth standard OLED display screen surface feature partition image data generated by performing image plane grid division processing on the standard OLED display screen surface feature image data A′, and μ represents the maximum value of the number of standard OLED display screen surface feature partition image numbers.

[0023] The present invention performs image plane grid division processing on the surface characteristic image parameters of the OLED display screen, so as to achieve the effect of scientific partition processing of the surface characteristic image of the OLED display screen.

[0024] Preferably, the steps of establishing a plane rectangular coordinate system according to the surface feature image data of the standard OLED display screen, measuring the plane coordinates of the surface feature partition image data of the standard OLED display screen in the plane rectangular coordinate system, and constructing the coordinate data of the surface feature partition image of the standard OLED display screen are as follows:

[0025] S41, establishing a plane rectangular coordinate system according to the standard OLED display screen surface feature image data A′, with the horizontal side of the right-angle side of the corresponding image as the horizontal coordinate axis and the vertical side as the vertical coordinate axis;

[0026] S42, measuring the standard OLED display screen surface feature partition image data b1 to b2 in the standard OLED display screen surface feature partition image data set B. μ The plane coordinates of the center point of the corresponding partition image in the plane rectangular coordinate system are used to construct the standard OLED display surface feature partition image coordinate data set C = (c1,,c m ,,c μ ), where cm Represents the surface feature partition image data of a standard OLED display screen b m The image coordinate data of the surface feature partition of the standard OLED display screen corresponding to the plane rectangular coordinate system, c m =(x m ,y m ), where x m Represents the surface feature partition image data of a standard OLED display screen b m The horizontal coordinate in the rectangular coordinate system, y m Represents the surface feature partition image data of a standard OLED display screen b m The ordinate in a rectangular coordinate system.

[0027] The present invention establishes a plane rectangular coordinate system to accurately measure the coordinate data of the surface feature partition image of the OLED display screen, thereby providing data support for subsequent surface defect type identification and defect location positioning of the OLED display screen.

[0028] Preferably, when the standard OLED display screen surface feature partition image coordinate data is constructed, the data recognition algorithm is used to match the standard OLED display screen surface feature partition image data with the OLED display screen surface defect feature image data, and the operation steps of analyzing and constructing the OLED display screen surface partition defect type data and outputting it are as follows:

[0029] S51, establish the OLED display surface defect feature image data set D = (d1,,d n ,,d ν ), n = 1, 2, 3,, ν; where d n represents the OLED display screen surface defect characteristic image data corresponding to the nth OLED display screen surface defect type, and ν represents the maximum number of OLED display screen surface defect types; the OLED display screen surface defect types include any one or more of scratch defects, pixel defects, bright spot defects, dark spot defects, and color deviation defects;

[0030] S52, when the standard OLED display screen surface feature partition image coordinate data set C is constructed, the standard OLED display screen surface feature partition image data set B is used to identify the standard OLED display screen surface feature partition image data set b m According to the standard OLED display screen surface feature partition image number m, the OLED display screen surface defect feature image data set D is sequentially compared with the OLED display screen surface defect feature image data d nImage feature matching is performed to analyze and construct the surface partition defect type data of the OLED display screen. The specific operation steps of the data recognition algorithm to analyze and construct the surface partition defect type data of the OLED display screen are as follows:

[0031] S521, initializing parameters, updating the number of defect type identification crow populations, the maximum number of iterations, and the flight distance Π;

[0032] S522, initializing the initial position and memory of the defect type recognition crow individual in the search space of the OLED display surface defect feature image data set D, N defect type recognition crow individuals are randomly distributed in a multidimensional search space, that is, N defect type recognition crow individuals are randomly distributed in the search space of the OLED display surface defect feature image data set D with a spatial dimension of ν; in the first iteration, it is assumed that the defect type recognition crow individual hides the food at the initial position; that is, in the first iteration, it is assumed that the defect type recognition crow individual hides the food at the initial position in the search space of the OLED display surface defect feature image data set D;

[0033] S523, calculating the individual fitness value of each defect type recognition crow, that is, calculating the standard OLED display screen surface feature partition image data set B of the standard OLED display screen surface feature partition image data set b m According to the standard OLED display surface feature partition image number m, the OLED display surface defect feature image data set D is sorted and the OLED display surface defect feature image data d n The fitness value at the initial position after matching;

[0034] S524, updating the position of the defect type identification crow individual in the search space of the OLED display surface defect feature image data set D, and the defect type identification crow individual position update formula is as follows: in represents the new position of the crow individual i in the search space of the OLED display surface defect feature image data set D in the t+1th iteration defect type identification, represents the position of the crow individual i in the search space D of the OLED display surface defect feature image data set for the t-th iteration defect type identification, Θ represents a random number uniformly distributed between

[01] , represents the flight distance of the crow individual i in the search space of the OLED display surface defect feature image data set D in the t-th iteration defect type identification, It represents the hidden location of food in the search space of the OLED display surface defect feature image data set D for the crow individual i in the t-th iteration defect type identification;

[0035] S525, determine the feasibility of the new position, and determine the feasibility of the new position of each defect type identification crow individual; if the new position of the defect type identification crow individual is feasible, the defect type identification crow individual will update its position, that is, search for the OLED display surface defect feature image data set D that is consistent with the standard OLED display surface feature partition image data b m Matching OLED display surface defect feature image data d n , the defect type identification crow individual is updated to the successfully matched OLED display surface defect feature image data d n position; otherwise, the defect type identifies that the crow individual stays at the current position and does not move to the new position;

[0036] S526, evaluate the fitness value of the new position, calculate the fitness value of the new position of each defect type recognition crow individual, that is, calculate the standard OLED display surface feature partition image data b in the standard OLED display surface feature partition image data set B m According to the standard OLED display surface feature partition image number m, the OLED display surface defect feature image data set D is sorted and the OLED display surface defect feature image data d n The fitness value at the new position after matching;

[0037] S527, updating memory. If the fitness value of the new position of the defect type recognition crow is greater than the fitness value of the initial position in the memory, the defect type recognition crow updates its memory through the new position, that is, searching for the standard OLED display surface feature partition image data set B that is consistent with the standard OLED display surface feature partition image data b. m OLED display surface defect feature image data d with the largest fitness value n ;

[0038] S528: When the maximum number of iterations is met, output the standard OLED display screen surface feature partition image data b m OLED display surface defect feature image data d n After matching, the corresponding OLED display screen surface partition defect type data is established to establish the OLED display screen surface partition defect type data set B'=(b' m1 ,,b' m2 ), 1≤m1≤m≤m2≤μ, where b' m1 Represents the surface feature partition image data of a standard OLED display screen b m1 Corresponding OLED display surface partition defect type data; b' m2 Represents the surface feature partition image data of a standard OLED display screen b m2Corresponding OLED display screen surface partition defect type data;

[0039] S53, outputting the OLED display screen surface partition defect type data set B'.

[0040] The present invention adopts the crow optimization algorithm of machine vision to match the surface feature partition image parameters of the standard OLED display screen with the surface defect feature image parameters of the OLED display screen, accurately analyzes and outputs the surface partition defect type of the OLED display screen, and achieves the effect of accurate detection of the surface defect type of the OLED display screen and autonomous output of the detection results.

[0041] Preferably, the operation steps of performing numerical comparison of OLED display screen surface feature partition image numbers based on the OLED display screen surface feature partition defect type data and the standard OLED display screen surface feature partition image coordinate data, searching for the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface feature partition defect type data, and identifying and generating the OLED display screen surface feature partition defect coordinate data are as follows:

[0042] S61, using a KD tree nearest neighbor search algorithm to search the OLED display screen surface partition defect type data set B' for the OLED display screen surface partition defect type data set B'. m1 to b' m2 The standard OLED display screen surface feature partition image data set B has the same characteristics as the standard OLED display screen surface feature partition image data b m Perform numerical comparison of the image number m of the surface feature partition of the OLED display screen, and search for the surface partition defect type data b' of the OLED display screen m1 to b' m2 The corresponding standard OLED display screen surface feature partition image data is used to identify and generate the defective OLED display screen surface feature partition image data set E = (e m1 ,,e m2 ), where e m1 Indicates the OLED display surface partition defect type data b' m1 Corresponding defective OLED display surface feature partition image data, e m2 Indicates the OLED display surface partition defect type data b' m2 Corresponding defective OLED display screen surface feature partition image data;

[0043] S62, using a KD tree nearest neighbor search algorithm to search the defective OLED display screen surface feature partition image data set E for the defective OLED display screen surface feature partition image data set E. m1 To e m2The standard OLED display screen surface feature partition image coordinate data c in the standard OLED display screen surface feature partition image coordinate data set C m According to the comparison of the OLED display screen surface feature partition image number m, search for the defective OLED display screen surface feature partition image data e m1 To e m2 The corresponding standard OLED display screen surface feature partition image coordinate data, and the generated OLED display screen surface partition defect coordinate data set F = (f m1 ,,f m2 ), where f m1 Represents the defective OLED display surface feature partition image data e m1 Corresponding OLED display surface partition defect coordinate data, f m2 Represents the defective OLED display surface feature partition image data e m2 Corresponding OLED display surface partition defect coordinate data.

[0044] The present invention uses the KD tree nearest neighbor search algorithm to compare the image number values ​​according to the surface partition defect type parameters of the OLED display screen and the surface feature partition image coordinate parameters of the standard OLED display screen, and accurately searches for the surface partition defect coordinate parameters of the OLED display screen, thereby achieving the effect of accurately locating the surface defect position of the OLED display screen.

[0045] Preferably, the standard OLED display screen surface feature partition image coordinate data and the standard OLED display screen surface feature image data are matched according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data; the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data are obtained and mapped to the standard OLED display screen surface feature image coordinate combination data according to the OLED display screen surface feature partition image number and coordinate value matching and displayed through a display device as feedback in the following operating steps:

[0046] S71, matching the standard OLED display screen surface feature partition image coordinate data set C with the standard OLED display screen surface feature image data A′ according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data G;

[0047] S72, obtaining a data set B' of surface partition defect types of the OLED display screen and a data set F of surface partition defect coordinates of the OLED display screen;

[0048] S73, mapping the OLED display screen surface partition defect type data in the OLED display screen surface partition defect type data set B' and the OLED display screen surface partition defect coordinate data in the OLED display screen surface partition defect coordinate data set F to the standard OLED display screen surface feature image coordinate combination data G according to the OLED display screen surface feature partition image number m and coordinate value matching, and displaying the mapped standard OLED display screen surface feature image coordinate combination data G through a display device, wherein the display device includes any one of a display screen, a screen, and a mobile terminal.

[0049] The present invention constructs standard OLED display surface feature image coordinate combination parameters and matches OLED display surface partition defect type parameters and OLED display surface partition defect coordinate parameters according to image numbers and coordinate values, accurately maps them to standard OLED display surface feature image coordinate combination parameters and displays them through display device feedback, thereby achieving visual and intuitive feedback on the surface defect type and defect position of the OLED display, thereby improving the intuitiveness and applicability of the OLED display surface defect detection results.

[0050] An OLED display screen surface defect automatic detection system based on machine vision for realizing the OLED display screen surface defect automatic detection method, the system comprising a display screen surface image data processing module, a display screen surface defect detection module, and a display screen surface defect information feedback module;

[0051] The display screen surface image data processing module includes a display screen surface feature image acquisition unit, a display screen surface feature image preprocessing unit, a display screen surface feature image partition processing unit, and a display screen surface feature image partition positioning unit;

[0052] The display screen surface feature image acquisition unit acquires surface feature image data of the OLED display screen through a shooting lens; the display screen surface feature image preprocessing unit performs image data preprocessing on the surface feature image data of the OLED display screen, and generates surface feature image data of a standard OLED display screen; the display screen surface feature image partition processing unit performs image plane grid division processing on the surface feature image data of the standard OLED display screen, and constructs surface feature partition image data of a standard OLED display screen; the display screen surface feature image partition positioning unit establishes a plane rectangular coordinate system according to the surface feature image data of the standard OLED display screen, measures the plane coordinates of the surface feature partition image data of the standard OLED display screen in the plane rectangular coordinate system, and constructs the coordinate data of the surface feature partition image of the standard OLED display screen;

[0053] The display screen surface defect detection module includes a display screen surface defect feature image storage unit, a display screen surface defect type recognition unit, and a display screen surface defect type recognition result output unit;

[0054] The display screen surface defect feature image storage unit is used to store the OLED display screen surface defect feature image data; the display screen surface defect type identification unit uses a machine vision data identification algorithm to match the standard OLED display screen surface feature partition image data with the OLED display screen surface defect feature image data, and analyzes and constructs the OLED display screen surface partition defect type data; the display screen surface defect type identification result output unit is used to output the OLED display screen surface partition defect type data;

[0055] The display screen surface defect information feedback module includes a display screen surface defect type position search unit, a display screen surface feature image information and partition image coordinate processing unit, a display screen surface defect information acquisition unit, and a display screen surface defect information feedback display unit;

[0056] The display screen surface defect type position search unit compares the OLED display screen surface feature partition image number value with the standard OLED display screen surface feature partition image coordinate data according to the OLED display screen surface partition defect type data and the standard OLED display screen surface feature partition image coordinate data, searches out the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface partition defect type data, and identifies and generates the OLED display screen surface partition defect coordinate data; the display screen surface feature image information and partition image coordinate processing unit matches the standard OLED display screen surface feature partition image coordinate data with the standard OLED display screen surface feature image data according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data; the display screen surface defect information acquisition unit is used to acquire the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data; the display screen surface defect information feedback display unit matches the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data according to the OLED display screen surface feature partition image number and coordinate value to the standard OLED display screen surface feature image coordinate combination data, and displays the feedback through the display device.

[0057] (III) Beneficial effects

[0058] The present invention provides an automatic detection system and method for surface defects of an OLED display screen based on machine vision, which has the following beneficial effects:

[0059] 1. Through the cooperation between the display screen surface feature image acquisition unit and the display screen surface feature image preprocessing unit, the surface feature image parameters of the OLED display screen are accurately collected online through the shooting lens, providing data support for the accurate detection of surface defects of the OLED display screen, and performing image data noise reduction preprocessing on the collected surface feature image parameters of the OLED display screen, thereby improving the quality of the collected surface feature image data of the OLED display screen; the display screen surface feature image partition processing unit performs image plane grid division processing on the surface feature image parameters of the OLED display screen, realizes scientific partition processing of the surface feature image of the OLED display screen, and improves the detection accuracy of surface defects of the OLED display screen; the display screen surface feature image partition positioning unit establishes a plane rectangular coordinate system to accurately measure the coordinate data of the surface feature partition image of the OLED display screen, realizes the accurate positioning of the surface partition image of the OLED display screen, provides data support for the subsequent surface defect type identification and defect position positioning of the OLED display screen, and improves the efficiency and accuracy of the surface defect identification results of the OLED display screen.

[0060] 2. Through the display screen surface defect feature image storage unit, the surface defect feature image parameters of the OLED display screen are scientifically preset and stored to realize the scientific and standardized operation of OLED display screen surface defect detection; the display screen surface defect type recognition unit and the display screen surface defect type recognition result output unit cooperate with each other, and use the intelligent recognition algorithm of machine vision to match the image features of the standard OLED display screen surface feature partition image parameters with the OLED display screen surface defect feature image parameters, accurately analyze and output the OLED display screen surface partition defect type, and realize the accurate detection of OLED display screen surface defect type and autonomous output of detection results.

[0061] 3. Through the display screen surface defect type position search unit, an intelligent search algorithm is used to compare the image number value based on the OLED display screen surface partition defect type parameter and the standard OLED display screen surface feature partition image coordinate parameter, and the OLED display screen surface partition defect coordinate parameter is accurately searched out to achieve accurate positioning of the OLED display screen surface defect position; the display screen surface feature image information and partition image coordinate processing unit and the display screen surface defect information feedback display unit cooperate with each other to match the standard OLED display screen surface feature partition image coordinate parameter with the standard OLED display screen surface feature image parameter reference plane rectangular coordinate system, and construct the standard OLED display screen surface feature image coordinate combination parameter, which provides data support for the OLED display screen surface defect type recognition result and defect position visualization display; the display screen surface defect information feedback display unit matches the OLED display screen surface partition defect type parameter and the OLED display screen surface partition defect coordinate parameter according to the image number and coordinate value, accurately maps them to the standard OLED display screen surface feature image coordinate combination parameter, and displays them through the display device feedback, so as to achieve visual and intuitive feedback of the OLED display screen surface defect type and defect position, thereby improving the intuitiveness and applicability of the OLED display screen surface defect detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A schematic diagram of a module of an automatic detection system for surface defects of an OLED display screen based on machine vision provided by the present invention;

[0063] Figure 2 This is a flow chart of the method for automatically detecting surface defects of an OLED display screen provided by the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The embodiments of the automatic detection system and method of OLED display surface defects based on machine vision are as follows:

[0066] See also Figure 1 - Figure 2 , an automatic detection method for surface defects of an OLED display screen, the method comprising the following steps:

[0067] S1, collecting surface feature image data of OLED display screen;

[0068] S2, performing image data preprocessing on the surface feature image data of the OLED display screen, and generating standard surface feature image data of the OLED display screen;

[0069] S3, performing image plane grid division processing on the surface feature image data of the standard OLED display screen to construct the surface feature partition image data of the standard OLED display screen;

[0070] S4, establishing a plane rectangular coordinate system according to the surface feature image data of the standard OLED display screen, measuring the plane coordinates of the surface feature partition image data of the standard OLED display screen in the plane rectangular coordinate system, and constructing the surface feature partition image coordinate data of the standard OLED display screen;

[0071] S5. When the standard OLED display screen surface feature partition image coordinate data is constructed, a data recognition algorithm is used to match the standard OLED display screen surface feature partition image data with the OLED display screen surface defect feature image data, and the OLED display screen surface partition defect type data is analyzed and constructed and output;

[0072] S6, comparing the numerical values ​​of the OLED display screen surface feature partition image numbers according to the OLED display screen surface feature partition defect type data and the standard OLED display screen surface feature partition image coordinate data, searching for the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface feature partition defect type data and identifying and generating the OLED display screen surface feature partition defect coordinate data;

[0073] S7, matching the standard OLED display screen surface feature partition image coordinate data with the standard OLED display screen surface feature image data according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data; obtaining the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data, mapping them to the standard OLED display screen surface feature image coordinate combination data according to the OLED display screen surface feature partition image number and coordinate value matching, and displaying them through a display device.

[0074] For further information, see Figure 1 - Figure 2 ,The steps for collecting the surface feature image data of the OLED display are as follows:

[0075] S11. Capturing a characteristic image of the surface of the OLED display screen online through a shooting lens and generating OLED display screen surface characteristic image data A, wherein the OLED display screen surface characteristic image data represents characteristic image data of the complete OLED display screen surface.

[0076] The steps for preprocessing the image data of the surface feature image data of the OLED display screen and generating the surface feature image data of the standard OLED display screen are as follows:

[0077] S21, using an adaptive filtering method to perform image data noise reduction preprocessing on the surface feature image data A of the OLED display screen, and generating standard surface feature image data A′ of the OLED display screen.

[0078] The steps of performing image plane grid division processing on the surface feature image data of the standard OLED display screen and constructing the surface feature partition image data of the standard OLED display screen are as follows:

[0079] S31, the surface feature image data A′ of the standard OLED display screen is divided into an image plane grid using a square with a side length a, and a standard OLED display screen surface feature partition image data set B is constructed. m ,,b μ ), m = 1, 2, 3,, μ; where b m represents the mth standard OLED display surface feature partition image data generated by performing image plane grid division processing on the standard OLED display surface feature image data A′, and μ represents the maximum value of the number of standard OLED display surface feature partition image numbers.

[0080] The steps of establishing a plane rectangular coordinate system based on the surface feature image data of the standard OLED display screen, measuring the plane coordinates of the surface feature partition image data of the standard OLED display screen in the plane rectangular coordinate system, and constructing the coordinate data of the surface feature partition image of the standard OLED display screen are as follows:

[0081] S41, establishing a plane rectangular coordinate system according to the standard OLED display screen surface feature image data A′, with the horizontal side of the right-angle side of the corresponding image as the horizontal coordinate axis and the vertical side as the vertical coordinate axis;

[0082] S42, measuring the standard OLED display screen surface feature partition image data b1 to b2 in the standard OLED display screen surface feature partition image data set B. μ The plane coordinates of the center point of the corresponding partition image in the plane rectangular coordinate system are used to construct the standard OLED display surface feature partition image coordinate data set C = (c1,,c m ,,c μ ), where c m Represents the surface feature partition image data of a standard OLED display screen b m The image coordinate data of the surface feature partition of the standard OLED display screen corresponding to the plane rectangular coordinate system, c m =(x m ,y m), where x m Represents the surface feature partition image data of a standard OLED display screen b m The horizontal coordinate in the rectangular coordinate system, y m Represents the surface feature partition image data of a standard OLED display screen b m The ordinate in a rectangular coordinate system.

[0083] The display screen surface feature image acquisition unit and the display screen surface feature image preprocessing unit cooperate with each other, and the surface feature image parameters of the OLED display screen are accurately collected online through the shooting lens, so as to provide data support for the accurate detection of surface defects of the OLED display screen, and the image data noise reduction preprocessing is performed on the collected surface feature image parameters of the OLED display screen, thereby improving the quality of the collected surface feature image data of the OLED display screen; the display screen surface feature image partition processing unit performs image plane grid division processing on the surface feature image parameters of the OLED display screen, realizes scientific partition processing of the surface feature image of the OLED display screen, and improves the detection accuracy of surface defects of the OLED display screen; the display screen surface feature image partition positioning unit establishes a plane rectangular coordinate system to accurately measure the coordinate data of the surface feature partition image of the OLED display screen, realizes the accurate positioning of the surface partition image of the OLED display screen, provides data support for the subsequent surface defect type identification and defect position positioning of the OLED display screen, and improves the efficiency and accuracy of the surface defect identification result of the OLED display screen.

[0084] For further information, see Figure 1 - Figure 2 When the standard OLED display screen surface feature partition image coordinate data is constructed, the data recognition algorithm is used to match the standard OLED display screen surface feature partition image data with the OLED display screen surface defect feature image data, and the steps of analyzing and constructing the OLED display screen surface partition defect type data and outputting it are as follows:

[0085] S51, establish the OLED display surface defect feature image data set D = (d1,,d n ,,d ν ), n = 1, 2, 3,, ν; where d n represents the OLED display surface defect characteristic image data corresponding to the nth OLED display surface defect type, ν represents the maximum number of OLED display surface defect types; the OLED display surface defect types include any one or more of scratch defects, pixel defects, bright spot defects, dark spot defects, and color deviation defects;

[0086] S52, when the standard OLED display screen surface feature partition image coordinate data set C is constructed, the standard OLED display screen surface feature partition image data set B is used to identify the standard OLED display screen surface feature partition image data b m According to the standard OLED display surface feature partition image number m, the OLED display surface defect feature image data set D is sorted and the OLED display surface defect feature image data d n Image feature matching is performed to analyze and construct the surface partition defect type data of the OLED display screen. The specific operation steps of the data recognition algorithm to analyze and construct the surface partition defect type data of the OLED display screen are as follows:

[0087] S521, initializing parameters, updating the number of defect type identification crow populations, the maximum number of iterations, and the flight distance Π;

[0088] S522, initializing the initial position and memory of the defect type recognition crow individual in the search space of the OLED display surface defect feature image data set D, N defect type recognition crow individuals are randomly distributed in a multidimensional search space, that is, N defect type recognition crow individuals are randomly distributed in the search space of the OLED display surface defect feature image data set D with a spatial dimension of ν; in the first iteration, it is assumed that the defect type recognition crow individual hides the food at the initial position; that is, in the first iteration, it is assumed that the defect type recognition crow individual hides the food at the initial position in the search space of the OLED display surface defect feature image data set D;

[0089] S523, calculating the individual fitness value of each defect type recognition crow, that is, calculating the standard OLED display screen surface feature partition image data set B of the standard OLED display screen surface feature partition image data set b m According to the standard OLED display surface feature partition image number m, the OLED display surface defect feature image data set D is sorted and the OLED display surface defect feature image data d n The fitness value at the initial position after matching;

[0090] S524, updating the position of the defect type identification crow individual in the search space of the OLED display surface defect feature image data set D, and the defect type identification crow individual position update formula is as follows: in represents the new position of the crow individual i in the search space of the OLED display surface defect feature image data set D in the t+1th iteration defect type identification, represents the position of the crow individual i in the search space D of the OLED display surface defect feature image data set for the t-th iteration defect type identification, Θ represents a random number uniformly distributed between

[01] , represents the flight distance of the crow individual i in the search space of the OLED display surface defect feature image data set D in the t-th iteration defect type identification, It represents the hidden location of food in the search space of the OLED display surface defect feature image data set D for the crow individual i in the t-th iteration defect type identification;

[0091] S525, determine the feasibility of the new position, and determine the feasibility of the new position of each defect type identification crow individual; if the new position of the defect type identification crow individual is feasible, the defect type identification crow individual will update its position, that is, search for the OLED display surface defect feature image data set D that is consistent with the standard OLED display surface feature partition image data b m Matching OLED display surface defect feature image data d n , the defect type identification crow individual is updated to the successfully matched OLED display surface defect feature image data d n position; otherwise, the defect type identifies that the crow individual stays at the current position and does not move to the new position;

[0092] S526, evaluate the fitness value of the new position, calculate the fitness value of the new position of each defect type recognition crow individual, that is, calculate the standard OLED display surface feature partition image data b in the standard OLED display surface feature partition image data set B m According to the standard OLED display screen surface feature partition image number m, the OLED display screen surface defect feature image data set D is sorted and the OLED display screen surface defect feature image data d n The fitness value at the new position after matching;

[0093] S527, updating memory. If the fitness value of the new position of the defect type recognition crow is greater than the fitness value of the initial position in the memory, the defect type recognition crow updates its memory through the new position, that is, searching for the standard OLED display surface feature partition image data set B that is consistent with the standard OLED display surface feature partition image data b. m OLED display surface defect feature image data d with the largest fitness value n ;

[0094] S528: When the maximum number of iterations is met, output the standard OLED display screen surface feature partition image data b m OLED display surface defect feature image data d nAfter matching, the corresponding OLED display screen surface partition defect type data is established to establish the OLED display screen surface partition defect type data set B'=(b' m1 ,,b' m2 ), 1≤m1≤m≤m2≤μ, where b' m1 Represents the surface feature partition image data of a standard OLED display screen b m1 Corresponding OLED display surface partition defect type data; b' m2 Represents the surface feature partition image data of a standard OLED display screen b m2 Corresponding OLED display screen surface partition defect type data;

[0095] S53, outputting the OLED display screen surface partition defect type data set B'.

[0096] Through the display screen surface defect feature image storage unit, the surface defect feature image parameters of the OLED display screen are scientifically preset and stored to realize the scientific and standardized operation of OLED display screen surface defect detection; the display screen surface defect type recognition unit and the display screen surface defect type recognition result output unit cooperate with each other, and use the intelligent recognition algorithm of machine vision to match the image features of the standard OLED display screen surface feature partition image parameters with the OLED display screen surface defect feature image parameters, accurately analyze and output the OLED display screen surface partition defect type, and realize the accurate detection of OLED display screen surface defect type and autonomous output of detection results.

[0097] For further information, see Figure 1 - Figure 2 , the operation steps of comparing the numerical values ​​of the OLED display screen surface feature partition image numbers with the standard OLED display screen surface feature partition image coordinate data according to the OLED display screen surface partition defect type data and searching for the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface partition defect type data and marking and generating the OLED display screen surface partition defect coordinate data are as follows:

[0098] S61, using the KD tree nearest neighbor search algorithm to search the OLED display screen surface partition defect type data set B' for the OLED display screen surface partition defect type data set b'. m1 to b' m2 The standard OLED display screen surface feature partition image data b in the standard OLED display screen surface feature partition image data set B m Perform numerical comparison of the image number m of the surface feature partition of the OLED display screen, and search for the surface partition defect type data b' of the OLED display screen m1 to b' m2The corresponding standard OLED display screen surface feature partition image data is used to identify and generate the defective OLED display screen surface feature partition image data set E = (e m1 ,,e m2 ), where e m1 Indicates the OLED display surface partition defect type data b' m1 Corresponding defective OLED display surface feature partition image data, e m2 Indicates the OLED display surface partition defect type data b' m2 Corresponding defective OLED display screen surface feature partition image data;

[0099] S62, using a KD tree nearest neighbor search algorithm to find the defective OLED display screen surface feature partition image data set E of the defective OLED display screen surface feature partition image data set E. m1 To e m2 The standard OLED display screen surface feature partition image coordinate data c in the standard OLED display screen surface feature partition image coordinate data set C m According to the comparison of the OLED display screen surface feature partition image number m, search for the defective OLED display screen surface feature partition image data e m1 To e m2 The corresponding standard OLED display screen surface feature partition image coordinate data, and the generated OLED display screen surface partition defect coordinate data set F = (f m1 ,,f m2 ), where f m1 Represents the defective OLED display surface feature partition image data e m1 Corresponding OLED display surface partition defect coordinate data, f m2 Represents the defective OLED display surface feature partition image data e m2 Corresponding OLED display surface partition defect coordinate data.

[0100] The standard OLED display surface feature partition image coordinate data and the standard OLED display surface feature image data are matched according to a plane rectangular coordinate system to generate standard OLED display surface feature image coordinate combination data; the OLED display surface partition defect type data and the OLED display surface partition defect coordinate data are obtained and mapped to the standard OLED display surface feature image coordinate combination data according to the OLED display surface feature partition image number and coordinate value matching and displayed through the display device as follows:

[0101] S71, matching the standard OLED display screen surface feature partition image coordinate data set C with the standard OLED display screen surface feature image data A′ according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data G;

[0102] S72, obtaining a data set B' of surface partition defect types of the OLED display screen and a data set F of surface partition defect coordinates of the OLED display screen;

[0103] S73. Map the OLED display surface partition defect type data in the OLED display surface partition defect type data set B' and the OLED display surface partition defect coordinate data in the OLED display surface partition defect coordinate data set F to the standard OLED display surface feature image coordinate combination data G according to the OLED display surface feature partition image number m and the coordinate value matching, and display the mapped standard OLED display surface feature image coordinate combination data G through a display device, which includes any one of a display screen, a screen, and a mobile terminal.

[0104] Through the display screen surface defect type position search unit, an intelligent search algorithm is used to compare the image number value based on the OLED display screen surface partition defect type parameter and the standard OLED display screen surface feature partition image coordinate parameter, and the OLED display screen surface partition defect coordinate parameter is accurately searched out to achieve accurate positioning of the OLED display screen surface defect position; the display screen surface feature image information and partition image coordinate processing unit and the display screen surface defect information feedback display unit cooperate with each other to match the standard OLED display screen surface feature partition image coordinate parameter with the standard OLED display screen surface feature image parameter reference plane rectangular coordinate system, and construct the standard OLED display screen surface feature image coordinate combination parameter, which provides data support for the OLED display screen surface defect type recognition result and defect position visualization display; the display screen surface defect information feedback display unit matches the OLED display screen surface partition defect type parameter and the OLED display screen surface partition defect coordinate parameter according to the image number and coordinate value, accurately maps them to the standard OLED display screen surface feature image coordinate combination parameter, and displays them through the display device feedback, so as to achieve visual and intuitive feedback of the OLED display screen surface defect type and defect position, thereby improving the intuitiveness and applicability of the OLED display screen surface defect detection result.

[0105] An automatic detection system for surface defects of an OLED display screen based on machine vision that realizes an automatic detection method for surface defects of an OLED display screen, the system comprising a display screen surface image data processing module, a display screen surface defect detection module, and a display screen surface defect information feedback module;

[0106] The display screen surface image data processing module includes a display screen surface feature image acquisition unit, a display screen surface feature image preprocessing unit, a display screen surface feature image partition processing unit, and a display screen surface feature image partition positioning unit;

[0107] A display screen surface feature image acquisition unit is used to acquire surface feature image data of an OLED display screen through a shooting lens; a display screen surface feature image preprocessing unit is used to perform image data preprocessing on the surface feature image data of an OLED display screen, and generate surface feature image data of a standard OLED display screen; a display screen surface feature image partition processing unit is used to perform image plane grid division processing on the surface feature image data of a standard OLED display screen, and construct surface feature partition image data of a standard OLED display screen; a display screen surface feature image partition positioning unit is used to establish a plane rectangular coordinate system according to the surface feature image data of a standard OLED display screen, and measure the plane coordinates of the surface feature partition image data of a standard OLED display screen in the plane rectangular coordinate system, and construct the coordinate data of the surface feature partition image of a standard OLED display screen;

[0108] The display screen surface defect detection module includes a display screen surface defect feature image storage unit, a display screen surface defect type recognition unit, and a display screen surface defect type recognition result output unit;

[0109] A display screen surface defect feature image storage unit is used to store the OLED display screen surface defect feature image data; a display screen surface defect type recognition unit uses a machine vision data recognition algorithm to match the standard OLED display screen surface feature partition image data with the OLED display screen surface defect feature image data, and analyzes and constructs the OLED display screen surface partition defect type data; a display screen surface defect type recognition result output unit is used to output the OLED display screen surface partition defect type data;

[0110] The display screen surface defect information feedback module includes a display screen surface defect type position search unit, a display screen surface feature image information and partition image coordinate processing unit, a display screen surface defect information acquisition unit, and a display screen surface defect information feedback display unit;

[0111] The display screen surface defect type position search unit compares the OLED display screen surface feature partition image number value based on the OLED display screen surface partition defect type data and the standard OLED display screen surface feature partition image coordinate data, searches out the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface partition defect type data, and identifies and generates the OLED display screen surface partition defect coordinate data; the display screen surface feature image information and partition image coordinate processing unit matches the standard OLED display screen surface feature partition image coordinate data with the standard OLED display screen surface feature image data according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data; the display screen surface defect information acquisition unit is used to acquire the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data; the display screen surface defect information feedback display unit maps the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data according to the OLED display screen surface feature partition image number and coordinate value matching to the standard OLED display screen surface feature image coordinate combination data, and feedbacks and displays them through a display device.

[0112] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The automatic detection method of surface defects of OLED display screen is characterized by: The method comprises the following steps: S1, collecting surface feature image data of the OLED display screen, wherein S1 comprises the following steps: S11, collecting a characteristic image of the surface of the OLED display screen online through a shooting lens and generating OLED display screen surface characteristic image data A, wherein the OLED display screen surface characteristic image data represents characteristic image data of the complete OLED display screen surface; S2, performing image data preprocessing on the surface feature image data of the OLED display screen, and generating standard surface feature image data of the OLED display screen, wherein S2 comprises the following steps: S21, using an adaptive filtering method to perform image data noise reduction preprocessing on the surface feature image data A of the OLED display screen, and generate standard surface feature image data A' of the OLED display screen; S3, performing image plane grid division processing on the surface feature image data of the standard OLED display screen to construct the surface feature partition image data of the standard OLED display screen, wherein S3 comprises the following steps: S31, the surface feature image data A' of the standard OLED display screen is processed by using a square with a side length a to perform image plane grid division processing, and a standard OLED display screen surface feature partition image data set B is constructed. m ,…,b μ ), m=1,2,3,…,μ; wherein bm represents the mth standard OLED display surface feature partition image data generated by performing image plane grid division processing on the standard OLED display surface feature image data A', and μ represents the maximum value of the number of standard OLED display surface feature partition image numbers; S4, establishing a plane rectangular coordinate system according to the surface feature image data of the standard OLED display screen, measuring the plane coordinates of the surface feature partition image data of the standard OLED display screen in the plane rectangular coordinate system, and constructing the surface feature partition image coordinate data of the standard OLED display screen, wherein S4 comprises the following steps: S41, establishing a plane rectangular coordinate system according to the standard OLED display screen surface feature image data A', with the horizontal side of the right-angle side of the corresponding image as the horizontal coordinate axis and the vertical side as the vertical coordinate axis; S42, measuring the standard OLED display screen surface feature partition image data b1 to b2 in the standard OLED display screen surface feature partition image data set B. μ The plane coordinates of the center points of the corresponding partition images in the plane rectangular coordinate system are used to construct the standard OLED display surface feature partition image coordinate data set C = (c1,…,c m ,…,c μ ), where cm represents the standard OLED display screen surface feature partition image data bm corresponding to the standard OLED display screen surface feature partition image coordinate data in the plane rectangular coordinate system, c m =(x m ,y m ), where x m Represents the surface feature partition image data of a standard OLED display screen b m The horizontal coordinate in the rectangular coordinate system, y m Represents the surface feature partition image data of a standard OLED display screen b m The ordinate in a rectangular coordinate system; S5, when the standard OLED display screen surface feature partition image coordinate data is constructed, the standard OLED display screen surface feature partition image data and the OLED display screen surface defect feature image data are matched with the image feature of the standard OLED display screen surface feature partition image data by using a data recognition algorithm, and the OLED display screen surface feature partition defect type data is analyzed and constructed and outputted, wherein S5 comprises the following steps: S51, establish the OLED display surface defect feature image data set D = (d1, ..., d n ,…,d ν ), n=1,2,3,…,ν; where d n represents the OLED display screen surface defect characteristic image data corresponding to the nth OLED display screen surface defect type, and ν represents the maximum number of OLED display screen surface defect types; the OLED display screen surface defect types include any one or more of scratch defects, pixel defects, bright spot defects, dark spot defects, and color deviation defects; S52, when the standard OLED display screen surface feature partition image coordinate data set C is constructed, the standard OLED display screen surface feature partition image data set B is used to identify the standard OLED display screen surface feature partition image data set b m According to the standard OLED display screen surface feature partition image number m, the OLED display screen surface defect feature image data set D is sequentially compared with the OLED display screen surface defect feature image data d n Image feature matching is performed to analyze and construct the surface partition defect type data of the OLED display screen. The specific operation steps of the data recognition algorithm to analyze and construct the surface partition defect type data of the OLED display screen are as follows: S521, initializing parameters, updating the number of defect type identification crow populations, the maximum number of iterations, and the flight distance Π; S522, initializing the initial position and memory of the defect type recognition crow individual in the search space of the OLED display surface defect feature image data set D, N defect type recognition crow individuals are randomly distributed in a multidimensional search space, that is, N defect type recognition crow individuals are randomly distributed in the search space of the OLED display surface defect feature image data set D with a spatial dimension of ν; in the first iteration, it is assumed that the defect type recognition crow individual hides the food at the initial position; that is, in the first iteration, it is assumed that the defect type recognition crow individual hides the food at the initial position in the search space of the OLED display surface defect feature image data set D; S523, calculating the individual fitness value of each defect type recognition crow, that is, calculating the standard OLED display screen surface feature partition image data set B of the standard OLED display screen surface feature partition image data set b m According to the standard OLED display surface feature partition image number m, the OLED display surface defect feature image data set D is sorted and the OLED display surface defect feature image data d n The fitness value at the initial position after matching; S524, updating the position of the defect type identification crow individual in the search space of the OLED display surface defect feature image data set D; S525, determine the feasibility of the new position, and determine the feasibility of the new position of each defect type identification crow individual; if the new position of the defect type identification crow individual is feasible, the defect type identification crow individual will update its position, that is, search for the OLED display surface defect feature image data set D that is consistent with the standard OLED display surface feature partition image data b m Matching OLED display surface defect feature image data d n , the defect type identification crow individual is updated to the successfully matched OLED display surface defect feature image data d n position; otherwise, the defect type identifies that the crow individual stays at the current position and does not move to the new position; S526, evaluate the fitness value of the new position, calculate the fitness value of the new position of each defect type recognition crow individual, that is, calculate the standard OLED display surface feature partition image data b in the standard OLED display surface feature partition image data set B m According to the standard OLED display surface feature partition image number m, the OLED display surface defect feature image data set D is sorted and the OLED display surface defect feature image data d n The fitness value at the new position after matching; S527, updating memory. If the fitness value of the new position of the defect type recognition crow is greater than the fitness value of the initial position in the memory, the defect type recognition crow updates its memory through the new position, that is, searching for the standard OLED display surface feature partition image data set B that is consistent with the standard OLED display surface feature partition image data b. m OLED display surface defect feature image data d with the largest fitness value n ; S528: When the maximum number of iterations is met, output the standard OLED display screen surface feature partition image data b m OLED display surface defect feature image data d n After matching, the corresponding OLED display screen surface partition defect type data is established to establish the OLED display screen surface partition defect type data set B'=(b' m1 ,…,b' m2 ), 1≤m1≤m≤m2≤μ, where b' m1 Represents the surface feature partition image data of a standard OLED display screen b m1 Corresponding OLED display surface partition defect type data; b' m2 Represents the surface feature partition image data of a standard OLED display screen b m2 Corresponding OLED display screen surface partition defect type data; S53, outputting the OLED display surface partition defect type data set B'; S6, comparing the numerical values ​​of the OLED display screen surface feature partition image numbers according to the OLED display screen surface feature partition defect type data and the standard OLED display screen surface feature partition image coordinate data, searching for the standard OLED display screen surface feature partition image coordinate data corresponding to the OLED display screen surface feature partition defect type data and identifying and generating the OLED display screen surface feature partition defect coordinate data; S7, matching the standard OLED display screen surface feature partition image coordinate data with the standard OLED display screen surface feature image data according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data; obtaining the OLED display screen surface partition defect type data and the OLED display screen surface partition defect coordinate data according to the OLED display screen surface feature partition image number and coordinate value matching, mapping them to the standard OLED display screen surface feature image coordinate combination data, and displaying them through a display device.

2. The method for automatically detecting surface defects of an OLED display screen according to claim 1, characterized in that: The S6 comprises the following steps: S61, using a KD tree nearest neighbor search algorithm to search the OLED display screen surface partition defect type data set B' for the OLED display screen surface partition defect type data set B'. m1 to b' m2 The standard OLED display screen surface feature partition image data set B has the same characteristics as the standard OLED display screen surface feature partition image data b m Perform numerical comparison of the image number m of the surface feature partition of the OLED display screen, and search for the surface partition defect type data b' of the OLED display screen m1 to b' m2 The corresponding standard OLED display screen surface feature partition image data is used to identify and generate the defective OLED display screen surface feature partition image data set E = (e m1 ,…,e m2 ), where e m1 Indicates the OLED display surface partition defect type data b' m1 Corresponding defective OLED display surface feature partition image data, e m2 Indicates the OLED display surface partition defect type data b' m2 Corresponding defective OLED display screen surface feature partition image data; S62, using a KD tree nearest neighbor search algorithm to search the defective OLED display screen surface feature partition image data set E for the defective OLED display screen surface feature partition image data set E. m1 To e m2 The standard OLED display screen surface feature partition image coordinate data c in the standard OLED display screen surface feature partition image coordinate data set C m According to the comparison of the OLED display screen surface feature partition image number m, search for the defective OLED display screen surface feature partition image data e m1 To e m2 The corresponding standard OLED display screen surface feature partition image coordinate data, and the generated OLED display screen surface partition defect coordinate data set F = (f m1 ,…,f m2 ), where f m1 Represents the defective OLED display surface feature partition image data e m1 Corresponding OLED display surface partition defect coordinate data, f m2 Represents the defective OLED display surface feature partition image data e m2 Corresponding OLED display surface partition defect coordinate data.

3. The method for automatically detecting surface defects of an OLED display screen according to claim 2, characterized in that: The S7 comprises the following steps: S71, matching the standard OLED display screen surface feature partition image coordinate data set C with the standard OLED display screen surface feature image data A' according to a plane rectangular coordinate system to generate standard OLED display screen surface feature image coordinate combination data G; S72, obtaining the OLED display screen surface partition defect type data set B' and the OLED display screen surface partition defect coordinate data set F; S73, mapping the OLED display screen surface partition defect type data in the OLED display screen surface partition defect type data set B' and the OLED display screen surface partition defect coordinate data in the OLED display screen surface partition defect coordinate data set F to the standard OLED display screen surface feature image coordinate combination data G according to the OLED display screen surface feature partition image number m and coordinate value matching, and displaying the mapped standard OLED display screen surface feature image coordinate combination data G through a display device, wherein the display device includes any one of a display screen, a screen, and a mobile terminal.

4. An automatic detection system for surface defects of an OLED display screen based on machine vision that implements the automatic detection method for surface defects of an OLED display screen as claimed in any one of claims 1 to 3, characterized in that: The system comprises a display screen surface image data processing module, a display screen surface defect detection module, and a display screen surface defect information feedback module.

Images