Backlight module detection method, electronic device and storage medium

Abstract

The application provides a backlight module detection method, an electronic device and a storage medium, and relates to the technical field of display. The method comprises the following steps: controlling a lamp panel to be detected to be in a lighting state, acquiring lamp panel optical data, inputting the lamp panel optical data into a prediction model, processing the lamp panel optical data through the prediction model, obtaining optical data after coating, performing quality detection processing on the lamp panel optical data, obtaining a lamp panel detection result, performing quality detection processing on the optical data after coating, obtaining a detection result after coating, and obtaining a BLU overall detection result according to the lamp panel detection result and the detection result after coating. Through the scheme, the speed and accuracy of BLU detection can be improved, and the detection cost can be reduced.

Description

Technical Field

[0001] This application relates to the field of display technology, and in particular to a backlight module detection method, electronic device, and storage medium. Background Technology

[0002] The backlight unit (BLU) is one of the key components of an LCD panel. It typically consists of a lamp plate and multiple optical film layers. Its function is to provide sufficient brightness and uniformly distributed light source so that images can be displayed normally. For the BLU as a whole, the most important indicator is the uniformity of illumination.

[0003] Currently, there are generally two methods for checking the uniformity or yield of BLU (Brilliant Light Unit) lamps: visual inspection or equipment testing. Visual inspection is highly subjective, resulting in inaccurate results and low efficiency. Equipment testing, on the other hand, typically requires a lighting test after each layer of film is applied to the lamp panel to determine if there are any lighting defects or if the uniformity meets the requirements. This method is cumbersome and inefficient.

[0004] Therefore, there is an urgent need for an efficient, accurate, and human-free method for detecting BLU. Summary of the Invention

[0005] This application provides a backlight module testing method, electronic device, and storage medium, which solves the problem of cumbersome and inefficient inspection procedures for BLU uniformity or yield in the prior art.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] In a first aspect, this application provides a backlight module inspection method, wherein the backlight module includes a lamp panel and a first optical film layer covering the light-emitting surface of the lamp panel. The method includes: acquiring an image of the lamp panel in a lit state to obtain lamp panel optical data; inputting the lamp panel optical data into a first prediction model corresponding to the first optical film layer to obtain backlight module optical data when the backlight module is in a lit state; and performing quality inspection on the backlight module based on the backlight module optical data to obtain a quality inspection result of the backlight module.

[0008] The backlight module inspection method provided in this application directly obtains the coated optical data through a prediction model by collecting only the optical data of the lamp board. Quality inspection processing is then performed on both the lamp board optical data and the coated optical data to obtain the overall BLU inspection result of the backlight module. This improves the speed and accuracy of BLU inspection and reduces inspection costs.

[0009] In some possible implementations, before inputting the optical data of the lamp panel into the first prediction model corresponding to the first optical film layer, a first sample dataset is obtained. The first sample dataset includes N subsets of first sample data, each subset containing multiple sets of first sample data. Each set of first sample data includes optical data of the lamp panel sample in its illuminated state, and optical data of the lamp panel sample in its illuminated state after the luminous surface is covered with the first optical film layer. Different types of lamp panel samples correspond to different subsets of sample data, while multiple lamp panel samples within the same subset have the same type. A first generative network model is obtained by training the model based on the first sample dataset. The first generative network model is iteratively trained using a discriminative network model, and the resulting second generative network model is used as the first prediction model. By establishing the first prediction model, the optical data of the lamp panel after film coating can be directly generated from the lamp panel optical data, eliminating the need for separate image acquisition of the coated lamp panel.

[0010] In some possible implementations, the first generative network model is iteratively trained using a discriminative network model. This includes: training the first generative network model based on multiple sets of first sample data in each subset of first sample data, enabling the first generative network model to generate multiple sets of backlight module optical data; each set of backlight module optical data includes optical data of the lamp panel sample in the lit state and optical data of the lamp panel after coating; training the discriminative network model based on the multiple sets of first sample data and the multiple sets of backlight module optical data generated by the first generative network model, such that the multiple sets of backlight module optical data generated by the first generative network model approximate the multiple sets of first sample data. By iteratively training the first generative network model using the discriminative network model, the first generative network model generates generated data that is infinitely close to the collected backlight module sample data.

[0011] In some possible implementations, obtaining the first sample dataset includes: acquiring multiple images of each of the N types of light panel samples to obtain multiple sets of first sample data.

[0012] In some possible implementations, the backlight module undergoes quality inspection based on its optical data to obtain the quality inspection result. This includes: detecting abnormal lamp points and uniformity between lamp beads based on the lamp board's optical data to obtain the lamp board inspection result; performing uniformity and defect detection on the optical data after the lamp board is coated to obtain the post-coating inspection result; and obtaining the overall quality inspection result of the backlight module based on both the lamp board inspection result and the post-coating inspection result. In other words, the quality inspection result of the backlight module is obtained by separately performing quality inspections on the lamp board's optical data and the optical data after the lamp board is coated.

[0013] Secondly, this application provides a backlight module detection method, wherein the backlight module includes a lamp panel and M types of optical film layers. The method includes: acquiring an image of the lamp panel in a lit state to obtain lamp panel optical data; inputting the lamp panel optical data into a second prediction model so that the second prediction model outputs M optical data of the lamp panel after film coating; the M optical data of the lamp panel after film coating are images presented when the lamp panel is lit and covered with M different optical film layers; and performing quality inspection on the M optical data of the lamp panel after film coating to obtain the M detection results of the backlight module after film coating.

[0014] The backlight module testing method provided in this application allows for the determination of the final BLU (Backlight Luminous Array) test results after different optical film layers are applied to an existing lamp panel, thereby selecting the optimal optical film layer. This eliminates the need to apply different optical film layers to the lamp panel sequentially using the equipment, improving the speed and accuracy of BLU testing and reducing testing costs.

[0015] In some possible implementations, after obtaining the detection results of the M lamp panels of the backlight module after coating, the method further includes: determining the optimal optical data after coating the M lamp panels from the optical data after coating the M lamp panels based on the detection results of the M lamp panels of the backlight module; and determining the optical film layer corresponding to the optimal optical data after coating the M lamp panels as the second optical film layer of the backlight module.

[0016] In some possible implementations, before inputting the optical data of the lamp panel into the second prediction model, the method further includes: acquiring a second sample dataset; the second sample dataset includes M subsets of first sample data, each subset of second sample data includes multiple sets of second sample data; each set of second sample data includes optical data of the lamp panel sample in the lit state, and optical data of the lamp panel sample in the lit state after the light-emitting surface is covered with an optical film layer; the types of optical films corresponding to different sample data subsets are different, and the types of multiple optical films corresponding to the same sample data subset are the same; training the model based on the second sample dataset to obtain a third generative network model; the third generative network model is a model that generates M optical data of the lamp panel after coating based on one lamp panel optical data; iteratively training the third generative network model through a discriminative network model, and using the trained fourth generative network model as the second prediction model.

[0017] In some possible implementations, the third generative network model is iteratively trained by a discriminative network model, including: training the third generative network model based on multiple sets of second sample data in each subset of second sample data, so that the third generative network model generates multiple sets of backlight module optical data; the multiple sets of backlight module optical data include optical data of multiple lamp panel samples in the lit state and optical data of multiple lamp panels after coating; training the discriminative network model based on the multiple sets of second sample data and the multiple sets of backlight module optical data generated by the third generative network model, so that the multiple sets of backlight module optical data generated by the third generative network model approach the multiple sets of second sample data.

[0018] In some possible implementations, obtaining the second sample dataset includes: acquiring multiple images of the backlight module corresponding to each of the M types of optical films to obtain multiple sets of second sample data.

[0019] In some possible implementations, the optical data of the M lamp panels after coating are subjected to quality inspection to obtain the inspection results of the M lamp panels of the backlight module after coating. The method is as follows: uniformity inspection and defect inspection are performed on the optical data of the M lamp panels after coating to obtain the inspection results of the M lamp panels of the backlight module after coating.

[0020] Thirdly, this application provides an electronic device including a processor, a memory, and a computer program stored in the memory, wherein the processor executes the computer program to cause the electronic device to perform the method as described in the first aspect.

[0021] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on an electronic device, causes the electronic device to perform the method as described in the first aspect.

[0022] The method described in the first aspect above can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules or units corresponding to the above functions. For example, a processing module or unit, a display module or unit, etc.

[0023] It is understood that the beneficial effects of the third and fourth aspects mentioned above can be found in the relevant descriptions in the first aspect above, and will not be repeated here. Attached Figure Description

[0024] Figure 1 A flowchart of a backlight module testing method disclosed in the prior art;

[0025] Figure 2 A flowchart illustrating a backlight module detection method provided in this application embodiment;

[0026] Figure 3A schematic diagram of the model application process and defect detection process corresponding to the backlight module detection method provided in the embodiments of this application;

[0027] Figure 4 A flowchart illustrating a backlight module detection method provided in this application embodiment;

[0028] Figure 5 This is a schematic diagram illustrating the acquisition of a light panel image using an image acquisition device as disclosed in an embodiment of this application;

[0029] Figure 6 A flowchart illustrating the generation process of the second generative network model provided in this application embodiment;

[0030] Figure 7 This is a schematic diagram illustrating the acquisition of the first sample dataset provided in an embodiment of this application;

[0031] Figure 8 This is a schematic diagram illustrating the iterative training of a first generative network model using a discriminative network model, as provided in an embodiment of this application.

[0032] Figure 9 This is a schematic diagram illustrating a quality inspection of a backlight module, provided as an embodiment of this application.

[0033] Figure 10 A schematic diagram showing how the optical data of the lamp panel provided in this application embodiment is obtained after coating through a prediction model;

[0034] Figure 11 This is a schematic diagram illustrating another method for quality inspection of a backlight module, provided as an embodiment of this application.

[0035] Figure 12 A flowchart illustrating another backlight module detection method provided in this application embodiment;

[0036] Figure 13 This is a schematic diagram illustrating how M optical film layers are applied to a lamp panel to form M backlight modules, as provided in an embodiment of this application. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] In this article, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The symbol " / " in this article indicates that the related objects are in an "or" relationship; for example, A / B means A or B.

[0039] The terms "first" and "second," etc., used in the specification and claims herein are used to distinguish different objects, not to describe a specific order of objects. In the description of embodiments in this application, unless otherwise stated, "multiple" means two or more; for example, multiple processing units refer to two or more processing units, etc.; multiple elements refer to two or more elements, etc.

[0040] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0041] To facilitate understanding of the embodiments of this application, some terms used in the embodiments of this application are explained below, so that those skilled in the art can understand them.

[0042] Backlight module or backlight unit (BLU) refers to a light source behind a liquid crystal display (LCD), whose light emission directly affects the visual effect of the liquid crystal display module (LCM). LCDs themselves do not emit light and must rely on a backlight. The main components of a backlight are: a light source (lamp board), a light guide plate, optical films, and structural components.

[0043] Generative Adversarial Networks (GANs) are deep learning models and one of the most promising unsupervised learning methods on complex distributions in recent years. The model produces relatively good outputs through the game-like learning of at least two modules: a generative model and a discriminative model. In the original GAN ​​theory, G and D are not required to be neural networks; they only need to fit the corresponding generation and discrimination functions. However, in practice, deep neural networks are generally used as G and D. Their functions are as follows: G is like a sample generator, taking noise as input and outputting a realistic sample; D is like a binary classifier, used to determine whether the input sample is real or fake. A good GAN application requires a good training method; otherwise, the freedom of the neural network model may lead to unsatisfactory outputs.

[0044] Currently, there are generally two methods for checking the uniformity or yield of BLU (Bright Luminous Unit) lights: visual inspection or equipment testing. Visual inspection is highly subjective, resulting in inaccurate results and low efficiency. Equipment testing, on the other hand, commonly involves inspecting the illuminated light panels to determine if there are any lighting defects or if the uniformity meets requirements, or inspecting the entire BLU to assess its uniformity.

[0045] See Figure 1 As shown, in existing technologies, the luminous uniformity detection of BLUs generally falls into two categories. One involves capturing images of the illuminated LED board at the LED chip station using vision devices such as cameras before the board is coated, and analyzing the optical data of the light emission to detect luminous uniformity or defects. The other involves capturing images of the coated LED board at the panel station using vision devices after the coating is applied, and analyzing the optical data of the coated LED board to detect luminous uniformity or defects. Since BLUs typically consist of an LED board and multiple optical film layers, the panel station usually requires multiple inspections—one inspection for each optical film layer applied. This method is cumbersome and inefficient.

[0046] Based on this, embodiments of this application provide a backlight module detection method, see [link to relevant documentation]. Figure 2 As shown, by collecting only the optical data of the lamp panel, the optical data after coating is obtained through a prediction model; based on the optical data of the backlight module, defect detection is performed to obtain the overall inspection result of the backlight module BLU. This solution can improve the speed and accuracy of BLU inspection and reduce inspection costs.

[0047] To better understand the embodiments of this application, the embodiments of this application will be briefly described below:

[0048] The method provided in this application is applied to a backlight module, wherein the backlight module includes a lamp panel and an optical film layer covering the light-emitting surface of the lamp panel. The method includes: acquiring an image of the lamp panel in a lit state to obtain lamp panel optical data; inputting the lamp panel optical data into a prediction model corresponding to the optical film layer to obtain optical data after coating; performing quality inspection processing on the lamp panel optical data and the optical data after coating to obtain lamp panel inspection results and coating inspection results; and summarizing the lamp panel inspection results and coating inspection results to obtain the quality inspection result of the backlight module.

[0049] See Figure 3 As shown in the embodiments of this application, a backlight module inspection method includes a model application process and a defect detection process. Through these two processes, when only the optical data of the LED panel is collected, the optical data of the LED panel is input into a prediction model, and the optical data after coating is obtained through the prediction model. Quality inspection processing is performed on the LED panel optical data and the optical data after coating, respectively, to obtain the LED panel inspection result and the post-coating inspection result. The LED panel inspection result and the post-coating inspection result are then summarized to obtain the overall inspection result of the backlight module (BLU). This solution can improve the speed and accuracy of BLU inspection; in the case of multiple optical coating layers, the optimal optical coating layer can also be selected based on the inspection results, reducing inspection costs.

[0050] The execution subject of the backlight module detection method provided in this application embodiment can be the aforementioned electronic device, or a functional module and / or functional entity within the electronic device capable of implementing the backlight module detection method. Furthermore, the solution of this application can be implemented through hardware and / or software, and the specific implementation can be determined according to actual usage requirements; this application embodiment does not impose any limitations. The following description uses an electronic device as an example, in conjunction with the accompanying drawings, to exemplarily illustrate the backlight module detection method provided in this application embodiment.

[0051] The following describes a backlight module testing method and system provided in this application, with reference to specific embodiments.

[0052] Example 1: Backlight Module Detection Method with Given Optical Coating Layers

[0053] Figure 4 This is a schematic flowchart of a backlight module detection method provided in an embodiment of this application. (Refer to...) Figure 4 As shown, the method includes the following steps S101-S103.

[0054] S101: Image acquisition is performed on the illuminated lamp panel to obtain optical data of the lamp panel.

[0055] Reference Figure 5As shown, an image acquisition device is used to capture images of a lit light panel. This image acquisition device may include, for example, a camera, scanner, radar, or laser scanner. These devices can convert real-world images into digital signals for computer processing and analysis. This embodiment of the application does not limit the scope of the invention.

[0056] Understandably, the acquired images should meet conditions such as high contrast, uniform overall grayscale, and moderate brightness to facilitate subsequent quality inspection based on the data.

[0057] S102, input the optical data of the lamp board into the first prediction model corresponding to the first optical film layer to obtain the optical data of the backlight module when the backlight module is lit.

[0058] For example, the obtained backlight module optical data includes lamp panel optical data and lamp panel coated optical data predicted by the first prediction model.

[0059] The first prediction model is a generative adversarial network, a deep learning model that produces fairly good outputs through the mutual learning between at least two models in the framework: a generative network model and a discriminative network model.

[0060] For example, in this embodiment of the application, the first prediction model is a model that generates optical data of a coated lamp panel based on optical data of a lamp panel; the output of the first prediction model is generated data that is infinitely close to the collected optical data of the coated lamp panel.

[0061] In the embodiments of this application, see Figure 6 As shown, before inputting the optical data of the lamp panel into the first prediction model corresponding to the first optical film layer, it is also necessary to establish a first generative network model and train the first generative network model to obtain a second generative network model.

[0062] For example, a first sample dataset is obtained; the first sample dataset includes N subsets of first sample data, and each subset of first sample data includes multiple sets of first sample data; each set of first sample data includes optical data of the lamp board sample in the lit state, and optical data of the lamp board sample in the lit state after the light-emitting surface of the lamp board sample is covered with the first optical film layer; the first optical film layer can be a single-layer film or a multi-layer composite film, and the specific type of film depends on the actual situation, which is not limited in this embodiment.

[0063] See Figure 7 As shown, multiple images were collected for each type of light panel sample in the N types of light panel sample platform to obtain N first sample data subsets; the first sample data subsets in the figure are represented by BLU dataset 1 to N; each first sample data subset includes multiple sets of first sample data.

[0064] Specifically, the model is trained based on the first sample dataset to obtain the first generative network model; the first generative network model is a model that generates optical data of a coated lamp based on optical data of a lamp.

[0065] In this embodiment, G is likened to a sample generator, which takes a light panel optical data as input and generates a set of backlight module optical data; D is likened to a binary classifier, which is used to determine whether the input backlight module optical data is true or false.

[0066] In the embodiments of this application, see Figure 8 As shown, the first generative network model also needs to be iteratively trained using a discriminative network model, and the trained second generative network model is used as the first prediction model. By establishing the first prediction model, the optical data of the coated light panel can be directly generated from the light panel's optical data, eliminating the need for separate image acquisition of the coated light panel.

[0067] For example, a first generative network model is trained based on multiple sets of first sample data in each first sample data subset, so that the first generative network model generates multiple sets of backlight module optical data; each set of backlight module optical data includes optical data of the lamp panel sample in the lit state and optical data of the lamp panel after coating.

[0068] Furthermore, a discriminant network model is trained based on multiple sets of first sample data and multiple sets of backlight module optical data generated by the first generative network model.

[0069] For example, multiple sets of backlight module optical data generated by the first generative network model are denoted as generated data, and multiple sets of backlight module sample data collected are denoted as real data. The generated data and real data are input into the discriminative network model, which determines whether the input data is real data (true if it is, false if it is not). The discriminative network outputs the probability value that the input data is real data. The loss function of the generative adversarial network is calculated based on the probability value. Based on the loss function of the generative adversarial network, the parameters of the generative network and the discriminative network are updated using the backpropagation algorithm, making the generated data approximate the real data.

[0070] For example, training stops when the discrimination network cannot distinguish between real and fake input data, that is, the probability output by the discrimination network is 0.5 regardless of whether the input is generated data or real data.

[0071] The generator network and the discriminator network are neural network models that can be CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks), or fully connected neural networks, etc., as long as they can complete the task. This application does not limit them.

[0072] By iteratively training the first generator network model with a discriminant network model, a second generator network model is obtained, which generates generated data that is infinitely close to the collected backlight module sample data.

[0073] The second generator network model is used as the first prediction model. The optical data of the lamp board is input into the first prediction model corresponding to the first optical film layer to obtain the optical data of the backlight module when the backlight module is lit.

[0074] S103, based on the optical data of the backlight module, perform quality inspection on the backlight module to obtain the quality inspection result of the backlight module.

[0075] For example, see details Figure 9 As shown, quality inspection of the backlight module includes inspecting the optical data of the lamp board and inspecting the optical data of the lamp board after it has been coated.

[0076] Based on the optical data of the LED panel, abnormal LED points and uniformity among LED chips are detected, resulting in the LED panel inspection results. Based on the optical data of the LED panel after coating, uniformity and defect detection are performed, resulting in the post-coating inspection results. The LED panel inspection results and the post-coating inspection results are then combined to obtain the backlight module quality inspection results. In other words, by performing quality inspections on both the LED panel optical data and the post-coating optical data, the quality inspection results of the backlight module are obtained.

[0077] In some embodiments, abnormal light spot detection is performed in the following ways: abnormal light spots are quickly and accurately determined by using the abnormal indicators of target light spots within a preset area and by using connected component detection; the uniformity between light beads can be detected by whether the uniformity of light illumination of light beads meets the requirements.

[0078] For example, uniformity detection and defect detection can be performed based on the optical data after the lamp panel is coated. Possible implementation methods include: acquiring a captured image of the display screen to be inspected; generating a defect detection map based on the captured image; and displaying the generated defect detection map in the inspection equipment so that users can intuitively view the defects.

[0079] Specifically, in this embodiment of the application, the display uniformity judgment and / or product grade classification of the display screen to be tested can be performed based on at least one of the following: the number of defective areas, the average defect score of all defective areas, and the location of the defective areas.

[0080] For example, in one application scenario of this application embodiment, a backlight module detection method provided by this application embodiment is described, see [link to relevant documentation]. Figure 10 As shown, when only the optical data of the lamp panel is collected, the optical data of the lamp panel is input into the first prediction model, and the optical data of the lamp panel after coating is obtained through the first prediction model; see [link to relevant documentation]. Figure 11 As shown, the optical data of the LED panel and the optical data after the LED panel is coated are subjected to quality inspection processing to obtain the LED panel inspection results and the post-coating inspection results. The LED panel inspection results and the post-coating inspection results are then summarized to obtain the overall inspection results of the backlight module (BLU). This solution can improve the speed and accuracy of BLU inspection and reduce inspection costs.

[0081] The scenario in Example 1 can be extended to another application: When actual screen manufacturers research and develop or manufacture backlight modules, they need to design for different numbers of LEDs, LED spacing, LED arrangement, and their coordination with optical films to achieve uniform overall light emission of the backlight module. If the backlight module is actually produced and then tested, and the uniformity is found to be unsatisfactory, it will result in a waste of time and materials. Based on this, this application provides a backlight module testing method. In the case of multiple optical films, optical data of the LED panel is collected. Using a second prediction model corresponding to each optical film layer, the optical data after applying different optical films to the light-emitting surface of the LED panel is predicted. The optical data after coating the LED panel is analyzed to determine the optimal coating scheme. This scheme is described in detail in Example 2 below.

[0082] Example 2: Backlight module testing method with M types of optical film layers

[0083] Figure 12 This is a schematic flowchart of a backlight module detection method provided in an embodiment of this application. (Refer to...) Figure 12 As shown, the method includes the following steps S201-S203.

[0084] S201: Acquire an image of the illuminated lamp panel to obtain the lamp panel's optical data.

[0085] An image acquisition device is used to acquire images of the illuminated light panel. Images with high contrast, uniform overall grayscale, and moderate brightness are selected to obtain the optical data of the light panel. The image acquisition method is the same as that described in Example 1 above.

[0086] S202, input the optical data of the lamp panel into the second prediction model so that the second prediction model outputs the optical data of M lamp panels after coating.

[0087] The second prediction model is a generative adversarial network, a deep learning model that produces fairly good outputs through the mutual learning between at least two models in the framework: a generative network model and a discriminative network model.

[0088] In this embodiment, the second prediction model is a model that generates optical data of a coated lamp panel based on optical data of a lamp panel; the output of the second prediction model is generated data that is infinitely close to the collected optical data of the coated lamp panel. It is understood that in this embodiment, M types of optical films correspond to different second prediction models, and there are a total of M second prediction models in this embodiment.

[0089] In this embodiment of the application, before inputting the optical data of the lamp board into the second prediction model, a third generative network model needs to be established and trained to obtain a fourth generative network model.

[0090] For example, see Figure 13 As shown, M optical films are applied to a given lamp panel sample to form M backlight modules as samples. Each backlight module includes one optical film and a given lamp panel.

[0091] Further, a second sample dataset is obtained; the second sample dataset includes M subsets of first sample data, and each subset of second sample data includes multiple sets of second sample data; each set of second sample data includes optical data of the lamp panel sample in the lit state, and optical data of the lamp panel sample in the lit state after the light-emitting surface is covered with an optical film layer; the types of optical film layers corresponding to different sample data subsets are different, and the types of multiple optical film layers corresponding to the same sample data subset are the same. Among them, the M types of optical film layers can be single-layer films or multi-layer composite films, and the specific type of film depends on the actual situation, which is not limited in this embodiment.

[0092] Furthermore, multiple images were acquired for each type of optical film corresponding to the backlight module to obtain multiple sets of second sample data.

[0093] Specifically, the model is trained based on the second sample dataset to obtain the third generative network model; the third generative network model is a model that generates M optical data of the coated lamp panels based on the optical data of one lamp panel.

[0094] In this embodiment, G is likened to a sample generator, which takes a light panel optical data as input and generates a set of backlight module optical data; D is likened to a binary classifier, which is used to determine whether the input backlight module optical data is true or false.

[0095] In this embodiment of the application, it is also necessary to iteratively train the third generative network model through the discriminative network model. By establishing the second prediction model, the optical data of the lamp board after coating can be directly generated from the optical data of the lamp board, without the need to perform multiple image acquisitions on the coated lamp board.

[0096] For example, a third generative network model is trained based on multiple sets of second sample data in each subset of second sample data, so that the third generative network model generates multiple sets of backlight module optical data; the multiple sets of backlight module optical data include optical data of multiple lamp panel samples in the lit state and optical data of multiple lamp panels after coating.

[0097] Furthermore, a discrimination network model is trained based on multiple sets of backlight module sample data collected and multiple sets of backlight module optical data generated by the third generative network model.

[0098] For example, multiple sets of backlight module optical data generated by the third generative network model are denoted as generated data, and multiple sets of backlight module sample data collected are denoted as real data. The generated data and real data are input into the discriminative network model, which determines whether the input data is real data (true if it is, false if it is not). The discriminative network outputs the probability value that the input data is real data. The loss function of the generative adversarial network is calculated based on the probability value. Based on the loss function of the generative adversarial network, the parameters of the generative network and the discriminative network are updated using the backpropagation algorithm, making the generated data approximate the real data.

[0099] For example, training stops when the discrimination network cannot distinguish between real and fake input data, that is, the probability output by the discrimination network is 0.5 regardless of whether the input is generated data or real data.

[0100] The generator network and the discriminator network are constructed neural network models, which can be CNN, RNN or fully connected neural network, etc., as long as they can complete the task. This application does not limit them.

[0101] By iteratively training the third generator network model with the discriminant network model, a fourth generator network model is obtained, which makes the multiple sets of backlight module optical data generated by the fourth generator network model approximate the multiple sets of backlight module sample data collected.

[0102] The trained fourth generative network model is used as the second prediction model. The optical data of the lamp panel is input into the second prediction model, and the second prediction model outputs the optical data of M lamp panels after coating.

[0103] S203, perform quality inspection on the optical data of M lamp panels after coating, and obtain the inspection results of the M lamp panels of the backlight module after coating.

[0104] For example, uniformity and defect detection are performed on the optical data of M lamp panels after coating to obtain the detection results of the M lamp panels of the backlight module after coating.

[0105] In this embodiment of the application, uniformity detection and defect detection are performed on the optical data of M lamp panels after coating. Possible implementation methods include: acquiring the captured image of the display screen to be inspected; generating a defect detection map based on the captured image; and displaying the generated defect detection map in the inspection equipment to facilitate users to intuitively view the defects.

[0106] Specifically, in this embodiment of the application, the display uniformity judgment and / or product grade classification of the display screen to be tested can be performed based on at least one of the following: the number of defective areas, the average defect score of all defective areas, and the location of the defective areas.

[0107] The backlight module testing method provided in this application collects optical data from the lamp board when multiple optical film layers are present. Using a second prediction model corresponding to each optical film layer, it predicts the optical data after applying different optical film layers. The method then analyzes the coated optical data to determine the optimal coating scheme. This approach improves the speed and accuracy of BLU (Backlight Luminous Unit) testing and reduces testing costs.

[0108] The foregoing mainly describes the solutions provided by the embodiments of this application from the perspective of method steps. It is understood that, in order to achieve the above functions, the electronic device implementing this method includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of protection of this application.

[0109] This application embodiment can divide an electronic device into functional modules based on the above method example. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; other feasible division methods may exist in actual implementation.

[0110] It should also be noted that in the embodiments of this application, "greater than" can be replaced with "greater than or equal to", "less than or equal to" can be replaced with "less than", or "greater than or equal to" can be replaced with "greater than", and "less than" can be replaced with "less than or equal to".

[0111] This application also provides a chip coupled to a memory, which is used to read and execute computer programs or instructions stored in the memory to perform the methods described in the above embodiments.

[0112] This application also provides an electronic device, which includes a chip for reading and executing computer programs or instructions stored in a memory, so that the methods in the various embodiments are executed.

[0113] This application also provides a computer-readable storage medium storing computer instructions. When the computer instructions are executed on an electronic device, the electronic device performs the aforementioned method steps to implement a method for obtaining wheel alignment parameters as described in the above embodiments.

[0114] This application also provides a computer program product, which is a computer-readable storage medium storing program code. When the computer program product is run on a computer, it causes the computer to perform the above-mentioned related steps to implement a method for obtaining wheel alignment parameters in the above embodiments.

[0115] In addition, embodiments of this application also provide an apparatus, which may specifically be a chip, component or module. The apparatus may include a connected processor and a memory; wherein the memory is used to store computer execution instructions. When the apparatus is running, the processor may execute the computer execution instructions stored in the memory to cause the chip to execute a method for obtaining wheel alignment parameters in the above-described method embodiments.

[0116] In this application, the electronic device, computer-readable storage medium, computer program product or chip provided in the embodiments are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0117] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0118] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection described in the claims.

Claims

1. A method for detecting a backlight module, characterized in that, The backlight module includes a lamp panel to be tested and a first optical film layer covering the light-emitting surface of the lamp panel. The method includes: Image acquisition is performed on the illuminated lamp panel to obtain optical data of the lamp panel; The optical data of the lamp panel is input into the first prediction model corresponding to the first optical film layer to obtain the optical data of the backlight module when the backlight module is lit. Based on the optical data of the backlight module, the quality of the backlight module is inspected to obtain the quality inspection result of the backlight module. Before inputting the optical data of the lamp panel into the first prediction model corresponding to the first optical film layer, the method further includes: acquiring a first sample dataset; the first sample dataset includes N first sample data subsets, each first sample data subset including multiple sets of first sample data; each set of first sample data includes optical data of the lamp panel sample in the lit state, and optical data of the lamp panel sample in the lit state after the light-emitting surface of the lamp panel sample is covered with the first optical film layer; different types of lamp panel samples correspond to different sample data subsets, and multiple lamp panel samples corresponding to the same sample data subset have the same type; training the model according to the first sample dataset to obtain a first generative network model; iteratively training the first generative network model through a discriminative network model, and using the trained second generative network model as the first prediction model.

2. The method according to claim 1, characterized in that, The iterative training of the first generative network model using a discriminative network model includes: The first generative network model is trained based on multiple sets of first sample data in each subset of the first sample data, so that the first generative network model generates multiple sets of backlight module optical data; each set of backlight module optical data includes optical data of the lamp panel sample in the lit state and optical data of the lamp panel after coating. Based on the multiple sets of first sample data and the multiple sets of backlight module optical data generated by the first generator network model, the discriminant network model is trained so that the multiple sets of backlight module optical data generated by the first generator network model are close to the multiple sets of first sample data.

3. The method according to claim 1 or 2, characterized in that, The process of obtaining the first sample dataset includes: acquiring multiple images of each of the N types of light panel samples to obtain multiple sets of first sample data.

4. The method according to claim 1, characterized in that, The step of performing quality inspection on the backlight module based on the optical data of the backlight module to obtain the quality inspection result of the backlight module includes: Abnormal lamp points and uniformity between lamp beads are detected on the optical data of the lamp board to obtain the lamp board detection results; Uniformity and defect detection are performed on the optical data of the lamp panel after coating to obtain the detection results after lamp panel coating; Based on the test results of the lamp panel and the test results after the lamp panel is coated, the quality test results of the backlight module are obtained.

5. A method for detecting a backlight module, characterized in that, The backlight module includes a lamp board to be tested and M types of optical films. The method includes: Images are captured of the illuminated lamp panel to obtain optical data of the lamp panel; The optical data of the lamp panel is input into the second prediction model so that the second prediction model outputs M optical data of the lamp panel after coating; the M optical data of the lamp panel after coating are the images presented when the lamp panel is lit and covered with M different optical film layers respectively. The optical data of the M lamp panels after coating are respectively subjected to quality inspection to obtain the inspection results of the M lamp panels of the backlight module after coating. Before inputting the optical data of the lamp panel into the second prediction model, the method further includes: acquiring a second sample dataset; the second sample dataset includes M subsets of first sample data, each subset of second sample data includes multiple sets of second sample data; each set of second sample data includes optical data of the lamp panel sample in the lit state, and optical data of the lamp panel sample in the lit state after the light-emitting surface is covered with an optical film layer; different types of optical film layers correspond to different subsets of sample data, and multiple optical film layers in the same subset of sample data have the same type; training the model based on the second sample dataset to obtain a third generative network model; the third generative network model is a model that generates M optical data of the lamp panel after coating based on one lamp panel optical data; iteratively training the third generative network model through a discriminative network model, and using the trained fourth generative network model as the second prediction model.

6. The method according to claim 5, characterized in that, After obtaining the detection results of the M lamp panels of the backlight module after coating, the method further includes: Based on the test results of the M lamp panels of the backlight module after coating, the optimal optical data after coating the M lamp panels is determined from the optical data of the M lamp panels after coating. The optical film layer corresponding to the optical data after the optimal lamp panel is coated is determined as the optimal optical film layer of the backlight module.

7. The method according to claim 5 or 6, characterized in that, The iterative training of the third generative network model using the discriminative network model includes: The third generative network model is trained based on multiple sets of second sample data in each subset of the second sample data, so that the third generative network model generates multiple sets of backlight module optical data; the multiple sets of backlight module optical data include optical data of the multiple lamp panel samples in the lit state and optical data of the multiple lamp panels after coating. The discriminant network model is trained based on the multiple sets of second sample data and the multiple sets of backlight module optical data generated by the third generative network model, so that the multiple sets of backlight module optical data generated by the third generative network model are close to the multiple sets of second sample data.

8. The method according to claim 5 or 6, characterized in that, The process of obtaining the second sample dataset includes: acquiring multiple images of the backlight module corresponding to each of the M types of optical film layers to obtain multiple sets of second sample data.

9. The method according to claim 5, characterized in that, The quality inspection of the optical data of the M lamp panels after coating is performed to obtain the inspection results of the M lamp panels of the backlight module after coating, including: Uniformity and defect detection are performed on the optical data of the M lamp panels after coating to obtain the detection results of the M lamp panels of the backlight module after coating.

10. An electronic device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory, the processor executing the computer program to cause the electronic device to perform the method as described in any one of claims 1 to 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when run on an electronic device, causes the electronic device to perform the method as described in any one of claims 1 to 9.

Images