Polaroid PSA glue for silk-screen printing ink glass and preparation method thereof

Abstract

The invention relates to a polaroid PSA glue for silk-screen printing ink glass and a preparation method thereof, a PSA main agent is diluted to obtain a diluted PSA main agent, and the PSA main agent is acrylic ester with the molecular weight of 9 million to 13 million; an additive is dropwise added into the diluted PSA main agent, the additive comprises a bridging agent, a silane coupling agent and a stripping agent, and the polaroid PSA glue for the silk-screen printing ink glass is obtained. Therefore, the quality, the performance and the stripping characteristic of the PSA glue can be ensured, the polaroid PSA glue is suitable for being attached to silk-screen printing ink black frame glass, and the bad conditions of layering, degumming and bubbles occurring in polaroid attaching are solved.

Description

Technical Field

[0001] The present application relates to the technical field of polarizer PSA adhesive, and more specifically, to a polarizer PSA adhesive for screen-printed ink glass and a preparation method thereof. Background Art

[0002] Polarizers are key components of liquid crystal displays (LCDs) and are used to control the polarization of light passing through the liquid crystal layer. In order to achieve the all-black effect of LCDs, customers usually screen-print ink on the liquid crystal glass to make the frame appear black, and polarizers are usually attached to the screen-printed black frame glass.

[0003] However, when laminating the polarizer to the black-frame glass with silk-screen ink, the traditional PSA glue can only adhere to the glass surface, and it is difficult to achieve effective bonding and lamination for the surface of this ink black frame. Over time, conventional PSA glue will cause problems such as delamination, degumming and bubbles in the polarizer. Since silk-screen ink requires a large peeling force, it is easy to break when tearing off the polarizer, and it may squeeze the liquid crystal glass and damage the display effect. In addition, the glue is also easy to peel off from the polarizer and remain on the liquid crystal module. Therefore, it is very important to solve the problem of laminating the polarizer PSA glue to the black frame, which is directly related to the quality stability and aesthetics of the polarizer product.

[0004] Therefore, an optimized preparation scheme for polarizer PSA adhesive for screen-printed ink glass is desired.

[0005] Application Contents

[0006] In order to solve the above technical problems, the present application is proposed. The present application provides a polarizer PSA glue for screen-printed ink glass and a preparation method thereof, wherein the PSA main agent is diluted to obtain a diluted PSA main agent, wherein the PSA main agent is an acrylate with a molecular weight between 900,000 and 1.3 million; and, an additive is dripped into the diluted PSA main agent, wherein the additive includes a bridging agent, a silane coupling agent and a stripping agent to obtain a polarizer PSA glue for screen-printed ink glass. In this way, the quality, performance and stripping characteristics of the PSA glue can be ensured, so that the polarizer PSA glue is suitable for bonding screen-printed ink black frame glass, and the undesirable conditions of delamination, debonding and bubbles in polarizer bonding are solved.

[0007] In a first aspect, a method for preparing a polarizer PSA adhesive for screen-printed ink glass is provided, comprising:

[0008] Diluting the PSA main agent to obtain a diluted PSA main agent, wherein the PSA main agent is an acrylate with a molecular weight between 900,000 and 1.3 million; and

[0009] Additives are added dropwise into the diluted PSA main agent, wherein the additives include a bridging agent, a silane coupling agent and a stripping agent to obtain a polarizer PSA adhesive for screen printing ink glass.

[0010] In the above-mentioned preparation method of polarizer PSA glue for screen-printed ink glass, the PSA main agent is diluted to obtain the diluted PSA main agent, including: pouring the PSA raw glue into a stainless steel barrel padded with a dust-free bag; weighing a certain amount of ethyl acetate and pouring it into the stainless steel barrel; and stirring by a motor to obtain the diluted PSA main agent.

[0011] In the above-mentioned preparation method of polarizer PSA glue for screen-printed ink glass, additives are added to the diluted PSA main agent, and the additives include a bridging agent, a silane coupling agent and a stripping agent to obtain the polarizer PSA glue for screen-printed ink glass, including: while the diluted PSA main agent is being stirred, the bridging agent is added; after the bridging agent is added, stirring is continued for a predetermined time; the silane coupling agent is continued to be added; after the silane coupling agent is added, stirring is continued for a predetermined time; the stripping agent is continued to be added; after the stripping agent is added, stirring is performed; and standing for half an hour or more to obtain the polarizer PSA glue for screen-printed ink glass.

[0012] In the above-mentioned preparation method of polarizer PSA glue for screen-printed ink glass, after the stripping agent is dripped, stirring is performed, including: obtaining a glue stirring state monitoring video collected by a camera; performing key frame sampling on the glue stirring state monitoring video to obtain a sequence of glue stirring state key image frames; passing each glue stirring state key image frame in the sequence of glue stirring state key image frames through a stirring state feature extractor based on a convolutional neural network model to obtain a time series of glue stirring state feature vectors; passing the time series of the glue stirring state feature vectors through an adaptive attention weight allocation network based on content importance to obtain a time series of adaptive glue stirring state feature vectors; passing the time series of the adaptive glue stirring state feature vectors through a stirring state temporal reasoner based on stirring state prior to obtain a stirring state temporal reasoning semantic expression feature vector; and passing the stirring state temporal reasoning semantic expression feature vector through a stirring state controller based on a classifier to obtain a control instruction, and the control instruction is used to indicate whether to stop stirring.

[0013] In the above-mentioned preparation method of polarizer PSA glue for screen-printed ink glass, the time series of the glue stirring state feature vector is passed through the content-importance-based adaptive attention weight allocation network to obtain the time series of the adaptive glue stirring state feature vector, including: the time series of the glue stirring state feature vector is processed through the content-importance-based adaptive attention weight allocation network with the following weight allocation formula to obtain the time series of the adaptive glue stirring state feature vector; wherein the weight allocation formula is:

[0014]

[0015] Among them, h i is the i-th glue stirring state feature vector in the sequence of glue stirring state feature vectors, and W i Represent the weight coefficient vector and weight coefficient matrix respectively, B i is the offset vector, Selu(·) represents the Selu function, e i is the attention score of the ith glue stirring state feature vector, λ and α are both hyperparameters, exp(·) represents exponential operation, e k is the attention score of the kth glue stirring state feature vector, t is the number of vectors in the sequence of the glue stirring state feature vector, a i is the weight value of the i-th glue stirring state feature vector, is the i-th adaptive glue stirring state feature vector in the sequence of the adaptive glue stirring state feature vectors.

[0016] In the above-mentioned preparation method of polarizer PSA glue for screen-printed ink glass, the time series of the adaptive glue stirring state feature vector is passed through a stirring state temporal reasoner based on stirring state prior to obtain a stirring state temporal reasoning semantic expression feature vector, including: the time series of the adaptive glue stirring state feature vector is processed by the stirring state temporal reasoner based on stirring state prior with the following stirring state reasoning formula to obtain the stirring state temporal reasoning semantic expression feature vector; wherein the stirring state reasoning formula is:

[0017]

[0018] Wherein, V1 is the adaptive glue stirring state feature vector of the last position in the time series of the adaptive glue stirring state feature vector, V rand is the adaptive glue stirring state feature vector at a random position in the time series of the adaptive glue stirring state feature vector, V *is the mean eigenvector of the time series of the adaptive glue stirring state eigenvector, p1 and p2 are weight hyperparameters, f1 and f2 are hyperparameters that use a Gaussian distribution random number function with a mean of 0 and a variance of 1 as the coefficient of the Gaussian distribution function, and ⊕ are vector subtraction and vector addition respectively, and V is the semantic expression feature vector of the stirring state temporal reasoning.

[0019] In the above-mentioned method for preparing polarizer PSA glue for silk-screen ink glass, a training step is also included: for training the stirring state feature extractor based on the convolutional neural network model, the content importance-based adaptive attention weight allocation network, the stirring state temporal reasoner based on stirring state prior, and the classifier-based stirring state controller.

[0020] In the above-mentioned method for preparing polarizer PSA glue for screen-printed ink glass, the training step includes: obtaining a training glue stirring state monitoring video collected by a camera; performing key frame sampling on the training glue stirring state monitoring video to obtain a sequence of training glue stirring state key image frames; passing each training glue stirring state key image frame in the sequence of training glue stirring state key image frames through the stirring state feature extractor based on the convolutional neural network model to obtain a time series of training glue stirring state feature vectors; passing the time series of training glue stirring state feature vectors through the content importance-based adaptive attention weight allocation network to obtain a time series of training adaptive glue stirring state feature vectors; The time series of the glue stirring state feature vector is passed through the stirring state temporal reasoning device based on stirring state prior to obtain the training stirring state temporal reasoning semantic expression feature vector; the training stirring state temporal reasoning semantic expression feature vector is clustered and optimized to obtain the optimized training stirring state temporal reasoning semantic expression feature vector; the optimized training stirring state temporal reasoning semantic expression feature vector is passed through the classifier-based stirring state controller to obtain the classification loss function value; based on the classification loss function value, the stirring state feature extractor based on the convolutional neural network model, the content importance-based adaptive attention weight allocation network, the stirring state temporal reasoning device based on stirring state prior and the classifier-based stirring state controller are trained.

[0021] In the above-mentioned method for preparing polarizer PSA glue for silk-screen ink glass, the training stirring state temporal reasoning semantic expression feature vector is clustered and optimized to obtain an optimized training stirring state temporal reasoning semantic expression feature vector, including: clustering each eigenvalue of the training stirring state temporal reasoning semantic expression feature vector based on the distance between eigenvalues ​​to obtain a clustering feature set; based on the clustering feature set, optimizing the training stirring state temporal reasoning semantic expression feature vector based on the clustered feature in-class and out-of-class representation to obtain the optimized training stirring state temporal reasoning semantic expression feature vector.

[0022] In a second aspect, a polarizer PSA glue for screen-printed ink glass is provided, which includes: the polarizer PSA glue for screen-printed ink glass is prepared by the preparation method of the polarizer PSA glue for screen-printed ink glass as described above.

[0023] Compared with the prior art, the present application provides a polarizer PSA adhesive for screen-printed ink glass and a preparation method thereof, wherein the PSA main agent is diluted to obtain a diluted PSA main agent, wherein the PSA main agent is an acrylate with a molecular weight between 900,000 and 1.3 million; and an additive is added dropwise to the diluted PSA main agent, wherein the additive includes a bridging agent, a silane coupling agent, and a stripping agent to obtain a polarizer PSA adhesive for screen-printed ink glass. In this way, the quality, performance, and stripping characteristics of the PSA adhesive can be ensured, so that the polarizer PSA adhesive is suitable for laminating screen-printed ink black frame glass, and the undesirable conditions of delamination, debonding, and bubbles in polarizer lamination are solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 The present invention is a flow chart of a method for preparing polarizer PSA glue for screen-printed ink glass according to an embodiment of the present application.

[0026] Figure 2 It is a schematic diagram showing the influence of molecular weight on various properties according to the examples of the present application.

[0027] Figure 3 This is a flow chart of step 120 in the method for preparing polarizer PSA glue for screen-printed ink glass according to an embodiment of the present application.

[0028] Figure 4Schematic diagram of the effect of the bridging agent according to an embodiment of the present application.

[0029] Figure 5 Schematic diagram of the reaction of an isocyanate bridging agent according to an embodiment of the present application.

[0030] Figure 6 This is a flow chart of step 126 in the method for preparing polarizer PSA glue for screen-printed ink glass according to an embodiment of the present application.

[0031] Figure 7 It is a block diagram of a system for preparing polarizer PSA glue for screen-printed ink glass according to an embodiment of the present application.

[0032] Figure 8 This is a schematic diagram of a scenario of a method for preparing polarizer PSA glue for screen-printed ink glass according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0034] Unless otherwise specified, all technical and scientific terms used in the embodiments of the present application have the same meaning as those commonly understood by those skilled in the art of the present application. The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit the scope of this application.

[0035] In the embodiments of the present application, it should be noted that, unless otherwise specified and limited, the term "connection" should be understood in a broad sense. For example, it can be an electrical connection or a connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0036] It should be noted that the terms "first\second\third" involved in the embodiments of the present application are only used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that the specific order or sequence of "first\second\third" can be interchanged where permitted. It should be understood that the objects distinguished by "first\second\third" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0037] In order to achieve an integrated black effect, clients usually screen-print ink on the LCD glass to make the frame appear black. Conventional PSA glue can only adhere to the glass surface, but it is difficult to effectively bond the surface of the black frame with such ink. After weathering, delamination, degumming and bubbles will occur.

[0038] Silk-screen ink requires a large peeling force, so the polarizer is prone to breakage during removal, squeezing the liquid crystal glass and damaging the display effect. The glue is easy to peel off the polarizer and remain on the liquid crystal module. Currently, most glues cannot meet the usage requirements. Therefore, solving the fitting of the polarizer PSA to the black frame is related to the quality stability of the polarizer products and the future growth of automotive products. Solving this problem is imperative.

[0039] In this application, the key technologies include:

[0040] (1) Selection of PSA main agent. There are many types of PSAs, but not many are suitable for attaching ink black frames. The functional group structure and molecular weight of PSA must be used as important references. Therefore, a lot of testing is required to carefully select a PSA main agent with excellent performance.

[0041] (2) Selection and ratio of bridging agents. The chemical reaction principles of different bridging agents are quite different. Isocyanate, epoxy, metal, and amino bridging agents all have their own advantages and disadvantages. Repeated experiments are required to screen the bridging agent and ratio that meet the composite requirements.

[0042] (3) The selection and ratio of silane coupling agent and release agent. The adhesion and high durability of ink cannot be guaranteed without these two auxiliary agents. Only by properly configuring the ratio of the two auxiliary agents can a perfect pressure-sensitive adhesive be obtained.

[0043] Figure 1 FIG. 1 is a flow chart of a method for preparing a polarizer PSA adhesive for screen printing ink glass according to an embodiment of the present application. Figure 1 As shown, the preparation method of the polarizer PSA glue for screen-printed ink glass includes: 110, diluting the PSA main agent to obtain a diluted PSA main agent, wherein the PSA main agent is an acrylate with a molecular weight between 900,000 and 1.3 million; and, 120, adding additives to the diluted PSA main agent, wherein the additives include a bridging agent, a silane coupling agent and a stripping agent to obtain the polarizer PSA glue for screen-printed ink glass.

[0044] In view of the above technical problems, in the technical solution of the present application, a method for preparing a polarizer PSA glue for screen-printed ink glass is proposed, which comprises the following steps: first, diluting the PSA main agent to obtain a diluted PSA main agent. It should be understood that the molecular weight of the acrylate has a decisive influence on the performance of the PSA glue. Theoretically, the larger the molecular weight of the acrylate main agent, the stronger and more excellent the rigidity of the PSA colloid prepared in the end, but if the molecular weight is too large, it is easy to peel off after weathering. Specifically, in a specific example of the present application, an acrylate with a molecular weight between 900,000 and 1.3 million is used as the PSA main agent. Then, additives are added dropwise to the diluted PSA main agent, and the additives include a bridging agent, a silane coupling agent, and a stripping agent to obtain a polarizer PSA glue for screen-printed ink glass.

[0045] It should be understood that the structures of pressure-sensitive adhesive used in polarizers include: TAC, PVA film, TAC, PSA pressure-sensitive adhesive suitable for screen printing ink, and PET release film.

[0046] The PSA pressure-sensitive adhesive suitable for screen printing ink is not much different from the conventional pressure-sensitive adhesive in coating process. To make the pressure-sensitive adhesive have the ability to adhere to ink and have high durability, first, the main agent of the glue should be selected. A main agent with a larger molecular weight should be selected, which has higher rigidity and is not prone to glue tearing and residual glue. Second, the PSA formula needs to be adjusted, and the amount of metal bridging agent and release agent should be appropriately increased, so that the glue can maintain a good balance between weather resistance and ink adhesion.

[0047] Among them, the selection of PSA main agent includes:

[0048] 1. Determination of molecular weight, such as Figure 2 As shown, the molecular weight of acrylate has a decisive influence on the performance of pressure-sensitive adhesive. Theoretically, the larger the molecular weight of the acrylate main agent, the stronger and more excellent the rigidity of the PSA colloid prepared in the end. However, if the molecular weight is too large, it is easy to peel off after weathering, which is counterproductive. The molecular weight of commonly used main agents is between 500,000 and 1.3 million, and that of J212 and J213 is between 700,000 and 1 million. According to experience and test data, the molecular weight of highly durable glue should be between 900,000 and 1.3 million. On this basis, the molecular weight should be carefully selected to achieve the desired suitable main agent.

[0049] Reducing the molecular weight of the main polymer of the pressure-sensitive adhesive can reduce its bulk viscosity, which is beneficial to the flow and wetting of the adhesive on the adhered surface, thereby improving the interfacial adhesion and increasing the peel strength of the pressure-sensitive adhesive product. However, reducing the molecular weight can also reduce the cohesive strength (tensile strength and elastic modulus) of the pressure-sensitive adhesive layer.

[0050] 2. Determination of the functional group of the main agent. Pressure sensitive adhesives are divided into acidic adhesives and acid-free adhesives. Acidic adhesives refer to colloids whose main agent functional groups are rich in carboxyl groups (-COOH). They are characterized by strong adhesion, hard colloids, and excellent weather resistance. The disadvantage is that the peeling force is too high and difficult to tear off. It is easy to corrode ITO conductive glass and cannot be used on specific liquid crystal modules. Acid-free adhesives refer to colloids obtained by replacing carboxyl groups with hydroxyl groups (-OH) in the main agent functional groups. They are characterized by low adhesion and easy to tear off. The colloid is soft and can be applied to ITO glass. The disadvantage is poor weather resistance. Hydroxyl-type acidic adhesives have a wide range of bridging agents to choose from due to their strong functional group activity. Hydroxyl-type acid-free adhesives have strong activity on amino-type bridging agents and are relatively inert to metal bridging agents. The main agents of the two functional groups have their own advantages and disadvantages, and need to be selected and tested according to specific circumstances.

[0051] Specifically, in the process of diluting the PSA main agent to obtain the diluted PSA main agent, that is, step 110, the PSA main agent is diluted to obtain the diluted PSA main agent, including: 111, pouring the PSA raw rubber into a stainless steel barrel padded with a dust-free bag; 112, weighing a certain amount of ethyl acetate and pouring it into the stainless steel barrel; and, 113, stirring by a motor to obtain the diluted PSA main agent.

[0052] Specifically, Figure 3 FIG. 1 is a flow chart of step 120 in the method for preparing polarizer PSA adhesive for screen printing ink glass according to an embodiment of the present application. Figure 3 As shown, 120, adding additives into the diluted PSA main agent, the additives including a bridging agent, a silane coupling agent and a stripping agent to obtain a polarizer PSA glue for screen-printed ink glass, including: 121, adding the bridging agent while the diluted PSA main agent is being stirred; 122, after the bridging agent is added, stirring is continued for a predetermined time; 123, continuing to add the silane coupling agent; 124, after the silane coupling agent is added, stirring is continued for a predetermined time; 125, continuing to add the stripping agent; 126, after the stripping agent is added, stirring is performed; and, 127, standing for half an hour or more to obtain the polarizer PSA glue for screen-printed ink glass.

[0053] Among them, in the selection of bridging agents, the role of bridging agents is to form links between linear polymer main agents through chemical reactions, so that PSA forms a network structure. At present, the types of bridging agents are divided into epoxy, amine, metal, and isocyanate. With the increase in the amount of cross-linking agent, the average molecular weight of the main polymer of the pressure-sensitive adhesive will increase rapidly, and the bulk viscosity and cohesive strength (reaction as tensile strength and elastic modulus, etc.) of the pressure-sensitive adhesive will increase significantly. This is the main reason why cross-linked pressure-sensitive adhesive products generally show complete interfacial adhesion failure during the 180° peel strength test, and the 180° peel strength value will gradually decrease with the increase in the amount of cross-linking agent, such as Figure 4 shown.

[0054] like Figure 5 As shown, taking isocyanate bridging agents as an example, they can react with both hydroxyl-based main agents and carboxyl-based main agents to form a stable network structure.

[0055] However, different cross-linking agents have their own advantages and disadvantages. Through comprehensive evaluation, if high workability is to be achieved, the proportion of metal hardeners needs to be increased, and epoxy or isocyanate type bridging agents need to be added in a certain proportion according to weather resistance conditions, so that they can have high durability. The addition ratio needs to be adjusted repeatedly to achieve a comprehensive balance, refer to Table 1.

[0056] Table 1

[0057]

[0058] In the selection and ratio of silane coupling agent and release agent, silane coupling agent is not a necessary additive for general PSA, but it is indispensable if high durability is to be achieved. Silane coupling agent can make PSA and glass substrate have stronger adhesion in a hot and humid environment, prevent water vapor from corroding the contact surface of PSA adhesive, and thus ensure better temperature and humidity weather resistance.

[0059] The function of the release agent is to adjust the initial peeling adhesion of PSA. It is generally used in conjunction with a silane coupling agent. It is also an important guarantee of peeling strength. Excessive addition of release agent will cause delamination of the adhesive layer after weathering, so repeated testing is required to determine the final usage amount.

[0060] Select the appropriate main agent, match it with the hardener and silane coupling agent, adjust the formula, coat PSA on the polarizer, perform a fitting test using black-frame ink glass, and screen out the best formula.

[0061] In one embodiment of the present application, the specific process flow is as follows:

[0062] (1) Dilution of PSA main agent: Pour a certain amount of PSA raw rubber into a stainless steel barrel lined with a dust-free bag, then weigh a certain amount of ethyl acetate and pour it into the barrel, and stir with a motor for 10 minutes.

[0063] (2) Addition of additives: Additives include bridging agent, silane coupling agent and stripping agent. Use a pipette to draw a certain amount of additives into a beaker. While the PSA main agent is being stirred, slowly add the additives with a dropper. After the addition is complete, continue stirring for 10 minutes, then add the second additive until all additives have been added. After stirring, let it stand for half an hour, take samples in a sealed container, and immediately test the viscosity and solid content.

[0064] (3) Use of PSA: The PSA is transferred to the glue coating section for coating to obtain a composite product of PSA and polarizer.

[0065] (4) Weathering test: Cut the finished polarizer into small pieces, attach them to the black-framed glass with ink, and then conduct tests at 95°C, 105°C, 70°C*95%RH, 85°C*85%RH, etc. At the same time, conduct glue reworkability test, peel strength test, initial adhesion test, and sustained adhesion test.

[0066] Accordingly, in the above-mentioned method for preparing polarizer PSA glue for screen-printed ink glass, the role of adding a stripping agent is to adjust the initial peeling adhesion and stripping performance of the PSA glue, that is, the polarizer can be easily stripped when needed without damaging other components. Stirring can ensure that the stripping agent is fully mixed in the PSA glue to ensure the uniformity and stability of the stripping performance. If the stirring time is too short, the stripping agent may not be fully dispersed, resulting in uneven performance of the PSA glue, affecting the quality of the final product. Stirring time that is too long may cause the colloidal stability of the PSA glue to decrease, phase separation or precipitation, and also waste energy and time. Therefore, in the process of stirring after adding the stripping agent, it is particularly important to determine the stopping time of stirring according to the stirring state of the glue.

[0067] Based on this, the technical concept of this application is to use a camera to monitor and collect the state monitoring video of the glue stirring process in real time, and introduce image processing and analysis algorithms in the back end to perform semantic analysis and time-series correlation feature extraction of each key image frame in the glue stirring state monitoring video, so as to automatically control the stirring process based on the time-series change of the glue stirring state, so as to ensure the quality, performance and peeling characteristics of the PSA glue, while avoiding the problem of energy waste caused by over-stirring. In this way, the polarizer PSA glue is suitable for laminating silk-screen ink black frame glass, solving the problem of delamination, degumming and bubbles in polarizer lamination.

[0068] Figure 6FIG. 1 is a flow chart of step 126 in the method for preparing polarizer PSA glue for screen printing ink glass according to an embodiment of the present application. Figure 6 As shown, 126, after the stripping agent is dripped, stirring is performed, including: 1261, obtaining a monitoring video of the glue stirring state collected by a camera; 1262, performing key frame sampling on the monitoring video of the glue stirring state to obtain a sequence of key image frames of the glue stirring state; 1263, passing each glue stirring state key image frame in the sequence of key image frames of the glue stirring state through a stirring state feature extractor based on a convolutional neural network model to obtain a time series of glue stirring state feature vectors; 1264, passing the time series of the glue stirring state feature vectors through an adaptive attention weight allocation network based on content importance to obtain a time series of adaptive glue stirring state feature vectors; 1265, passing the time series of the adaptive glue stirring state feature vectors through a stirring state temporal reasoner based on stirring state prior to obtain a stirring state temporal reasoning semantic expression feature vector; and, 1266, passing the stirring state temporal reasoning semantic expression feature vector through a stirring state controller based on a classifier to obtain a control instruction, and the control instruction is used to indicate whether to stop stirring.

[0069] More specifically, in the technical solution of the present application, first, a monitoring video of the glue stirring state collected by a camera is obtained. Next, considering that there is a large amount of redundant information in the monitoring video during the actual monitoring of the glue stirring state, in order to reduce the amount of video data, only the most representative and informative key frame images are retained, thereby reducing the amount of calculation and storage space for subsequent processing. In the technical solution of the present application, it is necessary to perform key frame sampling on the monitoring video of the glue stirring state to obtain a sequence of key image frames of the glue stirring state. In particular, here, the key frames usually contain the most important information and changes in the video. By sampling the key frames, the important moments and key information of the changes in the glue stirring state during the glue stirring process can be more effectively extracted. In this way, the changes in the glue stirring state over time can be observed more clearly, which is helpful for analyzing the dynamic changes and trends of the glue stirring state during the stirring process.

[0070] Then, a stirring state feature extractor based on a convolutional neural network model with excellent performance in extracting implicit features of images is used to perform feature extraction on each glue stirring state key image frame in the sequence of glue stirring state key image frames, so as to capture the implicit feature distribution information about the glue stirring state in each glue stirring state key image frame, thereby obtaining a time series of glue stirring state feature vectors.

[0071] Furthermore, considering that during the glue stirring process, its state may change in time series, in order to be able to highlight the key feature information of the change in the glue stirring state and improve the understanding and analysis capabilities of the stirring state and stirring process, in the technical solution of the present application, the time series of the glue stirring state feature vector is passed through an adaptive attention weight allocation network based on content importance to obtain the time series of the adaptive glue stirring state feature vector. It should be understood that the adaptive attention weight allocation network can automatically learn the time series correlation weights according to the glue stirring state characteristics at different time points, so as to highlight and manifest the important features related to the stirring state monitoring, and weaken the other irrelevant interference features, thereby improving the characterization capabilities of the features. At the same time, the model can adapt to different stirring state changes, improving the generalization and adaptability of the model.

[0072] In a specific embodiment of the present application, the time series of the glue stirring state feature vector is passed through an adaptive attention weight allocation network based on content importance to obtain a time series of the adaptive glue stirring state feature vector, including: the time series of the glue stirring state feature vector is passed through the adaptive attention weight allocation network based on content importance to process the time series with the following weight allocation formula to obtain the time series of the adaptive glue stirring state feature vector; wherein the weight allocation formula is:

[0073]

[0074] Among them, h i is the i-th glue stirring state feature vector in the sequence of glue stirring state feature vectors, and W i Represent the weight coefficient vector and weight coefficient matrix respectively, B i is the offset vector, Selu(·) represents the Selu function, e i is the attention score of the ith glue stirring state feature vector, λ and α are both hyperparameters, exp(·) represents exponential operation, e k is the attention score of the kth glue stirring state feature vector, t is the number of vectors in the sequence of the glue stirring state feature vector, a i is the weight value of the i-th glue stirring state feature vector, is the i-th adaptive glue stirring state feature vector in the sequence of the adaptive glue stirring state feature vectors.

[0075] It should be understood that the traditional glue stirring state monitoring method usually relies on semantic analysis of the image frames at each time point in the video to determine whether the stirring state of the glue meets the requirements. However, this method has hysteresis, that is, when it is detected that the stirring state of the glue meets the requirements, the stirring process will continue, which will not only waste energy, but also may reduce the stability of the colloid, and cause phase separation or precipitation. Therefore, in the technical solution of the present application, the time series of the adaptive glue stirring state feature vector is further passed through a stirring state time series reasoner based on a priori stirring state to obtain a stirring state time series reasoning semantic expression feature vector. The reasoner based on a priori stirring state can use the known stirring state time series information to guide the model to more accurately infer the current and predict the stirring state of the next stage, so as to better capture the dynamic changes and trends of the stirring state, improve the understanding and analysis capabilities of the stirring process, and thus improve the effect and accuracy of stirring monitoring and control.

[0076] In a specific embodiment of the present application, the time series of the adaptive glue stirring state feature vector is passed through a stirring state temporal reasoner based on stirring state prior to obtain a stirring state temporal reasoning semantic expression feature vector, including: the time series of the adaptive glue stirring state feature vector is passed through the stirring state temporal reasoner based on stirring state prior to obtain the stirring state temporal reasoning semantic expression feature vector by the following stirring state reasoning formula; wherein the stirring state reasoning formula is:

[0077]

[0078] Wherein, V1 is the adaptive glue stirring state feature vector of the last position in the time series of the adaptive glue stirring state feature vector, V rand is the adaptive glue stirring state feature vector at a random position in the time series of the adaptive glue stirring state feature vector, V * is the mean eigenvector of the time series of the adaptive glue stirring state eigenvector, p1 and p2 are weight hyperparameters, f1 and f2 are hyperparameters that use a Gaussian distribution random number function with a mean of 0 and a variance of 1 as the coefficient of the Gaussian distribution function, and ⊕ are vector subtraction and vector addition respectively, and V is the semantic expression feature vector of the stirring state temporal reasoning.

[0079] Then, the stirring state temporal reasoning semantic expression feature vector is passed through the stirring state controller based on the classifier to obtain a control instruction, and the control instruction is used to indicate whether to stop stirring. In other words, the temporal reasoning semantic expression feature information of the stirring state is used for classification processing, so as to determine the stopping time of stirring according to the temporal semantic reasoning state of the glue stirring state, thereby realizing a more intelligent and accurate stirring process. In this way, the quality, performance and peeling characteristics of the PSA glue can be ensured, while avoiding the problem of energy waste caused by over-stirring.

[0080] In the above technical solution, each glue stirring state feature vector in the time series of the glue stirring state feature vector respectively represents the image semantic features of each glue stirring state key image frame in the sequence of the glue stirring state key image frame. When the time series of the glue stirring state feature vector passes through the content importance-based adaptive attention weight allocation network, the content importance-based adaptive attention weight allocation network adaptively allocates weights based on the semantic correlation between each glue stirring state feature vector in the time series of the glue stirring state feature vector, so as to highlight the important part of the time series of the glue stirring state feature vector and suppress its unimportant part so that the time series of the glue stirring state feature vector has a relatively better significance in the time dimension. However, considering that each glue stirring state key image frame in the sequence of the glue stirring state key image frame has differences in the source image domain space, after image semantic feature extraction and adaptive attention weight allocation, the feature distribution imbalance and discreteness between each glue stirring state feature vector in the time series of the glue stirring state feature vector will be amplified. This means that when the time series of the adaptive glue stirring state feature vector is passed through a stirring state temporal reasoning device based on stirring state prior, the obtained stirring state temporal reasoning semantic expression feature vector has significant local feature distribution discreteness, which causes regression mode deviation, resulting in overflow of the control result obtained by the stirring state temporal reasoning semantic expression feature vector through a classifier-based stirring state controller, affecting its accuracy.

[0081] Based on this, in a preferred embodiment, the stirring state temporal reasoning semantic expression feature vector is passed through a classifier-based stirring state controller to obtain a control instruction, specifically including:

[0082] The stirring state temporal reasoning semantic expression feature vector is passed through a probabilistic activation function to obtain a probabilistic stirring state temporal reasoning semantic expression feature vector;

[0083] Determine a stop probability value indicating stopping stirring obtained by the stirring state controller based on a classifier, which is obtained by the stirring state temporal reasoning semantic expression feature vector of the stirring state;

[0084] After performing point subtraction on the unit feature vector and the probabilistic stirring state temporal reasoning semantic expression feature vector, the power function with the stop probability value as the base is calculated, and further point multiplication is performed with the probabilistic stirring state temporal reasoning semantic expression feature vector to obtain the probabilistic stirring state temporal reasoning semantic convergence limiting feature vector;

[0085] Taking the absolute value of each eigenvalue of the probabilistic stirring state temporal reasoning semantic expression feature vector as the base, calculating a power function with the inverse of the stop probability value as the exponent, and summing all eigenvalues ​​of the probabilistic stirring state temporal reasoning semantic expression feature vector to obtain the stirring state temporal reasoning semantic convergence bias eigenvalue;

[0086] The product of the probabilistic stirring state temporal reasoning semantic convergence limiting feature vector and the stirring state temporal reasoning semantic convergence bias feature value and the weight as a hyperparameter is added to obtain an optimized stirring state temporal reasoning semantic expression feature vector;

[0087] The optimized stirring state temporal reasoning semantic expression feature vector is passed through the classifier-based stirring state controller to obtain the control instruction.

[0088] Wherein, let the probabilistic stirring state temporal reasoning semantic expression feature vector be recorded as V, and the stopping probability value be p, then the optimized stirring state temporal reasoning semantic expression feature vector V' is expressed as:

[0089]

[0090] Where V is the probabilistic semantic expression vector of the stirring state temporal reasoning, α is the weight as a hyperparameter, and v i is the ith eigenvalue of the probabilistic stirring state temporal reasoning semantic expression feature vector, p is the stop probability value, (·) ⊙p It is a power function with the stopping probability value as the exponent to calculate the feature value of each position in the feature vector, ⊙ represents the point multiplication by position, represents difference by position, ⊕ represents addition by position, and V' is the optimized semantic expression feature vector of the stirring state temporal reasoning.

[0091] That is, for the stirring state temporal reasoning semantic expression feature vector, based on the convergence limit of the Cauchy-Hadamard form power series probability distribution, the feature set of the stirring state temporal reasoning semantic expression feature vector is represented by a formal power series based on the convergence probability, and a set convergence bounded bias represented by the convergence radius based on each eigenvalue of the stirring state temporal reasoning semantic expression feature vector is added. Under a unified probability convergence limit standard, this can avoid the invalid overflow of the classification result caused by the deviation of the distribution pattern of the stirring state temporal reasoning semantic expression feature vector during the regression process of the classifier-based stirring state controller due to the gradient propagation of the model parameters based on the convergence probability, thereby improving the accuracy of the recognition and monitoring results obtained by the classifier-based stirring state controller.

[0092] In one embodiment of the present application, the method for preparing polarizer PSA glue for screen-printed ink glass also includes a training step: for training the stirring state feature extractor based on the convolutional neural network model, the content-importance-based adaptive attention weight allocation network, the stirring state time series reasoner based on stirring state prior, and the classifier-based stirring state controller. The training step includes: obtaining a training glue stirring state monitoring video captured by a camera; performing key frame sampling on the training glue stirring state monitoring video to obtain a sequence of training glue stirring state key image frames; passing each training glue stirring state key image frame in the sequence of training glue stirring state key image frames through the stirring state feature extractor based on the convolutional neural network model to obtain a time series of training glue stirring state feature vectors; passing the time series of the training glue stirring state feature vectors through the content-importance-based adaptive attention weight allocation network to obtain a time series of training adaptive glue stirring state feature vectors; and passing the time series of the training adaptive glue stirring state feature vectors through the content-importance-based adaptive attention weight allocation network to obtain a time series of training adaptive glue stirring state feature vectors. The sequence passes through the stirring state temporal reasoning device based on stirring state prior to obtain a training stirring state temporal reasoning semantic expression feature vector; cluster optimization is performed on the training stirring state temporal reasoning semantic expression feature vector to obtain an optimized training stirring state temporal reasoning semantic expression feature vector; the optimized training stirring state temporal reasoning semantic expression feature vector is passed through the stirring state controller based on the classifier to obtain a classification loss function value; based on the classification loss function value, the stirring state feature extractor based on the convolutional neural network model, the content importance-based adaptive attention weight allocation network, the stirring state temporal reasoning device based on stirring state prior and the stirring state controller based on the classifier are trained.

[0093] Furthermore, the training stirring state temporal reasoning semantic expression feature vector is clustered and optimized to obtain an optimized training stirring state temporal reasoning semantic expression feature vector, including: clustering each eigenvalue of the training stirring state temporal reasoning semantic expression feature vector based on the distance between eigenvalues ​​to obtain a cluster feature set; based on the cluster feature set, optimizing the training stirring state temporal reasoning semantic expression feature vector based on the clustered feature in-class and out-of-class representation to obtain the optimized training stirring state temporal reasoning semantic expression feature vector.

[0094] In the above technical solution, each training glue stirring state feature vector in the time series of the training glue stirring state feature vector respectively represents the image semantic features of each training glue stirring state key image frame in the sequence of the training glue stirring state key image frame. When the time series of the training glue stirring state feature vector passes through the content importance-based adaptive attention weight allocation network, the content importance-based adaptive attention weight allocation network adaptively allocates weights based on the semantic correlation between each training glue stirring state feature vector in the time series of the training glue stirring state feature vector, so as to highlight the important part of the time series of the training glue stirring state feature vector and suppress its unimportant part so that the time series of the training glue stirring state feature vector has a relatively better significance in the time dimension. However, considering that each training glue stirring state key image frame in the sequence of the training glue stirring state key image frame has differences in the source image domain space, after image semantic feature extraction and adaptive attention weight allocation, the feature distribution imbalance and discreteness between each training glue stirring state feature vector in the time series of the training glue stirring state feature vector will be amplified. This ensures that when the time series of the training adaptive glue stirring state feature vector is passed through a stirring state temporal reasoner based on stirring state prior, the obtained training stirring state temporal reasoning semantic expression feature vector has significant local feature distribution discreteness.

[0095] In this way, when the training stirring state temporal reasoning semantic expression feature vector is classified and regressed as a whole through a classifier-based stirring state controller, the discreteness of the local feature distribution of the training stirring state temporal reasoning semantic expression feature vector will lead to difficulty in converging towards the predetermined class probability when the classification and regression is performed through the classifier-based stirring state controller, thereby affecting the training speed of the classifier-based stirring state controller and the accuracy of the final classification results.

[0096] Based on this, the applicant of the present application performs clustering optimization on the training stirring state temporal reasoning semantic expression feature vector, that is, first clustering each eigenvalue of the training stirring state temporal reasoning semantic expression feature vector, for example, clustering based on the distance between eigenvalues, and then optimizing based on the clustered feature in-class and out-of-class representations, expressed as: optimizing the training stirring state temporal reasoning semantic expression feature vector based on the clustered feature set based on the following optimization formula to obtain the optimized training stirring state temporal reasoning semantic expression feature vector; wherein, the optimization formula is:

[0097]

[0098] Among them, f is the individual eigenvalues ​​of the training stirring state temporal reasoning semantic expression feature vector, n is the number of feature sets corresponding to the training stirring state temporal reasoning semantic expression feature vector, that is, the length of the training stirring state temporal reasoning semantic expression feature vector, k is the number of cluster features, C represents the cluster feature set, and f' is the individual eigenvalues ​​of the optimized training stirring state temporal reasoning semantic expression feature vector.

[0099] Specifically, by taking the in-class features and out-of-class features of the training stirring state temporal reasoning semantic expression feature vector as different instance roles to perform class instance description based on clustering ratio distribution, and introducing the clustering response history based on the in-class and out-of-class dynamic context, a coordinated global perspective is maintained for the in-class distribution and out-of-class distribution of the overall features of the training stirring state temporal reasoning semantic expression feature vector, so that the optimized feature clustering operation for the training stirring state temporal reasoning semantic expression feature vector can maintain a consistent response of the in-class and out-of-class features, so that the regression convergence path based on feature clustering is kept consistent in the class regression process, and the convergence effect of the training stirring state temporal reasoning semantic expression feature vector towards the predetermined class probability is improved, so as to improve the training speed of the stirring state controller based on the classifier and the accuracy of the classification results. In this way, the stirring process can be automatically controlled based on the temporal changes of the stirring state of the glue, thereby ensuring the quality, performance and peeling characteristics of the PSA glue, while avoiding the problem of energy waste caused by over-stirring.

[0100] In summary, the intelligent control method of the lighting driving power supply based on the embodiment of the present application is explained, which uses a camera to monitor and collect the state monitoring video of the glue stirring process in real time, and introduces image processing and analysis algorithms at the back end to perform semantic analysis and time-series correlation feature extraction of each key image frame in the glue stirring state monitoring video, so as to automatically control the stirring process based on the time-series change of the glue stirring state, so as to ensure the quality, performance and peeling characteristics of the PSA glue, while avoiding the problem of energy waste caused by over-stirring. In this way, the polarizer PSA glue is suitable for laminating silk screen ink black frame glass, which solves the undesirable problems of stratification, debonding and bubbles in polarizer lamination.

[0101] In one embodiment of the present application, a polarizer PSA glue for screen-printed ink glass is provided, characterized in that the polarizer PSA glue for screen-printed ink glass is prepared by the preparation method of the polarizer PSA glue for screen-printed ink glass as described.

[0102] In one embodiment of the present application, Figure 7 FIG. 1 is a block diagram of a system for preparing polarizer PSA glue for screen printing ink glass according to an embodiment of the present application. Figure 7 As shown, according to an embodiment of the present application, a preparation system 200 of polarizer PSA glue for screen-printed ink glass includes: a main agent dilution module 210, used to dilute the PSA main agent to obtain a diluted PSA main agent, wherein the PSA main agent is an acrylate with a molecular weight between 900,000 and 1.3 million; and a polarizer PSA glue generation module 220, used to add additives to the diluted PSA main agent, wherein the additives include a bridging agent, a silane coupling agent and a stripping agent to obtain a polarizer PSA glue for screen-printed ink glass.

[0103] Here, those skilled in the art can understand that the specific functions and operations of each unit and module in the above-mentioned preparation system of polarizer PSA glue for screen printing ink glass have been referred to above. Figures 1 to 6 The preparation method of the polarizer PSA glue for screen printing ink glass has been described in detail, and therefore, its repeated description will be omitted.

[0104] As described above, the preparation system 200 of polarizer PSA glue for screen-printed ink glass according to the embodiment of the present application can be implemented in various terminal devices, such as a server for preparing polarizer PSA glue for screen-printed ink glass. In one example, the preparation system 200 of polarizer PSA glue for screen-printed ink glass according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the preparation system 200 of polarizer PSA glue for screen-printed ink glass can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the preparation system 200 of polarizer PSA glue for screen-printed ink glass can also be one of the many hardware modules of the terminal device.

[0105] Alternatively, in another example, the system 200 for preparing polarizer PSA glue for silk-screen ink glass and the terminal device may also be separate devices, and the system 200 for preparing polarizer PSA glue for silk-screen ink glass may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0106] Figure 8 Schematic diagram of a method for preparing polarizer PSA glue for screen-printed ink glass according to an embodiment of the present application. Figure 8 As shown, in this application scenario, first, a monitoring video of the glue stirring state collected by a camera is obtained (for example, Figure 8 C); then, the acquired glue stirring state monitoring video is input to a server (for example, as Figure 8 In S) shown in , the server is capable of processing the glue stirring state monitoring video based on the preparation algorithm of the polarizer PSA glue for screen-printed ink glass to obtain a control instruction, and the control instruction is used to indicate whether to stop stirring.

[0107] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0108] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0109] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for preparing polarizer PSA glue for screen printing ink glass, characterized in that: include: Diluting the PSA main agent to obtain a diluted PSA main agent, wherein the PSA main agent is an acrylate with a molecular weight between 900,000 and 1.3 million; as well as Additives are added dropwise into the diluted PSA main agent, wherein the additives include a bridging agent, a silane coupling agent and a stripping agent to obtain a polarizer PSA adhesive for screen printing ink glass.

2. The method for preparing polarizer PSA glue for screen printing ink glass according to claim 1, characterized in that: Diluting the PSA main agent to obtain a diluted PSA main agent includes: Pour the PSA raw rubber into a stainless steel bucket lined with a dust-free bag; Weigh a certain amount of ethyl acetate and pour it into the stainless steel barrel; and The mixture is stirred by a motor to obtain the diluted PSA main agent.

3. The method for preparing polarizer PSA glue for screen printing ink glass according to claim 2, characterized in that: Additives are added dropwise into the diluted PSA main agent, wherein the additives include a bridging agent, a silane coupling agent, and a stripping agent to obtain a polarizer PSA adhesive for screen printing ink glass, including: During the stirring of the diluted PSA main agent, adding the bridging agent dropwise; After the bridging agent is added dropwise, stirring is continued for a predetermined time; Continue to dropwise add the silane coupling agent; After the silane coupling agent is added dropwise, stirring is continued for a predetermined time; Continue to dropwise add the stripping agent; After the stripping agent is added dropwise, stirring is performed; and The mixture is allowed to stand for half an hour or more to obtain the polarizer PSA adhesive for screen printing ink glass.

4. The method for preparing polarizer PSA glue for screen printing ink glass according to claim 3, characterized in that: After the stripping agent is added dropwise, stirring is performed, including: Obtain the monitoring video of glue stirring status collected by the camera; Performing key frame sampling on the glue stirring state monitoring video to obtain a sequence of key image frames of the glue stirring state; Each glue stirring state key image frame in the sequence of glue stirring state key image frames is respectively passed through a stirring state feature extractor based on a convolutional neural network model to obtain a time series of glue stirring state feature vectors; Passing the time series of the glue stirring state feature vector through a content importance-based adaptive attention weight allocation network to obtain an adaptive time series of the glue stirring state feature vector; Passing the time series of the adaptive glue stirring state feature vector through a stirring state temporal reasoner based on stirring state prior to obtain a stirring state temporal reasoning semantic expression feature vector; and The stirring state temporal reasoning semantic expression feature vector is passed through a stirring state controller based on a classifier to obtain a control instruction, wherein the control instruction is used to indicate whether to stop stirring.

5. The method for preparing polarizer PSA glue for screen printing ink glass according to claim 4, characterized in that: The time series of the glue stirring state feature vector is passed through a content-importance-based adaptive attention weight allocation network to obtain a time series of the adaptive glue stirring state feature vector, including: the time series of the glue stirring state feature vector is processed through the content-importance-based adaptive attention weight allocation network with the following weight allocation formula to obtain the time series of the adaptive glue stirring state feature vector; Wherein, the weight distribution formula is: Among them, h i is the i-th glue stirring state feature vector in the sequence of glue stirring state feature vectors, and W i Represent the weight coefficient vector and weight coefficient matrix respectively, B i is the offset vector, Selu(·) represents the Selu function, e i is the attention score of the ith glue stirring state feature vector, λ and α are both hyperparameters, exp(·) represents exponential operation, e k is the attention score of the kth glue stirring state feature vector, t is the number of vectors in the sequence of the glue stirring state feature vector, a i is the weight value of the i-th glue stirring state feature vector, is the i-th adaptive glue stirring state feature vector in the sequence of the adaptive glue stirring state feature vectors.

6. The method for preparing polarizer PSA glue for screen printing ink glass according to claim 5, characterized in that: The time series of the adaptive glue stirring state feature vector is passed through a stirring state temporal reasoner based on stirring state prior to obtain a stirring state temporal reasoning semantic expression feature vector, including: the time series of the adaptive glue stirring state feature vector is passed through the stirring state temporal reasoner based on stirring state prior to obtain the stirring state temporal reasoning semantic expression feature vector by the following stirring state reasoning formula; Wherein, the stirring state reasoning formula is: Wherein, V1 is the adaptive glue stirring state feature vector of the last position in the time series of the adaptive glue stirring state feature vector, V rand is the adaptive glue stirring state feature vector at a random position in the time series of the adaptive glue stirring state feature vector, V * is the mean eigenvector of the time series of the adaptive glue stirring state eigenvector, p1 and p2 are weight hyperparameters, f1 and f2 are hyperparameters that use a Gaussian distribution random number function with a mean of 0 and a variance of 1 as the coefficient of the Gaussian distribution function, and are vector subtraction and vector addition respectively, and V is the semantic expression feature vector of the stirring state temporal reasoning.

7. The method for preparing polarizer PSA adhesive for screen printing ink glass according to claim 6, characterized in that: It also includes a training step: for training the stirring state feature extractor based on the convolutional neural network model, the content importance-based adaptive attention weight allocation network, the stirring state temporal reasoner based on stirring state prior, and the stirring state controller based on the classifier.

8. The method for preparing polarizer PSA adhesive for screen printing ink glass according to claim 7, characterized in that: The training step comprises: Obtain the training glue stirring state monitoring video collected by the camera; Performing key frame sampling on the training glue stirring state monitoring video to obtain a sequence of key image frames of the training glue stirring state; Each training glue stirring state key image frame in the sequence of training glue stirring state key image frames is respectively passed through the stirring state feature extractor based on the convolutional neural network model to obtain a time series of training glue stirring state feature vectors; Passing the time series of the training glue stirring state feature vector through the content importance-based adaptive attention weight allocation network to obtain the time series of the training adaptive glue stirring state feature vector; Passing the time series of the training adaptive glue stirring state feature vector through the stirring state temporal reasoner based on stirring state prior to obtain the training stirring state temporal reasoning semantic expression feature vector; Performing cluster optimization on the training stirring state temporal reasoning semantic expression feature vector to obtain an optimized training stirring state temporal reasoning semantic expression feature vector; Passing the optimized training stirring state temporal reasoning semantic expression feature vector through the classifier-based stirring state controller to obtain a classification loss function value; Based on the classification loss function value, the stirring state feature extractor based on the convolutional neural network model, the content importance-based adaptive attention weight allocation network, the stirring state temporal reasoner based on stirring state prior, and the classifier-based stirring state controller are trained.

9. The method for preparing polarizer PSA glue for screen printing ink glass according to claim 8, characterized in that: Clustering optimization is performed on the training stirring state temporal reasoning semantic expression feature vector to obtain an optimized training stirring state temporal reasoning semantic expression feature vector, including: Clustering each eigenvalue of the training stirring state temporal reasoning semantic expression feature vector based on the distance between eigenvalues ​​to obtain a cluster feature set; Based on the clustering feature set, the training stirring state temporal reasoning semantic expression feature vector is optimized based on the clustered feature in-class and out-of-class representation to obtain the optimized training stirring state temporal reasoning semantic expression feature vector.

10. A polarizer PSA adhesive for screen printing ink glass, characterized in that: The polarizer PSA adhesive for screen-printed ink glass is prepared by the preparation method of the polarizer PSA adhesive for screen-printed ink glass as described in any one of claims 1 to 9.

Images