Defect detection method and device based on dynamic double-threshold template and storage medium

Abstract

The invention discloses a defect detection method and device based on a dynamic dual-threshold template, and a storage medium, and aims to improve the defect detection precision of an image shot by a display screen. Collecting a plurality of batches of defect-free sample images of the target model display screen; counting gray extreme value boundary data of each pixel point on the plurality of batches of defect-free sample images; and generating a lower limit standard template and an upper limit standard template according to the gray extreme value boundary data. And obtaining a regional difference amplification coefficient and a compensation offset. And generating a dynamic detection threshold template for the lower limit standard template and the upper limit standard template through the regional difference amplification coefficient and the compensation offset. And obtaining a to-be-tested screen body image of the target model display screen. And carrying out pixel-level difference on the to-be-detected screen body image and the dynamic detection threshold template to generate a defect mask image. And performing defect detection on the display screen through the defect mask image to generate a first defect detection result.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of display screen detection, and in particular to a defect detection method and device based on a dynamic double-threshold template and a storage medium. BACKGROUND

[0002] In the automatic optical detection method of surface defects of a display panel (such as an OLED, a Micro-LED, and an LCD), the display panel body gray scale may be unevenly distributed. Therefore, when detecting surface defects of the display panel, the prior art usually collects a plurality of captured images in batches to obtain a plurality of gray scale values of each pixel point, sorts the plurality of gray scale values of the same pixel point from small to large, takes an average value and a median value as a gray scale value of the point on a standard template Golden image, ignores inherent gray scale differences (such as edge attenuation and non-uniform backlight) of different display areas of the display panel, directly performs difference between a to-be-detected image and the standard template Golden image, and then uses a full threshold to detect. The global fixed difference threshold leads to over-detection (such as misjudgment of a bright spot as a defect) in a high dynamic range or missed detection (such as a dark area fine scratch) in a low contrast area.

[0003] Nowadays, new display screens are continuously updated and iterated, and the display screen structure and pixel points are increasingly precise. The traditional gray scale uneven distribution processing method is obviously insufficient, especially for new display screens such as folding screens, spliced screens, display screens containing microcircuit modules, and quantum dot electroluminescent display screens QLED, etc. The traditional method cannot adapt to the gray scale baseline fluctuation of different specific display areas of the new display panel, such as edge brightness attenuation and non-uniform center area. At the same time, the median value and the average value obtained by sorting the gray scale will eliminate the area characteristics, leading to misjudgment in the high dynamic range area, which is easy to cause defect over-detection, missed detection, etc., and reduces the defect detection precision of the display screen captured image. SUMMARY

[0004] The present application discloses a defect detection method and device based on a dynamic double-threshold template and a storage medium, for improving the defect detection precision of a display screen captured image.

[0005] In a first aspect, embodiments of the present application provide a defect detection method based on a dynamic double-threshold template, comprising: Collect several batches of defect-free sample images of the target model display screen; count the grayscale extreme value boundary data of each pixel point on the several batches of defect-free sample images; generate a lower limit standard template and an upper limit standard template based on the grayscale extreme value boundary data; obtain the regional difference amplification coefficient and the compensation offset; generate a dynamic detection threshold template for the lower limit standard template and the upper limit standard template through the regional difference amplification coefficient and the compensation offset; obtain the screen body image to be tested of the target model display screen; perform pixel-level differentiation between the screen body image to be tested and the dynamic detection threshold template to generate a defect mask image; perform defect detection on the display screen through the defect mask image to generate a first defect detection result.

[0006] Optionally, the regional difference amplification coefficient includes an upper limit regional difference amplification coefficient and a lower limit regional difference amplification coefficient, the compensation offset includes an upper limit compensation offset and a lower limit compensation offset, and the dynamic detection threshold template includes a bright defect dynamic detection threshold template and a dark defect dynamic detection threshold template; the step of generating a dynamic detection threshold template from the lower limit standard template and the upper limit standard template through the regional difference amplification coefficient and the compensation offset includes: generating a bright defect dynamic detection threshold template through the upper limit regional difference amplification coefficient, the upper limit compensation offset, the lower limit standard template and the upper limit standard template; generating a dark defect dynamic detection threshold template through the lower limit regional difference amplification coefficient, the lower limit compensation offset, the lower limit standard template and the upper limit standard template.

[0007] Optionally, the target model display screen is a quantum dot electroluminescent display screen; the step of obtaining the regional difference magnification coefficient and the compensation offset includes: determining the radial thickness distribution of the quantum dot electroluminescent display screen corresponding to each defect-free sample image; dividing each defect-free sample image into a high-thickness area and a low-thickness area according to the radial thickness distribution, the high-thickness area being an area with a radial thickness higher than the upper limit of a reference thickness range, and the low-thickness area being an area with a radial thickness higher than the upper limit of a reference thickness range; calculating the upper-limit regional difference magnification coefficient and the upper-limit compensation offset based on the grayscale data, grayscale gradient data and thin film reflectivity data of the high-thickness area; calculating the lower-limit regional difference magnification coefficient and the lower-limit compensation offset based on the grayscale data, grayscale gradient data and thin film reflectivity data of the low-thickness area; and averaging the upper-limit regional difference magnification coefficient, upper-limit compensation offset, lower-limit regional difference magnification coefficient and lower-limit compensation offset corresponding to each defect-free sample image.

[0008] Optionally, the defect mask image includes a bright defect mask image and a dark defect mask image; performing pixel-level differentiation between the screen image to be tested and the dynamic detection threshold template to generate the defect mask image includes: performing pixel-level differentiation between the screen image to be tested and the bright defect dynamic detection threshold template to generate a bright defect mask image; performing pixel-level differentiation between the screen image to be tested and the dark defect dynamic detection threshold template to generate a dark defect mask image.

[0009] Optionally, after the step of performing defect detection on the display screen through the defect mask image and generating a first defect detection result, the defect detection method further includes: performing connected domain segmentation on the defect mask image to determine the abnormal area; determining the background area based on the abnormal area; performing grayscale difference analysis between the abnormal area and the background area, and performing defect threshold control processing to generate a second detection result; and performing defect integration analysis on the first defect detection result and the second defect detection result.

[0010] Optionally, the step of determining the background area based on the abnormal area includes: determining the minimum inscribed circle radius of the abnormal area, dilating the abnormal area with the radius of the inscribed circle to generate the background area; when the background area contains other abnormal areas, removing the part of the background area that intersects with other abnormal areas.

[0011] Optionally, the grayscale difference between the abnormal area and the background area is analyzed, and the threshold control processing of the defect is performed. The steps of generating the second detection result include: calculating the first grayscale mean and the second grayscale mean of the abnormal area and the background area; calculating the contrast attribute difference value of the defect based on the first grayscale mean and the second grayscale mean; performing threshold control processing on the contrast attribute difference value to generate the second detection result.

[0012] Optionally, the step of generating a lower limit standard template and an upper limit standard template based on the grayscale extreme boundary data includes: determining the boundary grayscale value of each pixel point based on the grayscale extreme boundary data and the boundary ratio; and generating a lower limit standard template and an upper limit standard template based on the boundary grayscale value of each pixel point.

[0013] In the second aspect, an embodiment of the present application provides a defect detection device based on a dynamic dual-threshold template, including: an acquisition unit for acquiring several batches of defect-free sample images of a target model display screen; a statistical unit for counting the grayscale extreme boundary data of each pixel point on several batches of defect-free sample images; a first generation unit for generating a lower limit standard template and an upper limit standard template based on the grayscale extreme boundary data; a first acquisition unit for acquiring a regional difference amplification coefficient and a compensation offset; a second generation unit for generating a dynamic detection threshold template for the lower limit standard template and the upper limit standard template through the regional difference amplification coefficient and the compensation offset; a second acquisition unit for acquiring a screen image to be tested of a target model display screen; a third generation unit for performing pixel-level differentiation between the screen image to be tested and the dynamic detection threshold template to generate a defect mask image; and a fourth generation unit for performing defect detection on the display screen through the defect mask image to generate a first defect detection result.

[0014] Optionally, the regional difference amplification coefficient includes an upper limit regional difference amplification coefficient and a lower limit regional difference amplification coefficient, the compensation offset includes an upper limit compensation offset and a lower limit compensation offset, and the dynamic detection threshold template includes a bright defect dynamic detection threshold template and a dark defect dynamic detection threshold template; the second generation unit includes: generating a bright defect dynamic detection threshold template through the upper limit regional difference amplification coefficient, the upper limit compensation offset, the lower limit standard template and the upper limit standard template; generating a dark defect dynamic detection threshold template through the lower limit regional difference amplification coefficient, the lower limit compensation offset, the lower limit standard template and the upper limit standard template.

[0015] Optionally, the target model display screen is a quantum dot electroluminescent display screen; the first acquisition unit specifically includes: determining the radial thickness distribution of the quantum dot electroluminescent display screen corresponding to each defect-free sample image; dividing each defect-free sample image into a high-thickness area and a low-thickness area according to the radial thickness distribution, the high-thickness area is an area with a radial thickness higher than the upper limit of a reference thickness range, and the low-thickness area is an area with a radial thickness higher than the upper limit of a reference thickness range; calculating the upper-limit area difference magnification coefficient and the upper-limit compensation offset according to the grayscale data, grayscale gradient data and thin film reflectivity data of the high-thickness area; calculating the lower-limit area difference magnification coefficient and the lower-limit compensation offset according to the grayscale data, grayscale gradient data and thin film reflectivity data of the low-thickness area; and averaging the upper-limit area difference magnification coefficient, upper-limit compensation offset, lower-limit area difference magnification coefficient and lower-limit compensation offset corresponding to each defect-free sample image.

[0016] Optionally, the defect mask image includes a bright defect mask image and a dark defect mask image; the third generation unit includes: performing pixel-level differentiation between the screen image to be tested and the bright defect dynamic detection threshold template to generate a bright defect mask image; performing pixel-level differentiation between the screen image to be tested and the dark defect dynamic detection threshold template to generate a dark defect mask image.

[0017] Optionally, after the fourth generation unit, the defect detection device also includes: a first determination unit, used to perform connected domain segmentation on the defect mask image to determine the abnormal area; a second determination unit, used to determine the background area based on the abnormal area; a fifth generation unit, used to analyze the grayscale difference between the abnormal area and the background area, and perform threshold control processing of the defect to generate a second detection result; an analysis unit, used to perform defect integration analysis on the first defect detection result and the second detection result.

[0018] Optionally, the second determining unit includes: determining the minimum inscribed circle radius of the abnormal area, dilating the abnormal area with the radius of the inscribed circle to generate a background area; when the background area contains other abnormal areas, removing the part of the background area that intersects with other abnormal areas.

[0019] Optionally, the fifth generation unit includes: calculating the first grayscale mean and the second grayscale mean of the abnormal area and the background area; calculating the contrast attribute difference value of the defect based on the first grayscale mean and the second grayscale mean; performing threshold control processing on the contrast attribute difference value to generate a second detection result.

[0020] Optionally, the first generating unit includes: determining a boundary grayscale value of each pixel point according to grayscale extreme value boundary data and a boundary ratio; and generating a lower limit standard template and an upper limit standard template according to the boundary grayscale value of each pixel point.

[0021] In a third aspect, an embodiment of the present application provides a defect detection device based on a dynamic dual-threshold template, comprising: processor, memory, input and output units, and buses; The processor is connected to the memory, input and output units, and the bus; The memory stores a program, and the processor calls the program to execute the first aspect and any optional defect detection method of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium on which a program is stored. When the program is executed on a computer, the program executes the first aspect and any optional defect detection method of the first aspect.

[0023] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: In the present application, several batches of defect-free sample images of the target model display screen are first collected. The grayscale extreme value boundary data of each pixel point on the several batches of defect-free sample images are counted. The lower limit standard template and the upper limit standard template are generated based on the grayscale extreme value boundary data. The regional difference magnification coefficient and the compensation offset are obtained. The dynamic detection threshold template is generated for the lower limit standard template and the upper limit standard template using the regional difference magnification coefficient and the compensation offset. The screen body image to be tested of the target model display screen is obtained. The screen body image to be tested is differentiated with the dynamic detection threshold template at the pixel level to generate a defect mask image. The display screen is subjected to defect detection using the defect mask image to generate a first defect detection result.

[0024] By statistically analyzing the grayscale extreme boundary data of each pixel in several batches of defect-free sample images, an allowable grayscale fluctuation range is established, and nonlinear dynamic expansion is performed at the same time. The corresponding lower limit standard template and upper limit standard template are established based on the corresponding grayscale value difference characteristics. Then, the corresponding grayscale value difference at the same pixel is determined based on the lower limit standard template and the upper limit standard template. The final dynamic detection threshold template is obtained by using the regional magnification coefficient and the compensation value offset. Finally, during detection, the surface abnormal area is obtained by differentiating the image of the side screen and the dynamic detection threshold template to generate a defect mask image. At the same time, the contrast parameter value of the defect abnormal area in the defect mask image generated after the difference is controlled by comparing it with the background area. This method can well cope with the precise structure of the new display screen, detect the surface defect abnormal area, reduce the situation of smearing regional characteristics, reduce the misjudgment of high dynamic range areas, avoid over-inspection and missed detection, and improve the defect detection accuracy of display screen images. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0026] Figure 1 This is a schematic diagram of a first embodiment of a defect detection method based on a dynamic dual-threshold template of the present application; Figure 2 A schematic diagram of a first embodiment of a method for adjusting a light source of an imaging system according to the present application; Figure 3 A schematic diagram of a first embodiment of a method for generating regional difference magnification coefficients and compensation offsets according to the present application; Figure 4 A schematic diagram of a first embodiment of a method for generating a defect mask image according to the present application; Figure 5 A schematic diagram of a first embodiment of a defect detection method of the present application; Figure 6 A schematic diagram of a first embodiment of a method for determining a background area of ​​the present application; Figure 7 A schematic diagram of a first embodiment of a method for generating a second detection result according to the present application; Figure 8 A schematic diagram of a first embodiment of a method for generating a lower limit standard template and an upper limit standard template according to the present application; Figure 9This is a schematic diagram of a first embodiment of a defect detection device based on a dynamic dual-threshold template of the present application; Figure 10 This is a schematic diagram of a second embodiment of a defect detection device based on a dynamic dual-threshold template of the present application. DETAILED DESCRIPTION

[0027] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0028] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0029] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0030] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0031] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0032] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0033] Different areas of the screen have different grayscale values. Therefore, we cannot directly use the grayscale values ​​of the same position in batch images to sort them. We need to obtain the median and average values ​​as the pixel value of the point on the standard Golden image and then perform differential detection between the original image and the standard Golden image. The grayscale value differences corresponding to different areas of the screen must be considered.

[0034] Nowadays, new display screens are constantly updated and iterated, and the display screen structure and pixel points are becoming more and more precise. Traditional methods for processing uneven grayscale distribution are obviously insufficient, especially when facing new display screens such as folding screens, spliced ​​screens, displays containing microcircuit modules, and quantum dot electroluminescent displays (QLED). Traditional methods cannot adapt to the grayscale baseline fluctuations of new display panels for different specific display areas, such as edge brightness attenuation and uneven center area. At the same time, taking the median and mean of grayscale sorting will erase regional characteristics, resulting in misjudgment of high dynamic range areas, which can easily lead to over-inspection or missed detection of defects, etc., reducing the defect detection accuracy of display screen images.

[0035] Based on this, the present application discloses a defect detection method, device and storage medium based on a dynamic dual-threshold template, which are used to improve the defect detection accuracy of images captured by display screens.

[0036] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0037] The method of the present application can be applied to a server, device, terminal or other device with logic processing capability, and the present application does not limit this. For the convenience of description, the following description is based on the example of the execution subject being a terminal.

[0038] See also Figure 1The present application provides an embodiment of a defect detection method based on a dynamic dual-threshold template, comprising: 101. Collect several batches of defect-free sample images of the target model display screen.

[0039] In this embodiment, images of the target display screen are first captured, and defect-free images are used as samples. Specifically, before using the defect detection equipment to perform defect detection on a new display screen, a standard, defect-free new display screen is placed for image capture. The captured images are then subjected to defect detection and image correction to obtain defect-free sample images free of display screen defects, lens defects, and light source defects. A minimum of 50 batches of defect-free sample images must be collected.

[0040] 102. Count the grayscale extreme boundary data of each pixel on several batches of defect-free sample images.

[0041] 103. Generate a lower limit standard template and an upper limit standard template according to the grayscale extreme value boundary data.

[0042] Specifically, after acquiring N batches of rectified defect-free sample images, the terminal obtains the grayscale value of each pixel in each image. Specifically, each pixel obtains a corresponding initial grayscale set, each containing N pixel values. The terminal then calculates the grayscale extreme boundary data for each pixel in the N batches of defect-free sample images. Specifically, to prevent the influence of image noise and stains and ensure the authenticity of the created Golden standard template image, the terminal calculates the grayscale extreme boundary data for each pixel position (x, y) by counting the grayscale extreme values ​​of each pixel in the N defect-free sample images. Based on this grayscale extreme boundary data, the terminal then selects boundary grayscale values ​​suitable for the new display screen. There are two boundary grayscale values: an upper boundary grayscale value and a lower boundary grayscale value. The upper and lower boundary grayscale values ​​are used to generate a lower and upper standard template. The effective area sizes of the lower and upper standard templates are the same as the effective area of ​​the display screen in the defect-free sample images.

[0043] 104. Obtain a regional difference amplification coefficient and a compensation offset.

[0044] 105. Generate a dynamic detection threshold template for the lower limit standard template and the upper limit standard template through the regional difference amplification coefficient and the compensation offset.

[0045] In this embodiment, the terminal needs to first obtain the regional difference amplification coefficient and the compensation offset, and then use the gold_max and gold_min differences at different positions in the lower limit standard template and the upper limit standard template to establish the allowable grayscale fluctuation range through the regional difference amplification coefficient and the compensation offset, and perform nonlinear dynamic expansion at the same time to regenerate the dynamic detection threshold template.

[0046] The regional difference amplification coefficient is calculated using the display characteristics of the specific area of ​​the new display screen. Compared with conventional display screens, the new display screen has different functional areas. These functional areas often involve more complex usage conditions during use, such as the folding area of ​​the folding screen, the splicing edge of the splicing screen, and the light conversion film structure of the quantum dot electroluminescent display screen, especially the light conversion film structure of the quantum dot electroluminescent display screen. These specific areas are called specific areas. Such specific areas usually require more complex defect detection projects to effectively detect new defects in these special structures during use, which is impossible with conventional detection methods. In this embodiment, the terminal calculates the regional difference amplification coefficient and the compensation offset based on the display characteristics of the new display screen in the specific area, and then makes targeted adjustments to the lower limit standard template and the upper limit standard template based on the regional difference amplification coefficient and the compensation offset to generate a dynamic detection threshold template. The dynamic detection threshold template can improve the accuracy of subsequent detection of defects in the specific area of ​​the new display screen. The specific calculation method will be described in detail later.

[0047] 106. Obtain a screen image of the target model display screen to be tested.

[0048] 107. Perform pixel-level differentiation between the screen image to be tested and the dynamic detection threshold template to generate a defect mask image.

[0049] The terminal acquires an image of the target display model to be tested. It then performs pixel-level differentiation between the image and the dynamic detection threshold template to generate a defect mask image, effectively detecting both screen and display defects on new display screens. This defect mask image improves detection and contrast in defective areas.

[0050] 108. Perform defect detection on the display screen using the defect mask image to generate a first defect detection result.

[0051] Finally, the terminal may perform defect detection on the display screen on the defect mask image by conventional means to generate a first defect detection result.

[0052] In this embodiment, several batches of defect-free sample images of the target model display screen are first collected. Grayscale extreme value boundary data for each pixel on the several batches of defect-free sample images are counted. A lower limit standard template and an upper limit standard template are generated based on the grayscale extreme value boundary data. A regional difference amplification factor and a compensation offset are obtained. A dynamic detection threshold template is generated for the lower limit standard template and the upper limit standard template using the regional difference amplification factor and the compensation offset. An image of the screen to be tested of the target model display screen is obtained. A pixel-level difference is performed between the screen to be tested and the dynamic detection threshold template to generate a defect mask image. Defect detection of the display screen is performed using the defect mask image to generate a first defect detection result.

[0053] By statistically analyzing the grayscale extreme boundary data of each pixel in several batches of defect-free sample images, an allowable grayscale fluctuation range is established, and nonlinear dynamic expansion is performed at the same time. The corresponding lower limit standard template and upper limit standard template are established based on the corresponding grayscale value difference characteristics. Then, the corresponding grayscale value difference at the same pixel is determined based on the lower limit standard template and the upper limit standard template. The final dynamic detection threshold template is obtained by using the regional magnification coefficient and the compensation value offset. Finally, during detection, the surface abnormal area is obtained by differentiating the image of the side screen and the dynamic detection threshold template to generate a defect mask image. At the same time, the contrast parameter value of the defect abnormal area in the defect mask image generated after the difference is controlled by comparing it with the background area. This method can well cope with the precise structure of the new display screen, detect the surface defect abnormal area, reduce the situation of smearing regional characteristics, reduce the misjudgment of high dynamic range areas, avoid over-inspection and missed detection, and improve the defect detection accuracy of display screen images.

[0054] See also Figure 2 The present application provides an embodiment of a method for generating a dynamic detection threshold template, comprising: 201. Generate a bright defect dynamic detection threshold template through the upper limit area difference magnification coefficient, the upper limit compensation offset, the lower limit standard template and the upper limit standard template.

[0055] 202. Generate a dark defect dynamic detection threshold template using the lower limit area difference magnification coefficient, the lower limit compensation offset, the lower limit standard template, and the upper limit standard template.

[0056] In this embodiment, the terminal generates a bright defect dynamic detection threshold template using the upper limit area difference magnification coefficient, the upper limit compensation offset, the lower limit standard template, and the upper limit standard template. The formula is as follows: NewMaxGolden=MaxGolden+HighFactor(MaxGolden-MinGolden)+HighDelat Next, the terminal generates a dark defect dynamic detection threshold template using the lower limit area difference amplification coefficient, the lower limit compensation offset, the lower limit standard template, and the upper limit standard template. The formula is as follows: NewMinGolden=MinGolden-LowFactor*(MaxGolden-MinGolden)-LowDelat MaxGolden is the upper limit standard template, which can also be written as MaxGolden(x, y), where (x, y) represents the corresponding pixel position. MinGolden is the lower limit standard template. LowFactor and HighFactor are the lower limit and upper limit regional difference amplification factors for each pixel area, respectively. LowDelat and HighDelat are the lower limit and upper limit compensation offsets for the corresponding grayscale value, respectively.

[0057] See also Figure 3 The target display model is a quantum dot electroluminescent display. This application provides an embodiment of a method for generating a regional difference amplification coefficient and a compensation offset, including: 301. Determine the radial thickness distribution of the quantum dot electroluminescent display corresponding to each defect-free sample image; This embodiment demonstrates the fabrication method for a quantum dot light-emitting diode (QLED) display. Essentially, it uses pre-prepared quantum dot material as a light conversion film, which is placed in a specific manner between the blue LED backlight and the front liquid crystal layer. The blue backlight excites the quantum dot film, emitting pure red and green light. This light, combined with the remaining blue light, creates high-quality white light, which is then controlled by liquid crystal pixels. QLEDs do not emit light themselves.

[0058] In this embodiment, the quantum dot electroluminescent display (QLED) uses quantum dot material in each pixel to actively emit light when directly driven by an electric current. This is similar to the operating principle of traditional OLEDs, replacing the traditional light-emitting layer with quantum dots. A standard QLED uses a "sandwich" structure. Similar to OLED, the QLED mainly includes an anode, a hole transport layer, a quantum dot light-emitting layer, an electron transport layer, and a cathode.

[0059] The quantum dot light-emitting layer, the core layer of a quantum dot electroluminescent display (QLED), is composed of red, green, and blue quantum dot materials. Under the influence of an electric field, electrons and holes recombine in this layer, emitting light of a specific color. The fabrication of QLEDs combines the device physics of OLEDs with the materials science of quantum dots. The core challenge lies in precisely and non-destructively integrating quantum dots, a nanomaterial that is extremely sensitive to water and oxygen and easily damaged, into a multilayer thin-film device.

[0060] Quantum dot ink is commonly used in the production of quantum dot electroluminescent displays (QLEDs). First, the synthesized quantum dot particles are separated from the original solvent and then dispersed into a solvent with suitable physical properties (such as boiling point, surface tension, and viscosity) to form "quantum dot ink." The film is then formed using inkjet printing. Specifically, on a substrate with pre-fabricated TFT circuits and electrodes (anodes), pixel pits are created using a photolithography process. High-precision inkjet printing equipment is then used to precisely spray red, green, and blue quantum dot inks into the corresponding pixel pits. Finally, the solvent is evaporated by annealing (heating), leaving a uniform, flat quantum dot film. However, in existing technologies, during the drying process, the edges of the ink evaporate faster than the center. This causes the quantum dot particles to aggregate toward the edges due to tension, forming a ring-shaped, uneven film that is thin in the middle and thick at the edges. Even by adjusting the solvent and evaporation rate, a small amount of ring-shaped uneven film still exists.

[0061] This type of annular non-uniform film will affect the uniformity of luminescence, causing bright defects and dark defects to be unable to be captured by the upper and lower limit standard templates when defect detection is performed on the image to be inspected of the quantum dot electroluminescent display screen using the upper and lower limit standard templates, thereby reducing the defect detection accuracy of the image to be inspected of the quantum dot electroluminescent display screen.

[0062] Therefore, this embodiment proposes a method for generating a dynamic detection threshold template, using a defect-free sample image of a quantum dot electroluminescent display to adjust the upper limit standard template and the lower limit standard template. Specifically, based on the radial thickness distribution in the defect-free sample image of the quantum dot electroluminescent display, the reflective areas of the display screen affected by different thicknesses are separated by radial thickness. Then, based on the reflectivity, grayscale gradient and other data of the reflective area of ​​the display screen, the difference amplification factor and compensation offset between the upper limit standard template and the lower limit standard template are calculated.

[0063] Specifically, the terminal first determines the radial thickness distribution of the quantum dot electroluminescent display corresponding to each defect-free sample image. The film thickness distribution can be inverted by analyzing the interference effect of light reflected at the film interface or changes in the reflection spectrum or polarization state. Alternatively, an X-ray reflectometer or white light interferometer can be used, but this is not limited here.

[0064] 302. Divide each defect-free sample image into a high-thickness region and a low-thickness region according to the radial thickness distribution, wherein the high-thickness region is a region where the radial thickness is higher than the upper limit of the reference thickness range, and the low-thickness region is a region where the radial thickness is higher than the upper limit of the reference thickness range; Next, the terminal divides each defect-free sample image into high-thickness regions and low-thickness regions based on the radial thickness distribution. High-thickness regions are regions where the radial thickness exceeds the upper limit of the reference thickness range, while low-thickness regions are regions where the radial thickness exceeds the upper limit of the reference thickness range. During the fabrication of quantum dot electroluminescent displays, the optimal thickness varies for displays of different sizes. Typically, a reference thickness range is generated centered around the optimal thickness. Regions within this reference thickness range are considered transition regions. Adjusted upper and lower limit standard templates can be used for the transition region, or unadjusted upper and lower limit standard templates can be used. This is because defects in the transition region are less affected by annular non-uniform thin films. Radial thickness affects defect detection by affecting reflectivity, and the grayscale gradients of high-thickness regions (usually the outer ring) and low-thickness regions (usually the central circular or elliptical region) can also affect the appearance of defect features. Therefore, in this embodiment, a quantum dot electroluminescent display without a severely non-uniform thin film, i.e., a medium-uniformity quantum dot electroluminescent display, was used to obtain defect-free sample images corresponding to this type of display. Each defect-free sample image was then divided into high-thickness regions and low-thickness regions based on the radial thickness distribution.

[0065] 303. Calculate the upper limit region difference amplification coefficient and the upper limit compensation offset based on the grayscale data, grayscale gradient data, and film reflectivity data of the high thickness region; 304. Calculate the lower limit region difference amplification coefficient and the lower limit compensation offset based on the grayscale data, grayscale gradient data, and film reflectivity data of the low thickness region; The terminal calculates the upper limit area difference amplification factor and the upper limit compensation offset based on the grayscale data, grayscale gradient data, and film reflectivity data of the high thickness area, and calculates the lower limit area difference amplification factor and the lower limit compensation offset based on the grayscale data, grayscale gradient data, and film reflectivity data of the low thickness area. The upper limit area difference amplification factor is for bright defects, and the lower limit area difference amplification factor is for dark defects. The amplification factor of the upper limit standard template is generated using parameters such as the grayscale data of the high thickness area and the reflectivity in the uneven film, and the amplification factor of the lower limit standard template is generated using parameters such as the grayscale data of the low thickness area and the reflectivity in the uneven film. The formula is as follows:

[0066]

[0067] Among them, HighFactor is the upper limit area difference amplification coefficient, LowFactor is the lower limit area difference amplification coefficient, is the average reflectivity of the film in the reference thickness range (i.e., transition region), is the average reflectivity of the film in the high thickness region, is the average reflectivity of the film in the low thickness region, is the maximum pixel value in the high thickness area, is the minimum pixel value in the high thickness area, is the maximum pixel value in the low thickness area, is the minimum pixel value in the low thickness area. It should be noted that 、 、 and Before obtaining, it is necessary to remove abnormal extreme values ​​and then obtain the maximum and minimum values ​​from the remaining pixel values. is the weighted reference gradient in the transition region, is the weighted grayscale gradient of the high thickness area, is the weighted grayscale gradient of the low thickness area, is the grayscale mean of the high thickness area, is the grayscale mean of the low thickness area. is the contrast amplification term for high thickness areas, This is the contrast amplification term for low thickness areas, is the sharpness amplification term.

[0068]

[0069]

[0070]

[0071] in, and They are the horizontal gradient weighted value and the vertical gradient weighted value, respectively. The horizontal size and vertical size (or resolution) of the quantum dot electroluminescent display are usually used to generate the horizontal gradient weighted value and the vertical gradient weighted value. and are the horizontal and vertical gradients in the transition region, and are the horizontal and vertical gradients in the high thickness region, and These are the horizontal and vertical gradients in low-thickness regions, respectively. Using gradient and reflectivity adjustments in the contrast amplification option, supplemented by the sharpness amplification option, allows for better alignment of the lower and upper limit standard templates, generating dynamic detection threshold templates for bright and dark defects. In addition to overall adjustments to the lower and upper limit standard templates using the regional difference amplification factor, appropriate compensation is also required to reduce the difference between the extreme value boundaries of the lower and upper limit standard templates.

[0072] The calculation formulas for the upper limit compensation offset HighDelat and the lower limit compensation offset LowDelat are as follows:

[0073]

[0074] HighDelat and LowDelat are the upper limit compensation offset and lower limit compensation offset of the corresponding gray value respectively. and are the compensation strength coefficients of high thickness area and low thickness area respectively. This parameter is set as a constant based on human experience, aiming to adjust the boundary extreme value offset of the standard template. and They are the standard deviation of the high thickness area and the standard deviation of the low thickness area, respectively.

[0075]

[0076]

[0077] in, and are the number of pixels in the high thickness area and the number of pixels in the low thickness area, respectively. For high thickness areas, For low thickness areas, is the gray value of the i-th pixel in the high thickness area, is the grayscale value of the j-th pixel in the high-thickness area. It should be noted that the extremely high grayscale values ​​in the high-thickness area need to be screened out first, and the extremely low grayscale values ​​in the low-thickness area need to be screened out first. The upper and lower compensation offsets need to calculate the deviation of the boundary extreme values ​​through the degree of offset determined by the overall pixel value and the mean.

[0078] The upper limit compensation offset and the lower limit compensation offset calculated by the above formula are used to calculate the extreme value deviation based on each pixel point in the high thickness area and the low thickness area. The generated upper limit compensation offset and the lower limit compensation offset can eliminate the overall deviation of the lower limit standard template and the upper limit standard template.

[0079] 305. Averaging the upper limit region difference magnification coefficient, the upper limit compensation offset, the lower limit region difference magnification coefficient, and the lower limit compensation offset corresponding to each defect-free sample image.

[0080] The terminal can average the upper limit area difference magnification coefficient, upper limit compensation offset, lower limit area difference magnification coefficient and lower limit compensation offset corresponding to each defect-free sample image to cope with situations where high precision is not required. If higher precision is required, it is necessary to determine the upper limit area difference magnification coefficient, upper limit compensation offset, lower limit area difference magnification coefficient and lower limit compensation offset based on the radial thickness distribution of the screen image to be measured, so as to improve the accuracy of the subsequent generation of bright defect dynamic detection threshold template and dark defect dynamic detection threshold template.

[0081] See also Figure 4 The present application provides an embodiment of a method for generating a defect mask image, comprising: 401. Perform pixel-level differentiation between the screen image to be tested and the bright defect dynamic detection threshold template to generate a bright defect mask image.

[0082] 402. Perform pixel-level differentiation between the screen image to be tested and the dark defect dynamic detection threshold template to generate a dark defect mask image.

[0083] In this embodiment, the screen image to be tested is subjected to pixel-level differential detection. The terminal performs pixel-level differential between the screen image to be tested and the bright defect dynamic detection threshold template to generate a bright defect mask image. Each time the camera captures an image, it first acquires the screen area, then performs correction, and finally performs pixel-level differential between the corrected image Image(x, y) and the created NewMinGolden and NewMaxGolden. When bright defects are detected, the corresponding differential image (bright defect mask image) is HighDefectMask(x, y), and the formula is as follows: HighDefectMask(x,y)=I(x,y)-NewMaxGolden(x,y) When detecting dark defects, the image of the screen to be tested is differentiated from the dark defect dynamic detection threshold template at the pixel level to generate a dark defect mask image. The corresponding differential image (dark defect mask image) is LowDefectMask (x, y), and the formula is as follows: LowDefectMask(x,y)=NewMinGolden(x,y)-Image(x,y) See also Figure 5 , the present application provides an embodiment of a defect detection method, comprising: 501. Perform connected domain segmentation on the defect mask image to determine abnormal areas.

[0084] 502. Determine a background area based on the abnormal area.

[0085] 503. Analyze the grayscale difference between the abnormal area and the background area, perform threshold control processing on the defect, and generate a second detection result.

[0086] 504. Perform defect integration analysis on the first defect detection result and the second defect detection result.

[0087] To further verify whether the defect is genuine, the terminal performs connected domain segmentation on the defect mask image to identify the abnormal region. The background region is then determined based on the location and size of the abnormal region. The grayscale difference between the abnormal region and the background region is analyzed, and a threshold control process is performed to generate a second detection result. Finally, the first and second detection results are combined for defect analysis.

[0088] Specifically, the differential result images HighDefectMask (bright defect mask image) and LowDefectMask (dark defect mask image) will be segmented into connected domains, and then the ratio of the local average grayscale of the abnormal area to the average grayscale of the background will be calculated. The degree of difference between the defect and the background will be further determined based on the size of the value, and the defect will be controlled.

[0089] The specific method of determining the abnormal area will be described in the following embodiments. The specific method of performing grayscale difference analysis on the abnormal area and the background area and performing defect threshold control processing to generate the second detection result will be described in detail in the following embodiments.

[0090] See also Figure 6 , the present application provides an embodiment of a method for determining a background area, comprising: 601. Determine the minimum inscribed circle radius of the abnormal area, and dilate the abnormal area with the radius of the inscribed circle to generate a background area.

[0091] 602. When the background area includes other abnormal areas, remove the part of the background area that intersects with the other abnormal areas.

[0092] In this embodiment, the terminal uses the radius of the smallest inscribed circle of the defect area as a reference for the size of the corresponding background area. The defect area is then expanded by the radius of the inscribed circle. This means that the background area is the expanded defect area. If the expanded area contains other defect areas, any intersecting defects are removed.

[0093] See also Figure 7 , the present application provides an embodiment of a method for generating a second detection result, comprising: 701. Calculate a first grayscale mean value and a second grayscale mean value of the abnormal area and the background area.

[0094] 702. Calculate a contrast attribute difference value of the defect according to the first grayscale mean and the second grayscale mean.

[0095] 703. Perform threshold control processing on the contrast attribute difference value to generate a second detection result.

[0096] In this embodiment, the terminal first calculates the first grayscale mean and the second grayscale mean of the abnormal area and the background area, and respectively calculates the average grayscale values ​​of the defect area (abnormal area) and the expanded area (background area), which are recorded as Mean and BackMean. The calculation formula for the contrast attribute difference value (Contrast value) of the corresponding defect is as follows: Contrast=|Mean-BackMean| / (BackMean+Bias) Where Bias = 0.0001. A larger Contrast value indicates a more pronounced surface defect and a greater difference from the background. Areas with Contrast > T are retained as final defects, where T is the preset defect threshold for the panel type. Finally, the terminal applies a threshold to the contrast attribute difference to generate a second detection result.

[0097] See also Figure 8 The present application provides an embodiment of a method for generating a lower limit standard template and an upper limit standard template, comprising: 801. Determine the boundary grayscale value of each pixel point according to the grayscale extreme value boundary data and the boundary ratio.

[0098] 802. Generate a lower limit standard template and an upper limit standard template according to the boundary grayscale value of each pixel point.

[0099] In this embodiment, the terminal calculates the grayscale extreme value boundary data for each pixel position (x, y) by counting the grayscale extreme values ​​of each pixel in N defect-free sample images. The terminal can select the front and back 10% scores (boundary ratio) of the grayscale value of the pixel in the batch data as the grayscale extreme value boundary data, and then generate the lower limit standard template and the upper limit standard template according to the boundary grayscale value of each pixel. For example, if there are 50 batches (N=50), first sort the 50 grayscale values ​​of the same point (such as point (1, 1)) from small to large, take the 5th (5=50*10%) value gold_min as the grayscale value of the point (point (1, 1)) on the MinGolden image (lower limit standard template), and the 46th value gold_max as the grayscale value of the point (point (1, 1)) on the MaxGolden image (upper limit standard template), and so on. The corresponding proportions of gold_min and gold_max values ​​of all points are used to generate the lower limit standard template and the upper limit standard template, where this proportion is based on the prior image acquisition and detection of several new display screens, and the extreme values ​​of the pixel points in the specific area of ​​the image are counted (because specific areas are more likely to cause extreme points. If there is no specific area, the conventional center area and / or edge area is directly used), the occurrence ratio of extreme points is detected, and this ratio is used as the boundary ratio.

[0100] See also Figure 9 The present application provides an embodiment of a defect detection device based on a dynamic dual-threshold template, comprising: The acquisition unit 901 is configured to acquire several batches of defect-free sample images of a target model display screen.

[0101] The statistical unit 902 is used to count the grayscale extreme value boundary data of each pixel point on several batches of defect-free sample images.

[0102] The first generating unit 903 is configured to generate a lower limit standard template and an upper limit standard template according to the grayscale extreme value boundary data.

[0103] Optionally, the first generating unit 903 includes: The boundary grayscale value of each pixel is determined based on the grayscale extreme value boundary data and boundary ratio.

[0104] Generate a lower limit standard template and an upper limit standard template according to the boundary grayscale value of each pixel.

[0105] The first acquiring unit 904 is configured to acquire a regional difference magnification coefficient and a compensation offset.

[0106] Optionally, the target model display screen is a quantum dot electroluminescent display screen.

[0107] The first acquiring unit 904 specifically includes: Determine the radial thickness distribution of the quantum dot electroluminescent display corresponding to each defect-free sample image.

[0108] According to the radial thickness distribution, each defect-free sample image is divided into a high-thickness area and a low-thickness area. The high-thickness area is an area where the radial thickness is higher than the upper limit of the reference thickness range, and the low-thickness area is an area where the radial thickness is higher than the upper limit of the reference thickness range.

[0109] The upper limit area difference amplification coefficient and the upper limit compensation offset are calculated based on the grayscale data, grayscale gradient data and film reflectivity data of the high thickness area.

[0110] The lower limit area difference amplification coefficient and the lower limit compensation offset are calculated based on the grayscale data, grayscale gradient data and film reflectivity data of the low thickness area.

[0111] The upper limit region difference magnification coefficient, upper limit compensation offset, lower limit region difference magnification coefficient and lower limit compensation offset corresponding to each defect-free sample image are averaged.

[0112] The second generating unit 905 is configured to generate a dynamic detection threshold template for the lower limit standard template and the upper limit standard template by using the regional difference magnification coefficient and the compensation offset.

[0113] The second acquiring unit 906 is configured to acquire a screen image of a target display screen to be tested.

[0114] Optionally, the regional difference amplification coefficient includes an upper limit regional difference amplification coefficient and a lower limit regional difference amplification coefficient, the compensation offset includes an upper limit compensation offset and a lower limit compensation offset, and the dynamic detection threshold template includes a bright defect dynamic detection threshold template and a dark defect dynamic detection threshold template.

[0115] The second generating unit 906 includes: A bright defect dynamic detection threshold template is generated through the upper limit area difference amplification coefficient, the upper limit compensation offset, the lower limit standard template and the upper limit standard template.

[0116] A dark defect dynamic detection threshold template is generated through the lower limit area difference amplification coefficient, the lower limit compensation offset, the lower limit standard template and the upper limit standard template.

[0117] The third generating unit 907 is configured to perform pixel-level difference between the screen image to be tested and the dynamic detection threshold template to generate a defect mask image.

[0118] Optionally, the defect mask image includes a bright defect mask image and a dark defect mask image.

[0119] The third generating unit 907 includes: Perform pixel-level difference between the screen image to be tested and the bright defect dynamic detection threshold template to generate a bright defect mask image.

[0120] Perform pixel-level difference between the screen image to be tested and the dark defect dynamic detection threshold template to generate a dark defect mask image.

[0121] The fourth generating unit 908 is configured to perform defect detection on the display screen using the defect mask image to generate a first defect detection result.

[0122] The first determining unit 909 is configured to perform connected component segmentation on the defect mask image to determine abnormal areas.

[0123] The second determining unit 910 is configured to determine a background area according to the abnormal area.

[0124] Optionally, the second determining unit 910 includes: Determine the minimum inscribed circle radius of the abnormal area, dilate the abnormal area with the radius of the inscribed circle, and generate the background area.

[0125] When the background area contains other abnormal areas, the part of the background area that intersects with other abnormal areas is removed.

[0126] The fifth generating unit 911 is configured to analyze the grayscale difference between the abnormal area and the background area, perform threshold control processing on the defects, and generate a second detection result.

[0127] Optionally, the fifth generating unit 911 includes: Calculate the first grayscale mean and the second grayscale mean of the abnormal area and the background area.

[0128] A contrast attribute difference value of the defect is calculated based on the first grayscale mean and the second grayscale mean.

[0129] A threshold control process is performed on the contrast attribute difference value to generate a second detection result.

[0130] The analyzing unit 912 is configured to perform defect integration analysis on the first defect detection result and the second defect detection result.

[0131] See also Figure 10 , the present application provides a defect detection device based on a dynamic dual-threshold template, comprising: Processor 1001 , memory 1002 , input / output unit 1003 , and bus 1004 .

[0132] The processor 1001 is connected to the memory 1002 , the input / output unit 1003 , and the bus 1004 .

[0133] The memory 1002 stores a program, and the processor 1001 calls the program to execute the following Figure 1 、 Figure 2 and Figure 3 、 Figure 4 、 Figure 5 、 Figure 6 、 Figure 7 and Figure 8 Defect detection method in .

[0134] The present application provides a computer-readable storage medium, wherein a program is stored on the computer-readable storage medium, and when the program is executed on a computer, the program performs the following operations: Figure 1 、 Figure 2 and Figure 3 、 Figure 4 、 Figure 5 、 Figure 6 、 Figure 7 and Figure 8 Defect detection method in .

[0135] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0136] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0137] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0138] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0139] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk, and other media that can store program code.

Claims

1. A defect detection method based on a dynamic dual-threshold template, characterized in that: include: Collect defect-free sample images of several batches of target model display screens; Counting grayscale extreme value boundary data of each pixel point on the defect-free sample images of the batches; Generate a lower limit standard template and an upper limit standard template according to the grayscale extreme value boundary data; Obtain regional difference amplification coefficient and compensation offset; Generate a dynamic detection threshold template for the lower limit standard template and the upper limit standard template by using the regional difference amplification coefficient and the compensation offset; Acquire a screen image of the target model display screen to be tested; Performing pixel-level differentiation between the screen image to be tested and the dynamic detection threshold template to generate a defect mask image; Defect detection of the display screen is performed using the defect mask image to generate a first defect detection result.

2. The defect detection method according to claim 1, characterized in that: The regional difference amplification coefficient includes an upper limit regional difference amplification coefficient and a lower limit regional difference amplification coefficient, the compensation offset includes an upper limit compensation offset and a lower limit compensation offset, and the dynamic detection threshold template includes a bright defect dynamic detection threshold template and a dark defect dynamic detection threshold template; The step of generating a dynamic detection threshold template for the lower limit standard template and the upper limit standard template by using the regional difference amplification coefficient and the compensation offset includes: Generate a bright defect dynamic detection threshold template through the upper limit area difference magnification coefficient, the upper limit compensation offset, the lower limit standard template and the upper limit standard template; A dark defect dynamic detection threshold template is generated by using the lower limit area difference magnification coefficient, the lower limit compensation offset, the lower limit standard template and the upper limit standard template.

3. The defect detection method according to claim 2, characterized in that: The target model display screen is a quantum dot electroluminescent display screen; The step of obtaining the regional difference amplification coefficient and the compensation offset comprises: determining the radial thickness distribution of the quantum dot electroluminescent display corresponding to each defect-free sample image; Dividing each defect-free sample image into a high-thickness region and a low-thickness region according to the radial thickness distribution, wherein the high-thickness region is a region having a radial thickness higher than an upper limit of a reference thickness range, and the low-thickness region is a region having a radial thickness higher than an upper limit of the reference thickness range; Calculating the upper limit region difference amplification coefficient and the upper limit compensation offset according to the grayscale data, grayscale gradient data and film reflectivity data of the high thickness region; Calculating the lower limit area difference amplification coefficient and the lower limit compensation offset according to the grayscale data, grayscale gradient data and film reflectivity data of the low thickness area; The upper limit region difference magnification coefficient, upper limit compensation offset, lower limit region difference magnification coefficient and lower limit compensation offset corresponding to each defect-free sample image are averaged.

4. The defect detection method according to claim 2, characterized in that: The defect mask image includes a bright defect mask image and a dark defect mask image; The step of performing pixel-level differentiation between the screen image to be tested and the dynamic detection threshold template to generate a defect mask image includes: Performing pixel-level differentiation between the screen image to be tested and the bright defect dynamic detection threshold template to generate a bright defect mask image; Perform pixel-level differentiation between the screen image to be tested and the dark defect dynamic detection threshold template to generate a dark defect mask image.

5. The defect detection method according to any one of claims 1 to 4, characterized in that: After the step of performing defect detection on the display screen using the defect mask image to generate a first defect detection result, the defect detection method further includes: Performing connected domain segmentation on the defect mask image to determine abnormal areas; determining a background area according to the abnormal area; Analyzing the grayscale difference between the abnormal area and the background area, and performing threshold control processing on the defect to generate a second detection result; Perform defect integration analysis on the first defect detection result and the second defect detection result.

6. The defect detection method according to claim 5, characterized in that: The step of determining the background area according to the abnormal area includes: Determine the minimum inscribed circle radius of the abnormal area, and dilate the abnormal area by the radius of the inscribed circle to generate a background area; When the background area includes other abnormal areas, the portion of the background area that intersects with the other abnormal areas is removed.

7. The defect detection method according to claim 5, characterized in that: The step of analyzing the grayscale difference between the abnormal area and the background area, performing threshold control processing for defects, and generating a second detection result includes: Calculating a first grayscale mean value and a second grayscale mean value of the abnormal area and the background area; Calculating a contrast attribute difference value of the defect based on the first grayscale mean and the second grayscale mean; A threshold control process is performed on the contrast attribute difference value to generate a second detection result.

8. The defect detection method according to any one of claims 1 to 4, characterized in that: The step of generating a lower limit standard template and an upper limit standard template according to the grayscale extreme value boundary data comprises: Determine the boundary grayscale value of each pixel point according to the grayscale extreme value boundary data and the boundary ratio; A lower limit standard template and an upper limit standard template are generated according to the boundary grayscale value of each pixel point.

9. A defect detection device based on a dynamic dual-threshold template, characterized in that: include: An acquisition unit, used for acquiring images of several batches of defect-free samples of the target model display screen; a statistical unit, configured to count grayscale extreme value boundary data of each pixel point on the defect-free sample images of the batches; A first generating unit, configured to generate a lower limit standard template and an upper limit standard template according to the grayscale extreme value boundary data; A first acquiring unit is used to acquire a regional difference amplification coefficient and a compensation offset; A second generating unit is configured to generate a dynamic detection threshold template for the lower limit standard template and the upper limit standard template by using the regional difference amplification coefficient and the compensation offset; A second acquisition unit is used to acquire a screen image to be tested of the target model display screen; A third generating unit is configured to perform pixel-level differentiation between the screen image to be tested and the dynamic detection threshold template to generate a defect mask image; The fourth generating unit is configured to perform defect detection on the display screen using the defect mask image to generate a first defect detection result.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and when the program is executed on a computer, the defect detection method according to any one of claims 1 to 8 is executed.

Images