Defect detection method and device based on phase deflection technology and storage medium

Abstract

The invention discloses a defect detection method and device based on phase deflection, and a storage medium, which are used for improving the detection capability of different significant defects. Obtaining a reflection fringe image of the display screen to be detected; performing phase extraction and phase unwrapping processing on the reflection fringe image to generate phase distribution data; extracting gradient data of the phase distribution data; according to the gradient data and a saliency threshold value, multi-level saliency recognition is carried out on the defect area of the reflection fringe image, multi-level saliency recognition results are generated, and the multi-level saliency recognition results comprise a high saliency detection result, a middle saliency detection result and a low saliency detection result. Different levels of saliency recognition results correspond to different types of defects; performing hierarchical classification processing on the multi-level saliency recognition result; and outputting a final detection result after grading and classification processing.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of display screen detection, and in particular to a defect detection method, device and storage medium based on phase deflectometry. Background Art

[0002] Defect detection of flexible display screens is a necessary process for adjustment. Defects of flexible display screens can be divided into display defects and screen body defects. For example, in the production process of flexible display screens, surface defects such as foreign matter, dirt, dust, scratches, and scratches are prone to occur, which affect the screen quality and user experience. Display defects refer to the discordant areas that appear in the display screen during the display screen display process. In the prior art, display defects usually require inputting a specific picture on the display screen to be tested, and then collecting the picture through the camera, and then performing image detection to analyze it. Screen body defects refer to areas where damage occurs on the surface or inside of the display screen itself (such as chipped corners, foreign matter, etc.). Such defects can be analyzed by collecting the picture with the camera and then performing image detection.

[0003] As the application fields of flexible displays become more and more extensive and more and more functions are added, the pixel structure of flexible displays is becoming more and more complex. For example, the pixel spacing is getting smaller and smaller, the curvature of the flexible display is getting larger and larger, and circuit areas are set on the flexible display. These make defect detection of flexible displays more difficult.

[0004] Existing detection technologies lack the ability to classify and grade defects of different significance, and are unable to provide detailed defect categories and significance levels during detection, which reduces the detection capability of defects of different significance. Summary of the invention

[0005] The present application discloses a defect detection method, device and storage medium based on phase deflectometry, which are used to improve the detection capability of defects with different significance.

[0006] In a first aspect, an embodiment of the present application provides a defect detection method based on phase deflectometry, comprising: acquiring a reflection stripe image of a display screen to be tested; performing phase extraction and phase unwrapping processing on the reflection stripe image to generate phase distribution data; extracting gradient data of the phase distribution data; performing multi-level significance recognition on a defective area of ​​the reflection stripe image according to the gradient data and a preset significance threshold value to generate a multi-level significance recognition result, wherein the multi-level significance recognition result comprises a high significance detection result, a medium significance detection result, and a low significance detection result, and significance recognition results of different levels correspond to different types of defects; performing hierarchical classification processing on the multi-level significance recognition results; and outputting the final detection result after the hierarchical classification processing.

[0007] Optionally, in some embodiments of the present application, after the step of performing multi-level saliency recognition on the defect area of ​​the reflective stripe image according to the gradient data and the preset significance threshold to generate the multi-level saliency recognition result, and before the step of performing hierarchical classification processing on the multi-level saliency recognition result, the defect detection method also includes: obtaining an enhancement coefficient of the low-saliency area; and performing adaptive enhancement processing on the area corresponding to the low-saliency detection result according to the enhancement coefficient.

[0008] Optionally, in some embodiments of the present application, the step of obtaining the enhancement coefficient of the low-saliency region includes: generating the enhancement coefficient of the low-saliency region according to the medium-saliency detection result and the low-saliency detection result.

[0009] Optionally, in some embodiments of the present application, the step of performing hierarchical classification processing on the multi-level saliency recognition results includes: performing saliency grading processing on the defects in the multi-level saliency recognition results; analyzing the geometric characteristics of the defects in the multi-level saliency recognition results to generate shape factor data; and classifying the defects according to the grading results and the shape factor data.

[0010] Optionally, in some embodiments of the present application, the step of obtaining a reflected fringe image of the display screen to be tested includes: presetting a light source configuration, the light source is used to project sine and cosine lights of different phases; setting a reflector group, the reflector group is used to adjust the light path; using a time-sharing strobe function to project sine and cosine lights of different phases onto the display screen to be tested; and collecting the reflected fringe image through a high-resolution camera and outputting it digitally.

[0011] Optionally, in some embodiments of the present application, phase extraction and phase unwrapping processing are performed on the reflected fringe image, and the step of generating phase distribution data includes: extracting 8 phase images from the collected reflected fringe image, wherein the 8 phase images include 4 x-direction and 4 y-direction phase images; using the phase shift method to extract and phase unwrapping the 8 phase images to generate phase distribution data.

[0012] Optionally, in some embodiments of the present application, the step of outputting the final detection result after hierarchical classification processing includes: visually displaying and statistically analyzing the multi-level significance recognition results after hierarchical classification processing, generating and outputting the final detection result, and the final detection result includes at least a small local defect image, defect category, defect feature quantity and the original image.

[0013] Optionally, in some embodiments of the present application, the display screen to be tested is a display screen having an internal circuit arranged in a pixel layer; before the step of performing multi-level significance recognition on the defect area of ​​the reflection stripe image according to the gradient data and the preset significance threshold to generate the multi-level significance recognition result, the defect detection method further comprises: lighting the spherical integrating light source with a preset brightness according to the average thickness value of the internal circuit of the display screen to be tested and the average thickness of the display screen to be tested; irradiating the back of the display screen to be tested with the spherical integrating light source, and photographing the display screen to be tested from the front using an acquisition device to generate a first photographed image; lighting the display screen to be tested so that the display screen to be tested displays a preset grayscale picture, and acquiring a second photographed image of the display screen to be tested through the acquisition device; binarizing the first photographed image to determine the internal circuit area on the first photographed image; determining the internal circuit area on the second photographed image according to the internal circuit area on the first photographed image; determining a first grayscale mean of the internal circuit area on the second photographed image and a second grayscale mean of the non-circuit area on the second photographed image; acquiring a preset significance threshold, adjusting the preset significance threshold by the first grayscale mean and the second grayscale mean, and generating a significance threshold of the internal circuit area.

[0014] In a second aspect, an embodiment of the present application provides a defect detection device based on phase deflectometry, comprising: an acquisition unit, used to acquire a reflection stripe image of a display screen to be tested; a first generation unit, used to perform phase extraction and phase unwrapping processing on the reflection stripe image to generate phase distribution data; an extraction unit, used to extract gradient data of the phase distribution data; a second generation unit, used to perform multi-level significance recognition on the defect area of ​​the reflection stripe image according to the gradient data and a preset significance threshold, and generate multi-level significance recognition results, the multi-level significance recognition results including high significance detection results, medium significance detection results, and low significance detection results, and significance recognition results of different levels correspond to different types of defects; a processing unit, used to perform hierarchical classification processing on the multi-level significance recognition results; and an output unit, used to output the final detection result after the hierarchical classification processing.

[0015] Optionally, after the second generating unit and before the processing unit, the device further includes: a third generating unit, used to obtain an enhancement coefficient of the low-saliency area; and an enhancement unit, used to perform adaptive enhancement processing on the area corresponding to the low-saliency detection result according to the enhancement coefficient.

[0016] Optionally, the third generating unit includes: generating an enhancement coefficient of the low-saliency region according to the medium-saliency detection result and the low-saliency detection result.

[0017] Optionally, the processing unit includes: performing saliency grading processing on defects in the multi-level saliency recognition results; analyzing geometric characteristics of the defects in the multi-level saliency recognition results to generate shape factor data; and classifying defects according to the grading processing results and the shape factor data.

[0018] Optionally, the acquisition unit includes: presetting a light source configuration, the light source is used to project sine and cosine lights of different phases; setting a reflector group, the reflector group is used to adjust the light path; using a time-sharing stroboscopic function to project sine and cosine lights of different phases onto the display screen to be tested; and collecting the reflected fringe image through a high-resolution camera and outputting it digitally.

[0019] Optionally, the first generating unit includes: extracting 8 phase images from the collected reflected fringe images, wherein the 8 phase images include 4 x-direction and 4 y-direction phase images; using a phase shift method to perform extraction processing and phase unwrapping processing on the 8 phase images to generate phase distribution data.

[0020] Optionally, the output unit includes: visually displaying and statistically analyzing the multi-level saliency recognition results after hierarchical classification processing, generating and outputting the final detection result, and the final detection result at least includes a small local defect image, defect category, defect feature quantity and the original image.

[0021] Optionally, the display screen to be tested is a display screen having an internal circuit arranged in the pixel layer; before the step of the second generating unit, the defect detection device further includes: a lighting unit, which is used to light up the spherical integrating light source with a preset brightness according to the average thickness value of the internal circuit of the display screen to be tested and the average thickness of the display screen to be tested; a fourth generating unit, which is used to illuminate the back of the display screen to be tested through the spherical integrating light source, and use the acquisition device to shoot the display screen to be tested from the front to generate a first captured image; a fifth generating unit, which is used to light up the display screen to be tested so that the display screen to be tested displays a preset grayscale picture, and obtain a second captured image of the display screen to be tested through the acquisition device; a first determining unit, which is used to perform binarization processing on the first captured image to determine the internal circuit area on the first captured image; a second determining unit, which is used to determine the internal circuit area on the second captured image through the internal circuit area on the first captured image; a third determining unit, which is used to determine the first grayscale mean of the internal circuit area on the second captured image and the second grayscale mean of the non-circuit area on the second captured image; a sixth generating unit, which is used to obtain a preset significance threshold, and adjust the preset significance threshold through the first grayscale mean and the second grayscale mean to generate a significance threshold of the internal circuit area.

[0022] In a third aspect, an embodiment of the present application provides a defect detection device based on phase deflectometry, comprising: Processor, memory, input-output unit, and bus; The processor is connected to the memory, the input and output unit, 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.

[0023] 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 performs the first aspect and any optional defect detection method of the first aspect.

[0024] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: In the present application, the reflection stripe image of the display screen to be tested is first obtained by a preset light source and an image acquisition device. Then, the reflection stripe image is subjected to phase extraction and phase unwrapping processing to generate phase distribution data to achieve phase continuity. Then, the gradient data of the phase distribution data is extracted. After obtaining a preset significance threshold, multi-level significance recognition is performed on the defective area of ​​the reflection stripe image according to the gradient data and the preset significance threshold, so as to generate a multi-level significance recognition result, wherein the multi-level significance recognition result includes a high significance detection result, a medium significance detection result, and a low significance detection result, and significance recognition results of different levels correspond to different types of defects. Then, the multi-level significance recognition results are subjected to hierarchical classification processing. And the final detection result after hierarchical classification processing is output.

[0025] Defects are detected in a graded manner based on the gradient amplitude in the gradient data, and high-significance, medium-significance and low-significance thresholds are set in advance. This allows different types of defects, such as deep scratches, large dents, shallow scratches, slight dents and dirt, to be fully covered during the detection process, thereby improving the detection capability of defects with different significances. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative labor.

[0027] Figure 1 A schematic diagram of a first embodiment of a defect detection method based on phase deflectometry of the present application; Figure 2 A schematic diagram of a first embodiment of the method for low-significant defect enhancement of the present application; Figure 3 A schematic diagram of a first embodiment of a method for generating enhancement coefficients for the present application; Figure 4A schematic diagram of a first embodiment of a method for hierarchically classifying multi-level saliency recognition results of the present application; Figure 5 A schematic diagram of a first embodiment of a method for obtaining a reflection fringe image in the present application; Figure 6 A schematic diagram of a first embodiment of a method for generating phase distribution data according to the present application; Figure 7 A schematic diagram of a first embodiment of the method for outputting detection results of the present application; Figure 8 A schematic diagram of a first embodiment of a method for generating a significance threshold value for an internal circuit region of the present application; Fig. 9 A schematic diagram of a first embodiment of a defect detection device based on phase deflectometry of the present application; Fig.10 It is a schematic diagram of a second embodiment of a defect detection device based on phase deflectometry of the present application. DETAILED DESCRIPTION

[0028] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also 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 prevent unnecessary details from obstructing the description of the present application.

[0029] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0030] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0031] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" 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 "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0032] 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.

[0033] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0034] In the prior art, as the application fields of flexible display screens become more and more extensive and more and more functions are added, the pixel structure of flexible display screens becomes more and more complex. For example, the pixel spacing becomes smaller and smaller, the curvature of the flexible display screen becomes larger and larger, and a circuit area is provided on the flexible display screen, etc. These make it more difficult to detect defects in flexible display screens.

[0035] Existing detection technologies lack the ability to classify and grade defects of different significance, and are unable to provide detailed defect categories and significance levels during detection, which reduces the detection capability of defects of different significance.

[0036] Based on this, the present application discloses a defect detection method, device and storage medium based on phase deflectometry, which are used to improve the detection capability of defects of different significance.

[0037] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0038] The method of the present application can be applied to a server, a device, a terminal or other devices with logic processing capabilities, and the present application does not limit this. For the convenience of description, the following description is made by taking the execution subject as an example of a terminal.

[0039] See also Figure 1 The present application provides an embodiment of a defect detection method based on phase deflectometry, comprising: 101. Obtain a reflection fringe image of the display screen to be tested.

[0040] In an embodiment of the present application, the terminal obtains the reflected fringe image of the display screen to be tested. It is necessary to project sine and cosine lights of different phases to the display screen to be tested through a set light source, and then photograph the display screen to be tested through an image acquisition device to obtain the reflected fringe image.

[0041] Phase deflectometry is a high-precision detection method based on interference fringe analysis and phase information calculation. It can be used to detect surface defects of display screens by extracting phase gradient information on the surface of the object being tested. However, the existing phase deflectometry detection methods for display screens have the same problems as conventional detection technologies, lacking the ability to classify and grade defects of different significance, and failing to provide detailed defect categories and significance levels during detection, which reduces the ability to detect defects of different significance.

[0042] 102. Perform phase extraction and phase unwrapping processing on the reflected fringe image to generate phase distribution data.

[0043] In an embodiment of the present application, the terminal extracts phase distribution from the reflected fringe image, specifically by performing phase extraction and phase unwrapping processing on the reflected fringe image through a phase deflection method to generate phase distribution data.

[0044] 103. Extract gradient data of the phase distribution data.

[0045] The terminal extracts the gradient data of the phase distribution data through the gradient calculation module, and extracts the gradient information of the phase distribution for saliency detection and quantification of surface saliency changes. Specifically, the gradient data is the gradient amplitude, and the phase distribution data is the phase information after phase unwrapping:

[0046]

[0047] The gradient amplitude is defined as , is the phase at (x, y) coordinate The parameters to be integrated, It is the phase information after phase unwrapping.

[0048] The physical significance of the gradient calculation of phase distribution data is mainly: 1. High gradient amplitude: usually corresponds to significant changes in surface height (such as deep scratches, depressions).

[0049] 2. Low gradient amplitude: corresponds to smooth changes or surface noise.

[0050] 104. Perform multi-level saliency recognition on the defect area of ​​the reflection stripe image according to the gradient data and the preset saliency threshold, and generate a multi-level saliency recognition result. The multi-level saliency recognition result includes a high saliency detection result, a medium saliency detection result, and a low saliency detection result. Different levels of saliency recognition results correspond to different types of defects.

[0051] The terminal performs multi-level saliency recognition on the defect area of ​​the reflection stripe image according to the gradient data and the preset saliency threshold, and generates a multi-level saliency recognition result.

[0052] Specifically, the defect area is segmented mainly by gradient amplitude and preset significance threshold, and multi-level significance recognition is performed on the defect area so as to separate defects of different significance levels.

[0053] Specifically, the threshold is set first. The basis of significance detection is the gradient amplitude. , which indicates the degree of phase change. By setting different significance thresholds, the defects are divided into high significance, medium significance and low significance areas: High significance threshold ( ): For significant defects such as deep scratches and large depressions, the gradient amplitude is higher.

[0054] The significance threshold ( ): For medium-significant defects such as shallow scratches and slight dents.

[0055] Low significance threshold ( ): For low-visibility defects such as dirt and small particles.

[0056] Specific classification formula:

[0057] The hierarchical detection process includes high-significance detection, medium-significance detection, and low-significance detection.

[0058] 1. High significance detection: Filter to meet Typically detects significant areas such as deep scratches or large dents.

[0059] Output: Binary image of high saliency regions, marked in red.

[0060] 2. Medium significance detection: Filter to meet Detects less significant defects such as shallow scratches or slight dents.

[0061] Output: A binary image of the salient regions, marked in yellow.

[0062] 3. Low saliency detection: Filter to meet Pixels. Detect tiny significant defects such as surface dirt and small particles.

[0063] Output: Binary image of low saliency regions, marked in blue.

[0064] The priority fusion mechanism integrates the results of multiple layers of saliency detection to generate a graded saliency distribution map, and uses color coding (red, yellow, blue) to intuitively display the defect distribution. Based on the saliency level, area, shape factor and texture characteristics (such as uniformity), a classification method for three types of defects, namely scratches, dents and dirt, is proposed, and a grading of saliency levels is provided.

[0065] 105. Perform hierarchical classification on the multi-level saliency recognition results.

[0066] In the embodiment of the present application, the terminal performs hierarchical classification processing on the multi-level saliency recognition results, that is, further subdividing the multi-level saliency recognition results to obtain more detailed defect detection results.

[0067] 106. Output the final detection result after grading and classification.

[0068] The terminal finally outputs the final inspection results after hierarchical classification, and while outputting the classified inspection data, it also outputs other information related to the defects, such as intermediate parameters and gradient parameters generated during the inspection process.

[0069] In an embodiment of the present application, a reflection stripe image of the display screen to be tested is first obtained by a preset light source and an image acquisition device. Then, the reflection stripe image is subjected to phase extraction and phase unwrapping processing to generate phase distribution data to achieve phase continuity. Then, the gradient data of the phase distribution data is extracted. After obtaining a preset significance threshold, multi-level significance recognition is performed on the defective area of ​​the reflection stripe image according to the gradient data and the preset significance threshold, thereby generating a multi-level significance recognition result, wherein the multi-level significance recognition result includes a high significance detection result, a medium significance detection result, and a low significance detection result, and significance recognition results of different levels correspond to different types of defects. Then, the multi-level significance recognition result is subjected to hierarchical classification processing. And the final detection result after hierarchical classification processing is output.

[0070] Defects are detected in a graded manner based on the gradient amplitude in the gradient data, and high-significance, medium-significance and low-significance thresholds are set in advance. This allows different types of defects, such as deep scratches, large dents, shallow scratches, slight dents and dirt, to be fully covered during the detection process, thereby improving the detection capability of defects with different significances.

[0071] See also Figure 2The present application provides an embodiment of a method for low-significant defect enhancement, comprising: 201. Obtain enhancement coefficients for low-significance regions.

[0072] 202. Adaptively enhance the area corresponding to the low-saliency detection result according to the enhancement coefficient.

[0073] In the embodiment of the present application, the low-saliency area can be adaptively enhanced. Specifically, the low-saliency area (such as dirt and small particles) is adaptively enhanced to improve the detection capability. The formula is as follows:

[0074] Among them, k is the enhancement coefficient, which can be set by human experience to adjust the amplification intensity of low-significance areas. is the enhanced phase distribution, is the original phase distribution, is the preset phase distribution change. Adaptive enhancement will significantly improve the detection capability in low gradient change areas, while having little effect on high gradient areas.

[0075] After completing the enhancement of low-saliency defects, saliency fusion is required to generate a complete saliency distribution map by fusing the high, medium, and low saliency detection results: 1. High-significance areas give priority to covering medium- and low-significance areas.

[0076] 2. Regions of different significance are displayed in color codes (e.g., red for high significance, yellow for medium significance, and blue for low significance).

[0077] 3. Output significance classification map for subsequent defect classification and grading.

[0078] The priority fusion mechanism integrates the results of multiple layers of saliency detection to generate a graded saliency distribution map, and uses color coding (red, yellow, blue) to intuitively display the defect distribution. Based on the saliency level, area, shape factor and texture characteristics (such as uniformity), a classification method for three types of defects, namely scratches, dents and dirt, is proposed, and a grading of saliency levels is provided.

[0079] See also Figure 3 The present application provides an embodiment of a method for generating an enhancement coefficient, comprising: 301. Generate an enhancement coefficient for a low-saliency region according to the medium-saliency detection result and the low-saliency detection result.

[0080] In the embodiment of the present application, the terminal generates an enhancement coefficient of the low-significance region according to the medium-significance detection result and the low-significance detection result. The enhancement coefficient k can also be set according to a preset threshold value. and To generate, the formula is as follows:

[0081] In order to make the system more robust, adaptive enhancement technology is introduced and a preset significance threshold is used. and , that is, combined with low-significance threshold areas for targeted enhancement. For low-significance defect areas (such as dirt and small particles), a dynamic enhancement algorithm is introduced to improve the detection ability of weak signals without affecting high-significance areas.

[0082] See also Figure 4 The present application provides an embodiment of a method for hierarchical classification of multi-level saliency recognition results, comprising: 401. Perform saliency classification processing on defects in multi-level saliency recognition results.

[0083] 402. Analyze the geometric characteristics of defects in the multi-level saliency recognition results to generate shape factor data.

[0084] 403. Defect classification is performed based on the grading processing results and the shape factor data.

[0085] In the embodiment of the present application, the terminal performs saliency grading processing on the defects in the multi-level saliency recognition results for pre-classification. The geometric characteristics of the defects in the multi-level saliency recognition results are analyzed to generate shape factor data. Then, the defects are classified (sub-classified) according to the grading processing results and the shape factor data.

[0086] Specifically, defect classification and grading are important steps to further refine the saliency detection results, which mainly include saliency grading, geometric characteristic analysis and classification processing.

[0087] Defect classification: Based on the significance level and the area of ​​the defect, the defects are classified according to the following rules: Highly significant defects: Gradient Amplitude , usually significant defects such as deep scratches and large dents, covering a large area.

[0088] Significant defects: Gradient Amplitude , such as shallow scratches, slight dents, and medium size.

[0089] Low significance defects: Gradient Amplitude , the surface is dirty, with small particles and a small area.

[0090] Next, we conduct geometric characteristic analysis and further analyze it in combination with the geometric characteristics of the salient area: 1. Area: Calculate the total number of pixels of the defect and convert it into a physical unit area :

[0091] Where N is the number of pixels. The actual physical size of a single pixel.

[0092] 2. Perimeter: Detect the boundary of the defect area and calculate the perimeter:

[0093] 3. Shape factor: Analyze the defect shape characteristics through the relationship between area and perimeter:

[0094] : Round defects (such as particles or dirt).

[0095] : Non-circular defects (such as scratches).

[0096] Next, we need to refer to the classification criteria and classify the defects into the following categories based on geometric and texture characteristics: 1. Scratches: High or medium saliency regions.

[0097] Long strip distribution, small shape factor.

[0098] Linear texture distribution is prominent.

[0099] 2. Depression: High or medium saliency regions.

[0100] The area is large and the shape factor is close to 1.

[0101] 3. Dirty: Low saliency area.

[0102] Isolated distribution, shape factor close to 1, and uneven texture characteristics.

[0103] The specific classification formula is as follows: 1. Texture uniformity analysis: Using the gray uniformity descriptor:

[0104] U: Uniformity descriptor.

[0105] : Standard deviation of grayscale in local area.

[0106] : The mean grayscale value of the local area.

[0107] The higher the U, the more likely it is a scratch or dent, and the lower the U, the more likely it is dirt.

[0108] The final classification rule can be expressed as follows: Scratch, if linear texture and Shape Factor<0.3 Dents, if Shape Factor>=0.7 Dirt, if uniformity U <= 0.5 See also Figure 5 The present application provides an embodiment of a method for acquiring a reflection fringe image, comprising: 501. The light source configuration is preset, and the light source is used to project sine and cosine lights of different phases.

[0109] 502. Set a reflector group, where the reflector group is used to adjust the light path.

[0110] 503. Use the time-sharing strobe function to project sine and cosine lights of different phases onto the display screen to be tested.

[0111] 504. The reflected fringe image is collected by a high-resolution camera and output digitally.

[0112] In the embodiment of the present application, the terminal first sets up an optical detection module, where the optical detection module is the front end of the entire system and is responsible for obtaining the reflected fringe image of the surface of the object under test through optical principles, providing high-quality input data for subsequent image processing. The key tasks of this module include: 1. Preset the light source configuration so that it can project sine and cosine light of different phases. The light source should have high brightness and high uniformity to ensure the clarity of the projected stripes. Use the time-sharing strobe function to project sine and cosine light of different phases onto the product (display to be tested).

[0113] 2. Optimize the optical path design and add a reflector group. The purpose is to adjust the optical path so that the interference fringes are evenly projected onto the surface of the object and optimize the reflection path of the fringes.

[0114] a) Specifically, high-quality reflectors are needed to ensure a stable optical path and efficient use of light energy.

[0115] b) Secondly, it is also necessary to use multi-layer coated lenses, which can reduce light energy loss and improve reflection efficiency.

[0116] 3. The reflected fringe images are collected by a high-resolution camera and output digitally to provide data for subsequent phase calculations.

[0117] See also Figure 6 The present application provides an embodiment of a method for generating phase distribution data, comprising: 601. Extract 8 phase images from the collected reflection fringe images, wherein the 8 phase images include 4 phase images in the x direction and 4 phase images in the y direction.

[0118] 602. Use a phase shift method to perform extraction processing and phase unwrapping processing on the eight phase images to generate phase distribution data.

[0119] In the embodiment of the present application, an image processing module is used to perform phase extraction and phase unwrapping processing on the collected reflection fringe image. The image processing module is the core part of the entire system, which is used to process the reflection fringe image obtained by the optical detection module, extract surface morphology information and identify defects. The specific subsystems are as follows: 1. Phase calculation module: Phase calculation is the first step in image processing and is responsible for extracting the initial phase distribution from the reflected fringe image. The details are as follows: a) First, image extraction is performed. Using the time-sharing stroboscopic function and the 4-step phase shift method, 8 phase images are extracted from the collected fringe phase image, including 4 phase images in the x-direction and 4 phase images in the y-direction.

[0120] Light intensity distribution of fringe image: The light intensity distribution of interference fringe image can be expressed as:

[0121] is the pixel point in the reflected fringe image The light intensity on.

[0122] is the background light intensity.

[0123] It is the fringe modulation degree, which is related to the fringe contrast.

[0124] is the phase distribution, reflecting the surface height information.

[0125] The phase extraction method is to use the phase shift method to extract the phase. Specifically, the phase shift method requires the display to display multiple images with different initial phases and obtain them by the camera. When sinusoidal fringes are used, assuming that the N-step phase shift method is used, the light intensity expression received by the camera (image acquisition device) is:

[0126] In the formula, is the image intensity when the phase shift is n steps, which is a known quantity obtained by the camera, represents the background light intensity distribution, Indicates the modulation distribution, phase is the unknown quantity to be solved, so we need to find At least three phase shifts are required, that is, N is at least 3. The least squares method can be used to solve it, where The result is:

[0127] Since the inverse tangent function has a range of ,so is the absolute phase value folded into Wrapped phase within the interval. There are multiple grayscale cutoffs in the folded phase image. The cutoffs are aliased with the object under test and the defects, and the defects cannot be identified. Therefore, the obtained wrapped phase must be unfolded. This process is called phase unfolding.

[0128] d) Phase unwrapping: Since the phase is usually limited to the interval , there may be jumps in the phase calculation, resulting in discontinuity. Phase unpacking achieves phase continuity by removing jumps. The phase information after bit unpacking is:

[0129] Where N is an integer multiple added during the unpacking process , For phase distribution, ensure phase continuity.

[0130] See also Figure 7 The present application provides an embodiment of a method for outputting a detection result, comprising: 701. Visually display and statistically analyze the multi-level saliency recognition results after hierarchical classification, generate and output the final detection results, and the final detection results at least include a small local defect image, defect category, defect feature quantity and original image.

[0131] In the embodiment of the present application, the terminal performs visual display and statistical analysis on the multi-level significance recognition results after hierarchical classification through the result output module.

[0132] This module is the terminal part of the entire system, responsible for visual display and statistical analysis of the inspection results, and generating easy-to-understand reports. The terminal will output key information such as a small local defect image, defect category, defect feature quantity, and original image.

[0133] See also Figure 8 The present application provides an embodiment of a method for generating a significance threshold of an internal circuit area, comprising: 801. Light up the spherical integrating light source with a preset brightness according to an average thickness value of an internal circuit of the display screen to be tested and an average thickness of the display screen to be tested.

[0134] 802 . Illuminate the back side of the display screen to be tested with a spherical integral light source, and photograph the display screen to be tested from the front side using a capture device to generate a first photographed image.

[0135] 803 . Light up the display screen to be tested, so that the display screen to be tested displays a preset grayscale image, and obtain a second captured image of the display screen to be tested through a capture device.

[0136] 804 . Perform binarization processing on the first captured image to determine an internal circuit region on the first captured image.

[0137] 805 . Determine an internal circuit area on the second captured image according to the internal circuit area on the first captured image.

[0138] 806. Determine a first grayscale mean value of the internal circuit area on the second captured image and a second grayscale mean value of the non-circuit area on the second captured image.

[0139] 807 . Obtain a preset significance threshold, and adjust the preset significance threshold by using the first grayscale mean and the second grayscale mean to generate a significance threshold of the internal circuit area.

[0140] In this embodiment, for an LCD screen with an internal circuit arranged at the pixel layer, a thin-sheet circuit is added to the pixel layer of the LCD screen to increase the functionality of the LCD screen body. Usually, it is arranged below the pixel layer. However, the structure for the internal circuit is complicated with the complexity of the function. There are overlapping circuit areas or thickened circuit areas. Such areas will cause changes in regional reflections, and thus cause abnormal detection, especially in the internal circuit area. Therefore, in the embodiment of the present application, it is necessary to first determine the part of the circuit area, and then adjust the preset significance threshold. The original preset significance threshold is used to detect the non-circuit area, and the adjusted significance threshold is used to detect the internal circuit area.

[0141] First, the spherical integral light source is lit with a preset brightness according to the average thickness value of the internal circuit of the display screen to be tested and the average thickness of the display screen to be tested, that is, parameters with higher contrast (light source brightness and light source type) are selected according to the average thickness value of the internal circuit and the average thickness of the display screen to be tested, so that the light projected by the integral light source can pass through the non-circuit area of ​​the LCD screen as much as possible and be blocked by the internal circuit area as much as possible. Too high brightness causes the light source to pass through the thin circuit, so it is necessary to limit the brightness and type of the spherical integral light source within the calculated range.

[0142] Next, the back of the display screen to be tested is illuminated by a spherical integral light source, and the display screen to be tested is photographed from the front using an acquisition device to generate a first captured image. Then, the display screen to be tested is illuminated so that the display screen to be tested displays a preset grayscale image, and a second captured image of the display screen to be tested is obtained by the acquisition device, and the detection is performed at the preset grayscale. The first captured image is binarized to determine the internal circuit area on the first captured image. Due to the parameter adjustment of the spherical integral light source, the contrast between the internal circuit area and the non-circuit area is increased, and the binarization can distinguish the two areas on the LCD screen. The internal circuit area on the second captured image is determined by the internal circuit area on the first captured image, that is, the display screen areas of the two images are overlapped, and then the internal circuit area on the first captured image is mapped to the second captured image, and then the first grayscale mean of the internal circuit area on the second captured image and the second grayscale mean of the non-circuit area on the second captured image are analyzed, and the significance threshold is adjusted according to the difference between the two.

[0143]

[0144]

[0145]

[0146] in, is the first grayscale mean value of the internal circuit area on the second captured image, is the second grayscale average of the non-circuit area on the second captured image. In the embodiment of the present application, since the internal circuit is located below the pixel layer, it can reflect a portion of the light of the pixel layer, so that Usually greater than Through the above adjustment, the defect detection problem inside the circuit area can be well solved, that is, an adapted significance threshold can also be obtained in the internal circuit area.

[0147] See also Fig. 9 The present application provides an embodiment of a device for detecting repeated defects of a display screen, comprising: The lighting unit 901 is used to light up the spherical integrating light source with a preset brightness according to the average thickness value of the internal circuit of the display screen to be tested and the average thickness of the display screen to be tested.

[0148] The fourth generating unit 902 is configured to illuminate the back side of the display screen to be tested by using a spherical integral light source, and photograph the display screen to be tested from the front side using a capture device to generate a first photographed image.

[0149] The fifth generating unit 903 is used to light up the display screen to be tested so that the display screen to be tested displays a preset grayscale image, and obtain a second captured image of the display screen to be tested through a capture device.

[0150] The first determining unit 904 is configured to perform binarization processing on the first captured image to determine an internal circuit region on the first captured image.

[0151] The second determining unit 905 is configured to determine the internal circuit area on the second captured image according to the internal circuit area on the first captured image.

[0152] The third determining unit 906 is used to determine a first grayscale mean value of the internal circuit area on the second captured image and a second grayscale mean value of the non-circuit area on the second captured image.

[0153] The sixth generating unit 907 is used to obtain a preset significance threshold, adjust the preset significance threshold by using the first grayscale mean value and the second grayscale mean value, and generate a significance threshold of the internal circuit area.

[0154] The acquisition unit 908 is used to acquire the reflection fringe image of the display screen to be tested.

[0155] Optionally, the acquiring unit 908 includes: The light source configuration is preset, and the light source is used to project sine and cosine lights of different phases.

[0156] A reflector group is provided, wherein the reflector group is used to adjust the light path.

[0157] The time-sharing strobe function is used to project sine and cosine lights of different phases onto the display screen to be tested.

[0158] The reflected fringe images are captured by a high-resolution camera and digitized for output.

[0159] The first generating unit 909 is used to perform phase extraction and phase unwrapping processing on the reflected fringe image to generate phase distribution data.

[0160] Optionally, the first generating unit 909 includes: Eight phase images are extracted from the collected reflection fringe images, wherein the eight phase images include four phase images in the x direction and four phase images in the y direction.

[0161] The eight phase images were extracted and phase unwrapped using the phase shift method to generate phase distribution data.

[0162] The extraction unit 910 is used to extract gradient data of the phase distribution data.

[0163] The second generating unit 911 is used to perform multi-level saliency recognition on the defect area of ​​the reflection stripe image according to the gradient data and the preset saliency threshold, and generate a multi-level saliency recognition result. The multi-level saliency recognition result includes a high saliency detection result, a medium saliency detection result, and a low saliency detection result. Different levels of saliency recognition results correspond to different types of defects.

[0164] The third generating unit 912 is used to obtain the enhancement coefficient of the low saliency area.

[0165] Optionally, the third generating unit 912 includes: The enhancement coefficient of the low saliency region is generated according to the medium saliency detection results and the low saliency detection results.

[0166] The enhancement unit 913 is used to perform adaptive enhancement processing on the area corresponding to the low-saliency detection result according to the enhancement coefficient.

[0167] The processing unit 914 is used to perform hierarchical classification processing on the multi-level saliency recognition results.

[0168] Optionally, the processing unit 914 includes: The defects in the multi-level saliency recognition results are processed by saliency classification.

[0169] The geometric characteristics of defects in the multi-level saliency recognition results are analyzed to generate shape factor data.

[0170] Defect classification is performed based on the grading results and shape factor data.

[0171] The output unit 915 is used to output the final detection result after the hierarchical classification processing.

[0172] Optionally, the output unit 915 includes: The multi-level saliency recognition results after hierarchical classification are visualized and statistically analyzed to generate and output the final detection results, which at least include a small local defect image, defect category, defect feature quantity and the original image.

[0173] See also Fig.10 The present application provides a defect detection device based on phase deflectometry, comprising: Processor 1001 , memory 1002 , input / output unit 1003 , and bus 1004 .

[0174] The processor 1001 is connected to the memory 1002 , the input and output unit 1003 , and the bus 1004 .

[0175] The memory 1002 stores a program, and the processor 1001 calls the program to execute the following steps: Figure 1 , Figure 2 and Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8 Defect detection method in .

[0176] The present application provides a computer-readable storage medium, on which a program is stored, and when the program is executed on a computer, the program performs the following steps: Figure 1 , Figure 2 and Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8 Defect detection method in .

[0177] Those skilled in the art can 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.

[0178] In the several embodiments provided in the present 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 only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, 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.

[0179] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0180] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0181] 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 is essentially 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, including several instructions to enable 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 phase deflectometry, characterized in that: include: Acquire a reflection fringe image of the display screen to be tested; Performing phase extraction and phase unwrapping processing on the reflected fringe image to generate phase distribution data; extracting gradient data of the phase distribution data; Performing multi-level saliency recognition on the defect area of ​​the reflection fringe image according to the gradient data and a preset saliency threshold, and generating a multi-level saliency recognition result, wherein the multi-level saliency recognition result includes a high saliency detection result, a medium saliency detection result, and a low saliency detection result, and different levels of saliency recognition results correspond to different types of defects; The multi-level saliency recognition results are processed in a hierarchical and classified manner; The final detection results after hierarchical classification are output.

2. The defect detection method according to claim 1, characterized in that: After the step of performing multi-level saliency recognition on the defect area of ​​the reflection fringe image according to the gradient data and the preset saliency threshold to generate a multi-level saliency recognition result, and before the step of performing hierarchical classification processing on the multi-level saliency recognition result, the defect detection method further includes: Get the enhancement coefficient of low-saliency areas; Adaptively enhance the area corresponding to the low-saliency detection result according to the enhancement coefficient.

3. The defect detection method according to claim 2, characterized in that: The step of obtaining the enhancement coefficient of the low-saliency region comprises: An enhancement coefficient of a low-saliency region is generated according to the medium-saliency detection result and the low-saliency detection result.

4. The defect detection method according to claim 1, characterized in that: The step of performing hierarchical classification processing on the multi-level saliency recognition results comprises: Performing saliency grading processing on defects in the multi-level saliency recognition results; Analyzing geometric characteristics of defects in the multi-level saliency recognition results to generate shape factor data; Defect classification is performed based on the grading results and shape factor data.

5. The defect detection method according to any one of claims 1 to 4, characterized in that: The steps of obtaining the reflection fringe image of the display screen to be tested include: A light source configuration is preset, wherein the light source is used to project sine and cosine lights of different phases; Setting a reflector group, wherein the reflector group is used to adjust the light path; The time-sharing strobe function is used to project sine and cosine lights of different phases onto the display screen to be tested; The reflected fringe images are captured by a high-resolution camera and digitized for output.

6. The defect detection method according to any one of claims 1 to 4, characterized in that: The steps of performing phase extraction and phase unwrapping processing on the reflection fringe image to generate phase distribution data include: Extract 8 phase images from the collected reflection fringe images, wherein the 8 phase images include 4 phase images in the x direction and 4 phase images in the y direction; The eight phase images were extracted and phase unwrapped using the phase shift method to generate phase distribution data.

7. The defect detection method according to any one of claims 1 to 4, characterized in that: The step of outputting the final detection result after the hierarchical classification process comprises: The multi-level saliency recognition results after hierarchical classification are visualized and statistically analyzed to generate and output the final detection results, which at least include a small local defect image, defect category, defect feature quantity and the original image.

8. The defect detection method according to any one of claims 1 to 4, characterized in that: The display screen to be tested is a display screen having an internal circuit arranged at the pixel layer; Before the step of performing multi-level saliency recognition on the defect area of ​​the reflection fringe image according to the gradient data and the preset saliency threshold to generate a multi-level saliency recognition result, the defect detection method further includes: Lighting up the spherical integrating light source with a preset brightness according to the average thickness value of the internal circuit of the display screen to be tested and the average thickness of the display screen to be tested; The back side of the display screen to be tested is illuminated by the spherical integrating light source, and the display screen to be tested is photographed from the front side by using a collection device to generate a first photographed image; Lighting up the display screen to be tested so that the display screen to be tested displays a preset grayscale image, and acquiring a second captured image of the display screen to be tested through a collection device; performing binarization processing on the first captured image to determine an internal circuit area on the first captured image; determining the internal circuit area on the second captured image by using the internal circuit area on the first captured image; Determine a first grayscale mean value of an internal circuit area on the second captured image and a second grayscale mean value of a non-circuit area on the second captured image; A preset significance threshold is obtained, and the preset significance threshold is adjusted by using the first grayscale mean value and the second grayscale mean value to generate a significance threshold of the internal circuit area.

9. A defect detection device based on phase deflectometry, characterized in that: include: An acquisition unit, used for acquiring a reflection fringe image of the display screen to be tested; A first generating unit is used to perform phase extraction and phase unwrapping processing on the reflection fringe image to generate phase distribution data; An extraction unit, used for extracting gradient data of the phase distribution data; a second generating unit, configured to perform multi-level saliency recognition on the defect area of ​​the reflection fringe image according to the gradient data and a preset saliency threshold, and generate a multi-level saliency recognition result, wherein the multi-level saliency recognition result includes a high saliency detection result, a medium saliency detection result, and a low saliency detection result, and different levels of saliency recognition results correspond to different types of defects; A processing unit, used for classifying and processing the multi-level saliency recognition results; The output unit is used to output the final detection result after hierarchical classification processing.

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 performed.

Images