Automatic testing methods, systems and storage media for LCD modules

Abstract

This application relates to the field of automated inspection, providing an automated inspection method, system, and storage medium for liquid crystal modules. The automated inspection method for liquid crystal modules can determine whether there are abnormal positions in the liquid crystal module based on a first image captured when the liquid crystal module displays a preset first pattern. If an abnormal position is determined, it further determines whether there is dirt at the abnormal position based on an initial image captured when the liquid crystal module is in its initial state. If dirt is present at the abnormal position, it is designated as the first dirt position. By combining the initial image captured when the liquid crystal module is in its initial state and the first image captured when the liquid crystal module displays the preset first pattern, the liquid crystal module can be inspected and the location of dirt on the liquid crystal module can be determined without cleaning the liquid crystal module during the inspection process. This simplifies the automated inspection process and improves the efficiency of automated liquid crystal module inspection.

Description

Technical Field

[0001] This application relates to the field of automated testing, and in particular to automated testing methods, systems and storage media for LCD modules. Background Technology

[0002] With the continuous development of automated testing technology, traditional LCD module manufacturers have also introduced automated testing equipment, which has greatly improved work efficiency, but also faces some problems. For example, the surface of the LCD module may have very small dust particles or debris attached to it, making it difficult for the testing system to distinguish between external interference and actual internal defects.

[0003] To address the aforementioned issues, a product appearance inspection method has been proposed in related technologies. This method employs a process of inspection, cleaning, and re-inspection, comparing images before and after cleaning to distinguish between interfering objects and genuine defects.

[0004] However, the technical solutions in the relevant technologies require cleaning the product during the testing process, which leads to the complexity of the testing system and the cumbersome operation, thus affecting the testing efficiency of the LCD module. Summary of the Invention

[0005] To help improve the detection efficiency of LCD modules, this application provides an automatic detection method, system, and storage medium for LCD modules.

[0006] Firstly, this application provides an automatic detection method for liquid crystal modules, employing the following technical solution: An automatic detection method for liquid crystal modules, the method comprising: In response to the detection command, acquire the initial image of the LCD module under test in its initial state; Control the liquid crystal module to display a preset first pattern, and acquire a first image captured when the liquid crystal module displays the preset first pattern; Based on the first image, determine whether there is an abnormal position in the liquid crystal module; If an abnormal location is determined in the liquid crystal module, it is determined whether there is dirt at the abnormal location based on the initial image; If dirt is found at the abnormal location based on the initial image, the abnormal location is identified as the first dirty location.

[0007] By adopting the above technical solution, the detection of the liquid crystal module can be achieved by combining the initial image captured in the initial state of the liquid crystal module and the first image captured when the liquid crystal module displays the preset first pattern, and the location of dirt on the liquid crystal module can be determined without cleaning the liquid crystal module during the detection process. This can help simplify the automatic detection process of the liquid crystal module and thus help improve the efficiency of automatic detection of the liquid crystal module.

[0008] Optionally, after determining whether there is dirt at the abnormal location based on the initial image when an abnormal location is found in the liquid crystal module, the method further includes: If, based on the initial image, it is determined that there is no dirt at the abnormal location, the liquid crystal module is controlled to display a preset second pattern, and a second image is acquired when the liquid crystal module displays the preset second pattern. The preset second pattern is different from the preset first pattern. The presence of defects at the abnormal location is verified based on the second image. If a defect is determined to exist at the abnormal location based on the second image, the abnormal location is identified as a defective location.

[0009] By adopting the above technical solution, when it is determined from the initial image that there is no dirt at the abnormal position, the second image acquired by the liquid crystal module when displaying the preset second pattern is used to verify whether there are defects at the abnormal position. This can help avoid misjudgment of defects caused by errors in the first image acquisition process, and thus help improve the accuracy of defect judgment.

[0010] Optionally, if the defect is determined to exist at the abnormal location based on the second image, the method further includes: Extract the first abnormal region image corresponding to the abnormal location from the first image; Extract the second abnormal region image corresponding to the abnormal location from the second image; The first abnormal region image and the second abnormal region image are input into the defect classification model to obtain the defect type corresponding to the abnormal location.

[0011] By adopting the above technical solution, when a defect is determined to exist at an abnormal location based on the second image, the type of defect corresponding to the abnormal location can be further determined by combining the first abnormal area image and the second abnormal area image corresponding to the abnormal location. This can help with subsequent processing of the defective display components and also help monitor screen quality.

[0012] Optionally, after verifying whether there are defects at the abnormal location based on the second image, the method further includes: If it is determined from the second image that there are no defects at the abnormal position, the liquid crystal module is controlled to display a preset third pattern, and a third image is acquired when the liquid crystal module displays the preset third pattern. The preset third pattern is different from the preset first pattern and the preset second pattern. The presence of dirt at the abnormal location is verified based on the third image. If dirt is found at the abnormal location based on the third image, the abnormal location is identified as the second dirty location.

[0013] By adopting the above technical solution, when it is determined from the second image that there are no defects at the abnormal position, the presence of dirt at the abnormal position can be checked from the third image acquired when the LCD module displays the preset third pattern. This can help avoid the influence of errors in the initial image acquisition process and the limitations of the content reflected by the initial image on the judgment of the abnormal position, and thus help to accurately determine the abnormal position.

[0014] Optionally, after verifying whether there is dirt at the abnormal location based on the third image, the method further includes: If it is determined from the third image that there is no abnormality at the abnormal location, the abnormal location is determined as a location to be determined. Extract the image of the undetermined position corresponding to the undetermined position from the initial image, the first image, the second image, and the third image respectively; A request for manual judgment is generated and output based on each of the images of the undetermined locations.

[0015] By adopting the above technical solution, it is easier to detect LCD modules with manual assistance, which can help to handle difficult judgment situations and thus help to ensure the efficiency of automatic detection.

[0016] Optionally, before controlling the liquid crystal module to display the preset second pattern, the method further includes: Anomaly analysis is performed on the abnormal location to obtain the abnormal parameters corresponding to the abnormal location; The preset second pattern is determined based on the abnormal parameters, and the correspondence between the abnormal parameters and the preset second image is preset.

[0017] By adopting the above technical solution, a preset second image can be determined based on the actual situation corresponding to the abnormal location. This can help improve the relevance of the determined second image, thereby improving the verification accuracy, while also helping to reduce detection costs and improve detection efficiency.

[0018] Optionally, the preset first pattern includes at least two, and the step of controlling the liquid crystal module to display the preset first pattern and acquiring a first image captured when the liquid crystal module displays the preset first pattern includes: The liquid crystal module is controlled to sequentially display each of the preset first patterns, and the first image of each preset first pattern displayed by the liquid crystal module is acquired respectively; The step of determining whether there is an abnormal position in the liquid crystal module based on the first image includes: For each of the preset first patterns, the presence of the abnormal position in the liquid crystal module is determined based on the first image corresponding to the preset first pattern.

[0019] By adopting the above technical solution, the limitations of using a single preset first pattern for abnormal position detection can be avoided, thereby enabling comprehensive detection of abnormal positions of the LCD module under different display states, which in turn helps to improve the accuracy of the detection results.

[0020] Optionally, the first pattern includes a full red pattern, a full blue pattern, and a full green pattern. The step of determining whether there is dirt at the abnormal location based on the initial image when an abnormal location is determined to exist in the liquid crystal module includes: Determine whether the abnormal positions identified based on the first images corresponding to the all-red pattern, the all-blue pattern, and the all-green pattern overlap; If there is overlap, determine whether there is dirt at the overlapping abnormal location based on the initial image.

[0021] By adopting the above technical solution, when the abnormal positions determined based on the first image corresponding to the all-red pattern, all-blue pattern and all-green pattern overlap, the initial image is combined to determine whether there is dirt at the overlapping abnormal positions. This can help to accurately determine the cause of the abnormality, and at the same time help to avoid unnecessary dirt judgment and improve detection efficiency.

[0022] Secondly, this application provides an automatic detection system for LCD modules, which adopts the following technical solution: An automatic detection system for LCD modules, the system comprising a carrier, an image acquisition component, and an industrial control computer, wherein the image acquisition component is signal-connected to the controller; The carrier is used to hold the liquid crystal module; The image acquisition component is used to acquire images of the liquid crystal module placed on the vehicle and transmit them to the industrial control computer; The industrial control computer is used to execute any of the automatic detection methods for LCD modules provided in the first aspect.

[0023] Thirdly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform any of the automatic detection methods for liquid crystal modules provided in the first aspect.

[0024] In summary, this application includes at least one of the following beneficial technical effects: 1. Since the detection of the LCD module and the determination of the location of dirt on the LCD module can be achieved by combining the initial image captured in the initial state of the LCD module and the first image captured when the LCD module displays the preset first pattern, the LCD module does not need to be cleaned during the detection process. This can help simplify the automatic detection process of the LCD module and thus help improve the efficiency of automatic detection of the LCD module.

[0025] 2. If the abnormal location is determined to be free of dirt based on the initial image, the presence of defects at the abnormal location is verified by using the second image acquired by the LCD module when displaying the preset second pattern. This helps to avoid misjudgment of defects caused by errors in the first image acquisition process, thereby improving the accuracy of defect judgment. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the structure of an automatic detection system for liquid crystal modules provided in an embodiment of this application; Figure 2 This is a flowchart illustrating an automatic detection method for a liquid crystal module provided in an embodiment of this application; Figure 3 This is a flowchart illustrating a defect verification method provided in an embodiment of this application; Figure 4 This is a flowchart illustrating a defect classification method provided in an embodiment of this application; Figure 5 This is a flowchart illustrating a method for verifying dirt and grime according to an embodiment of this application; Figure 6 This is a flowchart illustrating a manual judgment request generation method provided in an embodiment of this application; Figure 7 This is a flowchart illustrating a first image acquisition method provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0027] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figure 1-8 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0028] To help understand the automatic detection method for liquid crystal modules provided in the embodiments of this application, the automatic detection system for liquid crystal modules provided in the embodiments of this application will be introduced first below.

[0029] refer to Figure 1 The automatic detection system for LCD modules provided in this application includes: a carrier (not shown in the figure), an image acquisition component 110 and an industrial control computer 120, wherein the acquisition component 110 is connected to the industrial control computer 120 by signal.

[0030] The carrier is used to place the LCD module. In one example, the carrier is a stage. Optionally, the stage is equipped with a limiter to limit the LCD module, which helps to acquire images from the LCD module placed on the carrier.

[0031] Image acquisition component 110 is used to acquire images of the LCD module placed on the carrier and transmit them to the industrial control computer 120. In one example, image acquisition component 110 is an industrial camera. Optionally, during image acquisition, a limiting member set on the carrier can limit the position of the LCD module within the image acquisition range of the image component, which can help to achieve image acquisition of the LCD module.

[0032] In actual implementation, the system may include two or more carriers. Different carriers can be switched using a switching device (such as a conveyor belt, turntable, etc.). In this way, while the image acquisition component 110 is acquiring images of the liquid crystal module placed on one carrier, the liquid crystal module placed on other carriers can be switched. This can help reduce the idle time of the image acquisition component 110 and thus improve the detection efficiency.

[0033] The industrial computer 120 is used to detect the LCD module based on the image acquired by the image acquisition component 110.

[0034] In one example, the industrial computer 120 is also connected to the LCD module placed on the carrier, which facilitates the control of the content displayed on the LCD module during the testing process, thereby helping to test the display effect of the LCD module.

[0035] In actual implementation, the automatic detection system for LCD modules may also include other mechanisms, such as conveying mechanisms and loading / unloading mechanisms. This embodiment does not limit the specific composition of the automatic detection system for LCD modules.

[0036] This application also discloses an automatic detection method for liquid crystal modules. This embodiment uses the application of this method to the controller of the automatic detection system for liquid crystal modules provided in the above embodiment as an example for illustration. In actual implementation, this method can also be applied to other devices used for detecting liquid crystal modules.

[0037] refer to Figure 2 The automatic detection method for LCD modules includes the following steps: Step 201: In response to the detection command, acquire the initial image of the liquid crystal module to be detected in its initial state.

[0038] The detection command can be automatically generated, such as being generated at regular intervals, or being generated when the LCD module to be detected switches, or it can be generated based on user operation or signals sent by other devices. This embodiment does not limit the generation method of the detection command.

[0039] Optionally, the LCD module to be tested is an LCD module placed on the stage of the LCD module automatic detection system, and the images of the LCD module in different states are acquired by the image acquisition component of the LCD module automatic detection system.

[0040] It should be noted that the original image acquired by the image acquisition component may include images of both the area corresponding to the liquid crystal module and the surrounding area. In this case, to avoid interference from images of areas outside the liquid crystal module's area, preprocessing of the original image acquired by the image acquisition component is necessary to extract the image of the area corresponding to the liquid crystal module. In one example, a mask image can be pre-set based on the location of the liquid crystal module's area in the original image acquired by the image acquisition component. This allows extraction of the image of the liquid crystal module's area from the original image acquired by the image acquisition component using the mask image. In practical implementations, other methods can also be used to preprocess the image acquired by the image acquisition component, such as binarizing the original image.

[0041] Optionally, the initial state is the state when the LCD module is not powered on. In this case, in response to the detection command, the image information of the LCD module acquired by the image acquisition component is obtained even when the LCD module is not powered on, thus obtaining the initial image.

[0042] In one example, after obtaining the initial image, the process also includes: saving the initial image and performing step 202.

[0043] Step 202: Control the LCD module to display a preset first pattern, and acquire the first image captured when the LCD module displays the preset first pattern.

[0044] Optionally, acquiring the first image of the LCD module when displaying a preset first pattern includes: after controlling the LCD module to display the preset first pattern, sending an image acquisition command to an image acquisition device; and determining the first image based on the original image returned by the image acquisition device.

[0045] In actual implementation, the image acquisition device automatically acquires images for a preset time period. At this time, the acquisition of the first image when the liquid crystal module displays the preset first pattern includes: determining the first image based on the original image returned by the image acquisition device after controlling the liquid crystal module to display the preset first pattern.

[0046] Step 203: Determine whether there are any abnormal positions in the liquid crystal module based on the first image.

[0047] The abnormal location refers to an abnormal location in the liquid crystal module indicated by the first image.

[0048] Specifically, abnormal locations can be caused by defects in the LCD module or by dirt on the surface of the LCD module.

[0049] Defects refer to display abnormalities in LCD modules, such as dead pixels and light leakage. Defects are usually caused by abnormalities in the LCD module manufacturing process and are generally difficult to repair.

[0050] Dirt refers to external interference on the surface of the LCD module, such as small dust particles and debris. Generally speaking, dirt can be removed by cleaning the LCD module.

[0051] Since defects and dirt can both affect the display performance of an LCD module, anomaly detection can be achieved by capturing images of the LCD module. Furthermore, because the impact of defects and dirt on the display performance of an LCD module may vary under different display scenarios, the cause of the anomaly can be determined by combining images of the LCD module captured under different display scenarios.

[0052] In one example, the first pattern is preset to be a solid color pattern, that is, the colors of all positions in the first pattern are the same. In this case, determining whether there are abnormal positions in the liquid crystal module based on the first image includes: determining whether there are positions in the first image with a color difference greater than a preset error threshold with other positions; if so, the positions in the first image with a color difference greater than the preset error threshold with other positions are determined to be abnormal positions; if not, it is determined that there are no abnormal positions in the liquid crystal module.

[0053] In one instance, the first pattern includes a pure red pattern (rgb(255,0,0)), a pure blue pattern (rgb(0,0,255)), and / or a pure green pattern (rgb(0,255,0)).

[0054] It should be noted that in actual implementation, the colors in the first image are usually represented by RGB values. In order to facilitate color comparison, the first image can be converted into the corresponding grayscale image. At this time, the difference in grayscale values ​​at different positions is the color difference at different positions. The preset error threshold can also be set based on the maximum allowable difference in grayscale values.

[0055] In one instance, the first image and the preset first pattern are converted into corresponding grayscale images using a Gamma correction algorithm.

[0056] In another example, determining whether there is an abnormal position in the liquid crystal module based on the first image includes: comparing the first image with a preset first pattern to determine whether there is a difference between the first image and the preset first pattern; if so, the difference position between the first image and the preset first pattern is determined as an abnormal position; if not, it is determined that there is no abnormal position in the liquid crystal module.

[0057] The method of comparing the first image with the preset first pattern to determine whether there is a difference between the first image and the preset first pattern can be to determine whether there is a difference position between the first image and the preset first pattern by using a pre-set image comparison model, or it can be to determine whether there is a difference position in the first image where the color difference is greater than a preset error threshold compared with the corresponding position in the preset first pattern.

[0058] Optionally, if there are no abnormal locations, the LCD module can be directly judged as a qualified product. Further, the initial image can be used to determine whether the LCD module needs to be cleaned.

[0059] Step 204: If an abnormal position is found in the liquid crystal module, determine whether there is dirt or grime at the abnormal position based on the initial image.

[0060] In one example, the method for determining whether there is dirt at an abnormal location based on the initial image is the same as the method for determining whether there is an abnormal location in the liquid crystal module based on the first image in step 203 above.

[0061] In another example, determining whether there is dirt at an abnormal location based on the initial image includes: obtaining a binary edge image corresponding to the initial image using a preset edge detection method; and determining the location of the dirt based on the contour in the binary edge image.

[0062] The edge detection algorithm may include at least one of the following operators: Canny operator, Prewitt operator, Kirsch operator, etc.

[0063] Since the initial image is captured in the initial state of the LCD module, and the LCD module is not displaying an image at this time, the defects of the LCD module will not affect the capture of the initial image. Therefore, the abnormal location detected in the initial image is the location of the dirt.

[0064] Optionally, there may be multiple abnormal locations. In this case, each abnormal location needs to be judged separately, that is, it is necessary to determine whether there is dirt at the abnormal location based on the initial image. This includes: determining whether there is dirt at each abnormal location based on the initial image, and performing corresponding processing steps according to the judgment results.

[0065] Step 205: If dirt is found at an abnormal location based on the initial image, the abnormal location is identified as the first dirty location.

[0066] Since the location of dirt can affect the display effect of the LCD module, the location of the abnormal dirt is determined when the abnormal location determined based on the first image coincides with the location of the dirt determined based on the initial image. In this way, the abnormal location and abnormal type can be determined.

[0067] In addition, since the judgment is made based on the initial image after the first image is used, it can help avoid the problem of misjudgment caused by the error in the acquisition of the initial image when judging the location of dirt. This can help improve the accuracy of dirt location judgment.

[0068] In practice, the handling of the first dirty location includes cleaning the LCD module and re-inspecting it, i.e., repeating steps 201 to 203. This helps to avoid the problem of missed defects when the dirty and defect locations overlap.

[0069] In one example, after identifying the abnormal location as the first dirty location, the process also includes recording the coordinates corresponding to the first dirty location.

[0070] The implementation principle of an automatic detection method for liquid crystal modules according to an embodiment of this application is as follows: In response to a detection command, an initial image of the liquid crystal module to be detected in its initial state is acquired; the liquid crystal module is controlled to display a preset first pattern, and a first image is acquired when the liquid crystal module displays the preset first pattern; based on the first image, it is determined whether there is an abnormal position in the liquid crystal module; if an abnormal position is determined, it is determined whether there is dirt at the abnormal position based on the initial image; if dirt is determined to be present at the abnormal position based on the initial image, the abnormal position is identified as the first dirt position. In the above technical solution, since the detection of the liquid crystal module and the determination of the dirt position on the liquid crystal module can be achieved by combining the initial image acquired in the initial state of the liquid crystal module and the first image acquired when the liquid crystal module displays the preset first pattern, without the need to clean the liquid crystal module during the detection process, this simplifies the automatic detection process of the liquid crystal module and thus improves the efficiency of automatic detection of the liquid crystal module.

[0071] In some implementations, reference Figure 3 Optionally, in step 204, after determining whether there is dirt at the abnormal location based on the initial image, if an abnormal location is found to exist in the liquid crystal module, the following steps are also included: Step 301: If it is determined from the initial image that there is no dirt at the abnormal location, control the LCD module to display a preset second pattern, and acquire the second image captured when the LCD module displays the preset second pattern.

[0072] The preset second pattern is different from the preset first pattern.

[0073] In one example, the second pattern is preset to a solid color pattern, meaning that all positions of the pattern are the same color. In one instance, the second pattern is preset to solid cyan (rgb(0,255,255)) and / or solid purple (rgb(255,0,255)).

[0074] In another example, the second image is preset to a special graphic, such as: horizontal lines, vertical lines, diagonal lines, dot matrix, grid, etc.

[0075] In this embodiment, unless otherwise specified above, the implementation of step 301 is the same as step 202 above, and will not be repeated here.

[0076] Optionally, before controlling the LCD module to display the preset second pattern, the method further includes: performing anomaly analysis on the abnormal position to obtain the abnormal parameters corresponding to the abnormal position; determining the preset second pattern based on the abnormal parameters, wherein the correspondence between the abnormal parameters and the preset second image is preset.

[0077] In one example, the anomaly parameter is determined based on the region to which the anomaly location belongs. For instance, the display module is divided into three regions: left, center, and right. The anomaly parameter is the region identifier of the region to which the anomaly location belongs. Correspondingly, the preset second image is set based on the region identifier. Different region identifiers correspond to different second region images. This can help to perform targeted detection of the region to which the anomaly location belongs, thereby improving the verification accuracy.

[0078] In another example, the preset first pattern includes at least two. In this case, the anomaly parameter is determined based on the preset first pattern at the detected anomaly location. For example, the anomaly parameter is the pattern identifier of the preset first pattern corresponding to the detected anomaly location. Correspondingly, the preset second image is determined based on the pattern identifier. The second region images corresponding to different pattern identifiers are different. This can help to combine the situation of the detected anomaly location to further judge the anomaly location, thereby helping to improve the verification accuracy.

[0079] In the above technical solution, since the preset second image is determined based on the abnormal parameters obtained by analyzing the abnormal position, the preset second image can be determined in combination with the actual situation corresponding to the abnormal position, which can help improve the relevance of the determined second image, thereby helping to improve the verification accuracy, and also helping to reduce the detection cost and improve the detection efficiency.

[0080] Step 302: Verify whether there are defects at abnormal locations based on the second image.

[0081] Optionally, the verification of whether there are defects at the abnormal location is based on the second image includes: determining whether there are defects at the abnormal location based on the second image, and obtaining the verification result.

[0082] Specifically, the method for determining whether there are defects at abnormal locations based on the second image is the same as the method for determining whether there are abnormal locations in the liquid crystal module based on the first image in step 203 above, and will not be repeated here in this embodiment.

[0083] Step 303: If a defect is found at an abnormal location based on the second image, the abnormal location is identified as the defect location.

[0084] In the above embodiments, since it is determined that there is no dirt at the abnormal position based on the initial image, the LCD module is further controlled to display a preset second pattern, and the presence of defects at the abnormal position is verified based on the second image acquired when the LCD module displays the preset second pattern. If a defect is determined to exist at the abnormal position based on the second image, the abnormal position is identified as the defect position. This can help avoid misjudgment of defects caused by errors in the first image acquisition process, thereby helping to improve the accuracy of defect judgment.

[0085] In some implementations, reference Figure 4 Optionally, in step 303, if a defect is determined to exist at an abnormal location based on the second image, the following steps are also included: Step 401: Extract the first abnormal region image corresponding to the abnormal location from the first image.

[0086] Optionally, the first abnormal region image corresponding to the abnormal location is extracted from the first image, including: taking the undetermined location as the center, extracting an image region of a preset shape and preset size from the first image to obtain the first abnormal region image.

[0087] Step 402: Extract the second abnormal region image corresponding to the abnormal location from the second image.

[0088] Optionally, the second abnormal region image corresponding to the abnormal location is extracted from the second image, including: taking the undetermined location as the center, extracting an image region of a preset shape and preset size from the second image to obtain the second abnormal region image.

[0089] In one example, the first and second anomalous region images are the same size. This facilitates the analysis and processing of both images.

[0090] Step 403: Input the first abnormal region image and the second abnormal region image into the defect classification model to obtain the defect type corresponding to the abnormal location.

[0091] The defect classification model is pre-set.

[0092] In one example, the defect classification model includes a first judgment sub-model, a second judgment sub-model, and a defect determination model. In this case, the first abnormal region image and the second abnormal region image are input into the defect classification model to obtain the defect type corresponding to the abnormal location. This includes: inputting the first abnormal region image into the first judgment sub-model to obtain a first score distribution; inputting the second abnormal region image into the second judgment sub-model to obtain a second score distribution; and inputting the first score distribution and the second score distribution into the defect determination model to obtain the defect type.

[0093] The first score distribution is used to indicate the probability of each preset defect type in the first abnormal region image; the second score distribution is used to indicate the probability of each preset defect type in the second abnormal region image.

[0094] The first and second judgment sub-models are obtained by training the initial neural network using training data. Each set of training data includes sample images and the types of defects contained in the sample images. Optionally, the training data for the first and second judgment sub-models are different. Specifically, the sample images for the first judgment sub-model are collected when the display component displays a preset first pattern, and the sample images for the second judgment sub-model are collected when the display component displays a preset second pattern.

[0095] Schematic, the training process of the first judgment sub-model and the second judgment sub-model includes: creating an initial network model; inputting sample images into the initial network model to obtain model results; iteratively updating the parameters of the initial network model based on the model results and the types of defects contained in the sample images; and obtaining the first judgment sub-model when the number of iterations reaches a preset number or the updated model converges.

[0096] Optionally, the defect type is used to determine the cumulative score corresponding to each preset defect type based on the first score distribution model and the second score distribution model, and to determine the defect type corresponding to the abnormal region based on the cumulative score of each preset defect type. For example, the preset defect type with the highest cumulative score is determined as the defect type corresponding to the abnormal region.

[0097] In one instance, determining the cumulative score corresponding to each defect type includes: for each preset defect type, determining the cumulative score corresponding to the defect type as the weighted sum of the scores corresponding to the defect type in the first score distribution and the scores corresponding to the defect type in the second score distribution.

[0098] The weights of the scores in the first score distribution and the scores in the second score distribution are preset. In one instance, the weights of the scores in the first score distribution and the scores in the second score distribution are the same.

[0099] In another example, the first and second abnormal region images are input into the defect classification model to obtain the defect type corresponding to the abnormal location. This includes: fusing the first and second abnormal region images to obtain the input image; and inputting the input image into the defect classification model to obtain the defect classification result.

[0100] The defect classification model is a neural network model. The specific training method for the defect classification model is the same as the training method of the first judgment sub-model mentioned above, and will not be repeated here.

[0101] Optionally, fusing the first abnormal region image and the second abnormal region image to obtain the input image includes: superimposing the first abnormal region image and the second abnormal region image pixels to obtain the fused image.

[0102] In the above embodiments, since the defect is determined to exist at the abnormal position based on the second image, the defect type corresponding to the abnormal position is further determined by combining the first abnormal region image corresponding to the abnormal position cropped from the first image and the second abnormal region image corresponding to the abnormal position cropped from the second image. This can help to process the display component with defects in the subsequent processing, and can also help to monitor the screen quality.

[0103] In some implementations, reference Figure 5 Optionally, after verifying whether there are defects at abnormal locations based on the second image in step 302, the following steps are also included: Step 501: If it is determined from the second image that there are no defects at the abnormal position, control the LCD module to display a preset third pattern, and acquire the third image captured when the LCD module displays the preset third pattern.

[0104] The preset third pattern is different from the preset first pattern and the preset second pattern.

[0105] In one example, the third pattern is preset to a solid color pattern, meaning that all positions of the pattern are the same color. In another example, the third pattern is preset to a completely black pattern (rgb(0,0,0)) or a completely white pattern (rgb(255,255,255)).

[0106] In actual implementation, the preset third pattern can also be set in other ways, such as setting it as a special graphic. This embodiment does not limit the setting method of the preset third pattern.

[0107] In this embodiment, unless otherwise specified above, the implementation of step 401 is the same as that of steps 202 and 301 above, and will not be repeated here.

[0108] Step 502: Verify whether there is dirt at the abnormal location based on the third image.

[0109] Optionally, the verification of whether there is dirt at the abnormal location is based on the third image includes: determining whether there is dirt at the abnormal location based on the third image, and obtaining the verification result.

[0110] Specifically, the method for determining whether there is dirt at an abnormal location based on the third image is the same as the method for determining whether there is dirt at an abnormal location based on the initial image in step 204 above, and will not be repeated here in this embodiment.

[0111] Step 503: If dirt is found at an abnormal location based on the third image, the abnormal location is identified as the second dirty location.

[0112] In the above embodiments, when it is determined that there are no defects at the abnormal position based on the second image, the LCD module is controlled to display a preset third image. The presence of dirt at the abnormal position is checked based on the third image acquired when the LCD module displays the preset third image. If dirt is found at the abnormal position based on the third image, the abnormal position is identified as the second dirty position. This helps to avoid the influence of errors in the initial image acquisition process and the limitations of the content reflected by the initial image on the judgment of the abnormal position, thereby helping to accurately determine the abnormal position.

[0113] Based on the above technical solution, further reference... Figure 6 After step 402 above, which verifies whether there is dirt at the abnormal location based on the third image, the following steps are also included: Step 601: If it is determined from the third image that there is no abnormality at the abnormal location, the abnormal location is determined as a location to be determined.

[0114] Step 602: Extract the image of the position to be determined from the initial image, the first image, the second image, and the third image.

[0115] In one example, the method for cropping the image at the unknown location includes: cropping an image region of a preset shape and size centered on the unknown location to obtain the image at the unknown location.

[0116] Step 603: Generate and output manual judgment requests based on the images of each pending location.

[0117] In one example, the output of a manual judgment request includes sending a manual judgment request to the administrator device.

[0118] Furthermore, after outputting the manual judgment request, the method also includes: responding to the manual judgment result; if the manual judgment result indicates that there is no abnormality at the pending location, correcting the preset first pattern and / or the abnormal location judgment method; if the manual judgment result indicates that the pending location is a dirty location, correcting the preset third image and / or the dirty location judgment method; if the manual judgment result only indicates that the pending location is a defective location, correcting the preset second image and / or the defective location judgment method. This helps to correct the automatic detection process through manual assistance, thereby improving the success rate of the automatic detection process.

[0119] In the above technical solution, since it is still impossible to determine the location of dirt based on the third image, a manual judgment request is generated and output based on the images of the undetermined location extracted from the initial image, the first image, the second image and the third image, respectively. This facilitates the detection of the liquid crystal module with human assistance, which helps to handle difficult judgment situations and thus helps to ensure the efficiency of automatic detection.

[0120] In some implementations, reference Figure 7 The preset first pattern includes at least two elements. Step 202 involves controlling the LCD module to display the preset first pattern and acquiring a first image captured when the LCD module displays the preset first pattern, including the following steps: Step 701: Control the LCD module to display each preset first pattern in sequence, and acquire the first image of each preset first pattern displayed by the LCD module.

[0121] In one example, the first pattern includes a pure red pattern (rgb(255,0,0)), a pure blue pattern (rgb(0,0,255)), and a pure green pattern (rgb(0,255,0)).

[0122] Accordingly, step 203, determining whether there are abnormal positions in the liquid crystal module based on the first image, includes the following steps: Step 702: For each preset first pattern, determine whether there is an abnormal position in the liquid crystal module based on the first image corresponding to the preset first pattern.

[0123] In one example, the first pattern includes a full red pattern, a full blue pattern, and a full green pattern. When an abnormal position is determined to exist in the liquid crystal module, it is determined whether there is dirt at the abnormal position based on the initial image, including: determining whether the abnormal positions determined based on the first image corresponding to the full red pattern, the full blue pattern, and the full green pattern overlap; if there is overlap, it is determined whether there is dirt at the overlapping abnormal position based on the initial image.

[0124] For locations where there is no overlap, dirt determination is not performed based on the initial image.

[0125] Since dirt will affect image acquisition regardless of the image displayed by the LCD module when there is dirt at abnormal locations, and this dirt will ultimately be reflected in the acquired image, the above technical solution only determines whether there is dirt at the overlapping abnormal locations when the abnormal locations determined based on the first images corresponding to the all-red, all-blue, and all-green patterns overlap. This helps to improve the accuracy of abnormality cause determination and also helps to avoid unnecessary dirt determination, thereby improving detection efficiency.

[0126] In addition, since all-red, all-blue, and all-green patterns are used as preset first patterns for abnormal location judgment, this can help to initially distinguish the defective and dirty locations of the LCD module in terms of color display, thereby helping to determine different abnormal verification methods and thus helping to accurately determine the cause of the abnormality.

[0127] In actual implementation, if an abnormal position is detected in the first image corresponding to any preset first pattern, the initial image can be used to determine whether the abnormal position is dirty.

[0128] In the above embodiments, since the preset first pattern includes at least two, and the first image of each preset first pattern displayed by the liquid crystal module can be obtained during the anomaly judgment process, and the anomaly position detection is performed on each first image, this can help avoid the limitation of using a single preset first pattern for anomaly position detection, thereby helping to comprehensively detect the anomaly position of the liquid crystal module in different display states, and thus helping to improve the accuracy of the detection results.

[0129] This application also provides an electronic device. In one example, the electronic device is the one described above. Figure 1 The industrial control computer 110 of the automatic detection system for LCD modules shown can be implemented as other devices in practice; this embodiment does not limit the type of electronic device. For example... Figure 8 As shown, Figure 8 The illustrated electronic device 800 includes a processor 801 and a memory 803. The processor 801 and the memory 803 are connected, for example, via a bus 802. Optionally, the electronic device 800 may also include a transceiver 804. It should be noted that in practical applications, the transceiver 804 is not limited to one type, and the structure of this electronic device 800 does not constitute a limitation on the embodiments of this application.

[0130] Processor 801 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 801 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0131] Bus 802 may include a pathway for transmitting information between the aforementioned components. Bus 802 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 802 can be divided into address bus, data bus, etc. For ease of representation, Figure 8 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0132] The memory 803 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0133] The memory 803 stores application code that executes the scheme of this application, and its execution is controlled by the processor 801. The processor 801 executes the application code stored in the memory 803 to implement the content shown in the foregoing method embodiments.

[0134] Electronic devices include, but are not limited to: mobile terminals such as mobile phones, laptops, PDAs (personal digital assistants), and PADs (tablet computers), as well as fixed terminals such as digital TVs and desktop computers. They can also serve as server-side components. Figure 8 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0135] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed in a computer, causes the computer to perform the automatic detection method for liquid crystal modules provided in the above embodiments.

[0136] It should be understood that although the steps in the flowcharts in the accompanying drawings are shown sequentially as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order in which these steps are performed, and they may be performed in other orders.

[0137] The above are only some embodiments of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An automatic detection method for liquid crystal modules, characterized in that, The method includes: In response to the detection command, acquire the initial image of the LCD module under test in its initial state; Control the liquid crystal module to display a preset first pattern, and acquire a first image captured when the liquid crystal module displays the preset first pattern; Based on the first image, determine whether there is an abnormal position in the liquid crystal module; If an abnormal location is determined in the liquid crystal module, it is determined whether there is dirt at the abnormal location based on the initial image; If dirt is found at the abnormal location based on the initial image, the abnormal location is identified as the first dirty location. If, based on the initial image, it is determined that there is no dirt at the abnormal location, an anomaly analysis is performed on the abnormal location to obtain the anomaly parameters corresponding to the abnormal location. Based on the anomaly parameters, a preset second pattern is determined. The correspondence between the anomaly parameters and the preset second pattern is preset. The anomaly parameters are determined based on the region to which the abnormal location belongs. The anomaly parameters are the region identifier of the region to which the abnormal location belongs. The preset second pattern is set based on the region identifier. The preset second pattern is different from the preset first pattern. The preset second pattern is a solid color pattern or a special graphic. The special graphic includes horizontal lines, vertical lines, diagonal lines, dot matrix, or grid. Control the liquid crystal module to display a preset second pattern, and acquire a second image captured when the liquid crystal module displays the preset second pattern; The presence of defects at the abnormal location is verified based on the second image. If a defect is determined to exist at the abnormal location based on the second image, the abnormal location is identified as a defect location. Extract the first abnormal region image corresponding to the abnormal location from the first image; A second abnormal region image corresponding to the abnormal location is extracted from the second image; the first abnormal region image and the second abnormal region image are image regions of a preset shape and size extracted with the abnormal location as the center, and the first abnormal region image and the second abnormal region image have the same shape and size; The first abnormal region image and the second abnormal region image are input into a defect classification model to obtain the defect type corresponding to the abnormal location. The defect classification model includes a first judgment sub-model, a second judgment sub-model, and a defect determination model. The first judgment sub-model is used to obtain a first score distribution based on the first abnormal region image. The first score distribution is used to indicate the probability that there are defects corresponding to each preset defect type in the first abnormal region image. The second judgment sub-model is used to obtain a second score distribution based on the second abnormal region image. The second score distribution is used to indicate the probability that there are defects corresponding to each preset defect type in the second abnormal region image. The defect determination model is used to determine the cumulative score corresponding to each preset defect type based on the first score distribution model and the second score distribution model, and determine the preset defect type with the highest cumulative score as the defect type corresponding to the abnormal region.

2. The method according to claim 1, characterized in that, After verifying whether there are defects at the abnormal location based on the second image, the method further includes: If it is determined from the second image that there are no defects at the abnormal position, the liquid crystal module is controlled to display a preset third pattern, and a third image is acquired when the liquid crystal module displays the preset third pattern. The preset third pattern is different from the preset first pattern and the preset second pattern. The presence of dirt at the abnormal location is verified based on the third image. If dirt is found at the abnormal location based on the third image, the abnormal location is identified as the second dirty location.

3. The method according to claim 2, characterized in that, After verifying whether there is dirt at the abnormal location based on the third image, the process further includes: If it is determined from the third image that there is no abnormality at the abnormal location, the abnormal location is determined as a location to be determined. Extract the image of the undetermined position corresponding to the undetermined position from the initial image, the first image, the second image, and the third image respectively; A request for manual judgment is generated and output based on each of the images of the undetermined locations.

4. The method according to claim 1, characterized in that, The preset first pattern includes at least two, the abnormal parameter is the pattern identifier of the preset first pattern at the detected abnormal position, the preset second pattern is determined based on the pattern identifier, and the step of controlling the liquid crystal module to display the preset first pattern and acquiring a first image captured when the liquid crystal module displays the preset first pattern includes: The liquid crystal module is controlled to sequentially display each of the preset first patterns, and the first image of each preset first pattern displayed by the liquid crystal module is acquired respectively; The step of determining whether there is an abnormal position in the liquid crystal module based on the first image includes: For each of the preset first patterns, the presence of the abnormal position in the liquid crystal module is determined based on the first image corresponding to the preset first pattern.

5. The method according to claim 4, characterized in that, The first pattern includes a full red pattern, a full blue pattern, and a full green pattern. The step of determining whether there is dirt at the abnormal location based on the initial image when an abnormal location is identified in the liquid crystal module includes: Determine whether the abnormal positions identified based on the first images corresponding to the all-red pattern, the all-blue pattern, and the all-green pattern overlap; If there is overlap, determine whether there is dirt at the overlapping abnormal location based on the initial image.

6. An automatic detection system for liquid crystal modules, characterized in that, The system includes a vehicle, an image acquisition component, and an industrial control computer, wherein the image acquisition component is signal-connected to the industrial control computer. The carrier is used to hold the liquid crystal module; The image acquisition component is used to acquire images of the liquid crystal module placed on the vehicle and transmit them to the industrial control computer; The industrial control computer is used to execute the automatic detection method for liquid crystal modules according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed in the computer, the computer is instructed to perform the automatic detection method for liquid crystal modules according to any one of claims 1 to 5.

Images