Preparation method of high-precision ultrathin glass substrate

Abstract

The invention provides a preparation method of a high-precision ultra-thin glass substrate, and relates to the field of glass preparation, and the preparation method comprises the following steps: obtaining a surface feature vector of each sampling area of a to-be-treated glass substrate; if the number of the sampling areas of which the classification labels are key labels is greater than a preset threshold value, acquiring a preset classification vector with the highest matching degree with the surface feature vector of each sampling area from a plurality of preset classification vectors as a target classification vector corresponding to the sampling area; according to the influence fluctuation value of the target classification vector corresponding to each sampling area, obtaining a normalized weight of each sampling area; and according to the surface feature vector and the normalized weight of each sampling area, obtaining a fusion feature vector of the to-be-processed glass substrate. Compared with the target pressure curve obtained by the fusion feature vector which is directly fused without setting the weight, the target pressure curve obtained by the method is more adaptive to the to-be-processed glass substrate, so that the processing effect of the to-be-processed glass substrate is better.

Description

Background Art

[0002] In the glass manufacturing industry, glass processing is a key process to improve the performance and quality of glass, and its importance is self-evident. Different processing steps such as fine grinding, rough polishing, and fine polishing play a decisive role in the final flatness, smoothness, and strength of the glass, and precise control of pressure is the core point.

[0003] At present, in the technical application of glass processing, a fixed pressure value is usually set in each processing link. Although this method takes into account the differences between different links, there are significant defects in the setting of fixed pressure values ​​in a single processing link. In the same processing link, as the processing time progresses, the processing tool will gradually wear, and its contact state with the glass and cutting ability will continue to change. If a fixed pressure value is always used, in the early stage of processing tool wear, it may not be able to fully exert its cutting effect, resulting in low processing efficiency; and in the later stage, due to severe tool wear, the same fixed pressure may make the processing force insufficient and fail to achieve the expected processing effect. In addition, during the processing of glass, the material removal and microstructure of its surface continue to change, and a single fixed pressure value cannot flexibly adapt to these changes. For example, in the fine grinding stage, the glass surface gradually becomes smooth from the initial rough state. If the pressure is always fixed, it may cause excessive processing of the flat surface in the later stage, damage the surface quality, and affect the optical properties and mechanical strength of the glass. Therefore, it is urgent to develop a more advanced pressure control technology to optimize the glass processing process and meet the growing demand for high-quality glass production. Summary of the invention

[0004] In response to the above technical problems, the present application provides a method for preparing a high-precision ultra-thin glass substrate, which at least partially solves the problems existing in the prior art.

[0005] In a first aspect of the present application, a method for preparing a high-precision ultra-thin glass substrate is provided, the method comprising: Obtaining a surface feature vector of each sampling area of ​​the glass substrate to be processed; wherein each sampling area has a classification label; the classification label represents the degree of influence of the pressure curve category on the performance parameter of the sampling area; the classification label includes a key label and a non-key label; the glass substrate to be processed is a glass substrate before being processed in the light pressure stage; If the number of sampling areas whose classification labels are key labels is greater than a preset threshold, a preset classification vector with the highest matching degree with the surface feature vector of each sampling area is obtained from a number of preset classification vectors as a target classification vector corresponding to the sampling area; each glass sample corresponding to each preset classification vector is a qualified glass sample; each glass sample corresponding to each preset classification vector has an influence fluctuation value; the influence fluctuation values ​​corresponding to any two preset classification vectors are different; the influence fluctuation value represents the degree to which the performance parameters of the glass sample are affected by the preset pressure curve category; According to the influence fluctuation value of the target classification vector corresponding to each sampling area, the normalized weight of each sampling area is obtained; wherein the influence fluctuation value is proportional to the normalized weight; According to the surface feature vector and normalized weight of each sampling area, a fused feature vector of the glass substrate to be processed is obtained; wherein the fused feature vector is used to determine the target pressure curve list with the highest matching degree for the glass substrate to be processed in the pressure curve list corresponding to several historical qualified glass samples, and the glass substrate to be processed is prepared based on the target pressure curve list.

[0006] This application has at least the following beneficial effects: The high-precision ultra-thin glass substrate preparation method provided by the present application firstly performs a rough classification of a key label / non-key label on the surface feature vector corresponding to each sampling area to determine whether it is necessary to assign different weights to each sampling area for feature fusion later; secondly, when the number of sampling areas with the classification label as the key label is greater than the preset threshold, that is, when most of the sampling areas in the glass substrate to be processed are sensitive to the pressure curve, a plurality of preset classification vectors are used to assign the most suitable influence fluctuation value to each sampling area, because, after the rough classification of the classification label in the first step, it is only possible to preliminarily know whether each sampling area is sensitive to the pressure curve category, and further processing can obtain the sensitivity of each sampling area to different pressure curves. Then, the influence fluctuation value of the target classification vector corresponding to each sampling area is used as the weight of each sampling area, and then normalized so that the sum of the normalized weights of all sampling areas corresponding to the glass substrate to be processed is 1, and finally, the surface feature vector corresponding to each sampling area of ​​the glass substrate to be processed is fused to obtain the fused feature vector of the glass substrate to be processed. The present application determines a weight for each sampling area, that is, when fusion is performed, a corresponding weight is assigned to the surface feature vector corresponding to each sampling area and then fused, so that the fused feature vector obtained can more comprehensively and objectively highlight the sensitivity to different pressure curves. Furthermore, according to the obtained fused feature vector, the target pressure curve with the highest matching degree with the glass substrate to be processed is determined in the list of pressure curves corresponding to several historical qualified glass samples. The target pressure curve obtained by the present application is more suitable for the glass substrate to be processed than the target pressure curve obtained by directly fusion of the fused feature vector without setting weights, so that the processing effect of the glass substrate to be processed is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments 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 creative work.

[0008] Figure 1 Flow chart of the method for preparing a high-precision ultra-thin glass substrate provided in an embodiment of the present application. DETAILED DESCRIPTION

[0009] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0010] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable 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. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0011] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present application, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0012] Please refer to Figure 1 As shown, an embodiment of the present application provides a method for preparing a high-precision ultra-thin glass substrate, the method comprising: S100, obtaining the surface feature vector of each sampling area of ​​the glass substrate to be processed; wherein each sampling area has a classification label; the classification label characterizes the influence degree of the pressure curve category on the performance parameters of the sampling area; the classification label includes a key label and a non-key label; the glass substrate to be processed is a glass substrate before the light pressure stage treatment.

[0013] Specifically, the glass substrate to be processed may have several sampling areas, wherein infrared images of each sampling area of ​​the glass substrate to be processed are obtained respectively, and CNN is used to extract features of the infrared images of each sampling area, so as to obtain the surface feature vector corresponding to each sampling area. And the glass substrate to be processed is a glass substrate before being processed in the light pressure stage. Here. In the preparation process of high-precision ultra-thin glass substrates, starting from the light pressure stage, the fine grinding, rough polishing, and fine polishing links of the light pressure stage; the fine grinding, rough polishing, and fine polishing links of the heavy pressure stage; the fine grinding, rough polishing, and fine polishing links of the repair stage all require pressure.

[0014] As an example: if the glass substrate to be processed is a square glass substrate, its sampling area can be the areas corresponding to the four corners and the area corresponding to the geometric center. The sum of the sampling areas is smaller than the area of ​​the entire glass substrate to be processed. When selecting the sampling areas, they should be relatively evenly distributed at different positions of the glass substrate, and they should be able to represent the comprehensive surface characteristics of the entire glass substrate.

[0015] In addition, each sampling area has a corresponding classification label, where the classification label includes a key label and a non-key label, wherein the classification label can be obtained through a classification model, and the classification label represents the degree of influence of the pressure curve category of the sampling area on the performance parameters of the sampling area when the sampling area is processed. Further, the key label indicates that the pressure curve category of the sampling area has a greater influence on the performance parameters of the sampling area when the sampling area is processed; conversely, the non-key label indicates that the pressure curve category of the sampling area has a smaller influence on the performance parameters of the sampling area when the sampling area is processed. That is, the key label indicates that when the sampling area is processed, different pressure curves are used, and the numerical values ​​of the performance parameters finally obtained by the sampling area are not much different; conversely, the non-key label indicates that when the sampling area is processed, different pressure curves are used, and the numerical values ​​of the performance parameters finally obtained by the sampling area are greatly different.

[0016] The pressure curve can be a number of pre-set pressure curves, which are used to set the grinding pressure in a grinding process according to the pressure curve instead of using a constant pressure. This is because the surface state of the glass continues to change as the grinding process proceeds. Using the pressure curve can accurately adjust the pressure according to the actual needs of different stages. Compared with a constant pressure value, the performance of the glass obtained by using the pressure curve is better.

[0017] It should be noted that the grinding process in this embodiment may include: fine grinding, rough polishing, and fine polishing in the light pressure stage; fine grinding, rough polishing, and fine polishing in the heavy pressure stage; fine grinding, rough polishing, and fine polishing in the repair stage, etc.

[0018] In this embodiment, a rough classification of key label / non-key label is performed on the surface feature vector corresponding to each sampling area to determine whether it is necessary to assign different weights to each sampling area for feature fusion later.

[0019] S200, if the number of sampling areas whose classification labels are key labels is greater than a preset threshold, then the preset classification vector with the highest matching degree with the surface feature vector of each sampling area is obtained from a number of preset classification vectors as the target classification vector corresponding to the sampling area; the glass samples corresponding to each preset classification vector are all qualified glass samples; the glass samples corresponding to each preset classification vector have an influence fluctuation value; the influence fluctuation values ​​corresponding to any two preset classification vectors are different; the influence fluctuation value characterizes the degree to which the performance parameters of the glass sample are affected by the preset pressure curve category.

[0020] Specifically, if the number of sampling areas whose classification labels are key labels is greater than a preset threshold, it means that most of the sampling areas in the glass substrate to be processed are sensitive to the pressure curve. Such sampling areas have a greater impact on the performance parameters of the glass after being processed using different pressure curves, while some sampling areas may be insensitive to the pressure curve, that is, the performance parameters of the glass obtained after such sampling areas are processed using different pressure curves are basically the same. At this time, it means that most areas of the glass sample to be tested are sensitive to the pressure curve, while some sampling areas may be insensitive to the pressure curve. At this time, a suitable target classification vector is matched for each sampling area. The target classification vector is one of the preset classification vectors. Here, each preset classification vector has a corresponding qualified glass sample. A qualified glass sample means that each performance parameter of the glass obtained after processing is qualified. The preset classification vector is a surface feature vector obtained by extracting features using CNN based on the infrared image of the corresponding qualified glass sample. It should be noted that in this embodiment, the size of the qualified glass sample can be the same as the size of a sampling area of ​​the glass substrate to be processed, so that the matching result obtained is more accurate.

[0021] In addition, each glass sample corresponding to a preset classification vector has an influence fluctuation value, where the influence fluctuation value represents the degree to which the performance parameters of the glass sample are affected by the preset pressure curve category. Here, as above, the larger the influence fluctuation, the greater the fluctuation in the performance parameters of the glass obtained after the qualified glass sample is processed using different pressure curves. Conversely, the smaller the influence fluctuation, the smaller the fluctuation in the performance parameters of the glass obtained after the qualified glass sample is processed using different pressure curves.

[0022] In the present embodiment, when the number of sampling areas whose classification labels are key labels is greater than a preset threshold, that is, when most of the sampling areas in the glass substrate to be processed are relatively sensitive to the pressure curve, a number of preset classification vectors are used to assign the most suitable influence fluctuation value to each sampling area. This is because, after the rough classification of the classification labels in the first step, it is only possible to preliminarily know whether each sampling area is sensitive to the pressure curve category, while further processing in the present embodiment can obtain the sensitivity of each sampling area to different pressure curves.

[0023] It should be noted that when the surface feature vector of each sampling area is matched with each preset classification vector, the Euclidean distance, cosine distance, etc. may be used to obtain the matching degree.

[0024] S300, obtaining a normalized weight of each sampling area according to the influence fluctuation value of the target classification vector corresponding to each sampling area; wherein the influence fluctuation value is proportional to the normalized weight.

[0025] Specifically, the influence fluctuation value of the target classification vector corresponding to each sampling area is used as the weight of each sampling area, and then normalized so that the sum of the normalized weights of all sampling areas corresponding to the glass substrate to be processed is 1.

[0026] S400, obtaining a fused feature vector of the glass substrate to be processed according to the surface feature vector and normalized weight of each sampling area; wherein the fused feature vector is used to determine a target pressure curve list with the highest matching degree for the glass substrate to be processed in a pressure curve list corresponding to several historical qualified glass samples, and the glass substrate to be processed is prepared based on the target pressure curve list.

[0027] Specifically, the surface feature vectors corresponding to each sampling area of ​​the glass substrate to be processed are fused to obtain a fused feature vector of the glass substrate to be processed. In the present embodiment, a weight is determined for each sampling area, that is, a corresponding weight is assigned to the surface feature vector corresponding to each sampling area during feature fusion and then fused, so that the fused feature vector obtained can more comprehensively and objectively highlight the sensitivity to different pressure curves. Furthermore, according to the obtained fused feature vector, the target pressure curve with the highest matching degree with the glass substrate to be processed is determined in the pressure curve list corresponding to several historical qualified glass samples. The target pressure curve obtained in the present embodiment is more suitable for the glass substrate to be processed than the target pressure curve obtained by directly fusion of the fused feature vector without setting weights. Finally, the glass substrate to be processed is prepared based on the target pressure curve list.

[0028] In an exemplary embodiment of the present application, after obtaining the surface feature vector corresponding to each sampling area of ​​the glass substrate to be processed, the method further includes: S500: If the number of sampling regions whose classification labels are key labels is equal to or less than a preset threshold, a fusion feature vector of the glass substrate to be processed is obtained according to the surface feature vector corresponding to each sampling region.

[0029] Specifically, if the number of sampling areas whose classification labels are key labels is equal to or less than a preset threshold, it means that most of the sampling areas in the glass substrate to be processed are relatively insensitive to the pressure curve, and such sampling areas have little effect on the performance parameters of the glass after being processed using different pressure curves. Therefore, in this embodiment, a fused feature vector is directly obtained. Then, a target pressure curve is determined for the glass substrate to be tested based on the fused feature vector.

[0030] In an exemplary embodiment of the present application, the fused feature vectors are used to determine a target pressure curve list with the highest matching degree for a glass substrate to be processed from a pressure curve list corresponding to a number of historical qualified glass samples according to the following steps: S410, obtaining the historical surface feature vectors corresponding to the historical qualified glass samples before the light pressing stage processing, so as to obtain the historical surface feature vector list LT = (LT1, LT2, ..., LT i ,…,LT n ), i=1, 2, ..., n, where n is the number of historical qualified glass samples; LT i is the historical surface feature vector corresponding to the i-th historical qualified glass sample before the light pressing stage; each historical qualified glass sample has a corresponding historical pressure curve list; the historical pressure curve list includes each pressure curve corresponding to each process link requiring pressure starting from the light pressing stage.

[0031] Specifically, the historical qualified glass sample can be the same size as the glass substrate to be processed, and the sampling area is also divided in the same way. The historical surface feature vector is obtained by extracting features from the infrared image of each sampling area of ​​the corresponding historical qualified glass sample before the light pressure stage using CNN and then performing feature fusion. Each historical qualified glass sample has a corresponding historical pressure curve list. Here, each historical qualified glass sample may have several corresponding pressure curve lists so that its final performance parameters can be qualified. In this embodiment, the historical pressure curve list corresponding to each historical qualified glass sample can be the best performance parameter corresponding to all the pressure curve lists corresponding to the historical qualified glass sample so that its final performance parameters can be qualified. The historical pressure curve list includes each pressure curve corresponding to each process link that requires pressure starting from the light pressure stage. That is, each of the above-mentioned process links that require pressure has a corresponding pressure curve.

[0032] It should be noted that, here, the performance parameters may include substrate parallelism, flatness, finish and roughness. Among them, the closer the substrate parallelism is to zero, the more uniform and stable the light refraction and propagation of the glass in optical applications, such as in precision optical instruments, high parallelism substrates can ensure imaging quality and reduce aberrations. If the parallelism deviation is large, light propagation will produce deviations, affecting the use of glass in the optical field. The closer the flatness is to zero, the smoother the surface. In fields such as semiconductor manufacturing that require extremely high flatness, glass substrates with high flatness can ensure the accuracy of processes such as lithography and improve the yield rate of chip manufacturing. If the flatness is poor, it will cause problems such as uneven coating and deformation of lithography patterns. The higher the finish is usually, the better. High finish means that the glass surface is smooth and flawless, which can reduce light scattering and improve the light transmittance and reflection performance of the glass. In applications such as optical lenses and display screens, high-finish glass can present clearer images and better visual effects. Generally speaking, the smaller the roughness, the better. Roughness is the opposite of smoothness. Small roughness indicates that the microscopic fluctuations on the glass surface are small, which can improve the mechanical strength of the glass and reduce the risk of breakage caused by stress concentration due to surface defects. At the same time, in terms of optical performance, it can also avoid problems such as diffuse reflection of light caused by rough surfaces.

[0033] It is understandable that, for the glass substrate to be processed, the criteria for judging the quality of the performance parameters can be adjusted according to its corresponding use, but the numerical value of each performance parameter can be used for quantitative judgment.

[0034] S420, clustering LT to obtain a vector cluster list YJ=(YJ1, YJ2, ..., YJ p , …, YJ q ), p = 1, 2, ..., q, where q is the number of vector clusters; YJ p is the pth vector cluster; the matching degree between any two historical surface feature vectors in the same vector cluster is greater than the preset matching degree threshold; YJ p =(YJ p，1 , Y.J. p，2 , …, YJ p，a , …, YJ p，f（p） ); a = 1, 2, ..., f (p); f (p) is the number of historical surface feature vectors contained in the p-th vector cluster; YJ p，a is the ath historical surface feature vector contained in the pth vector cluster; each vector cluster has a corresponding marker vector; the marker vector is the center vector of the corresponding vector cluster; each vector cluster has a corresponding key historical pressure curve list; the key historical pressure curve list is a historical pressure curve list corresponding to the historical qualified glass samples with the smallest comprehensive difference with the standard performance parameters in the corresponding vector cluster.

[0035] Specifically, clustering is performed according to the historical surface feature vector corresponding to each historical qualified glass sample among a number of historical qualified glass samples, and each obtained vector cluster contains one type of historical qualified glass samples, that is, the surface features of the historical qualified glass samples in the same vector cluster are similar.

[0036] The central vector can be the average vector of all historical surface feature vectors in the corresponding vector clustering cluster, so it can be the historical surface feature vector with the highest matching degree with the corresponding average vector in the vector clustering cluster. The key historical pressure curve list is a list of historical pressure curves corresponding to the historical qualified glass samples with the smallest comprehensive difference with the standard performance parameters in the corresponding vector clustering cluster. The standard performance parameter is the normalized sum of the standard values ​​of each performance parameter corresponding to the substrate to be tested. The comprehensive difference refers to the total difference obtained after normalizing the difference between the values ​​of all performance parameters and the corresponding standard values. Here, the purpose of normalization is to unify the dimensions between the numerical values ​​of each indicator. So that the obtained comprehensive difference is more referenceable. That is, the key historical pressure curve list is a list of historical pressure curves corresponding to the historical qualified glass samples with the best comprehensive performance parameters in the corresponding vector clustering cluster.

[0037] S430, according to DT and YJ, obtain the first matching list YP=(YP1, YP2, ..., YP p , …, YP q ), among which, YP p BYT is the symbol vector of the p-th vector cluster p The matching degree with DT; DT is the fused feature vector of the glass substrate to be processed.

[0038] Specifically, YP p =(BYT p ·DT) / (|BYT p |×|DT|).

[0039] S440, determining MAX(YP) as the target vector cluster.

[0040] S450: Determine the key historical pressure curve list corresponding to the target vector cluster as the target pressure curve list.

[0041] In this implementation, in order to reduce the amount of data processing, historical qualified glass samples with similar surface features are first clustered through clustering. Then, each vector clustering cluster has a corresponding marker vector for matching with the glass substrate to be processed, and each vector clustering cluster also has a list of historical pressure curves corresponding to the historical qualified glass samples with the best comprehensive performance parameters as the representative curve list of the clustering cluster. In this way, surface feature matching between the glass substrate to be processed and each historical qualified glass sample is avoided, the amount of data processing is reduced, and data processing time is saved.

[0042] In an exemplary embodiment of the present application, after determining MAX(YP) as the target vector cluster, the method further includes: S460, obtaining the matching degree between DT and each historical surface feature vector in the target vector cluster, so as to obtain a second matching degree list.

[0043] S470: Determine the historical surface feature vector corresponding to the maximum matching degree in the second matching degree list as the target historical surface feature vector.

[0044] S480: Determine the historical pressure curve list corresponding to the target historical surface feature vector as the target pressure curve list.

[0045] Specifically, in this embodiment, in order to reduce the amount of data processing, firstly, the historical qualified glass samples with similar surface features are clustered by clustering, and then, each vector clustering cluster has a corresponding marker vector, which is used to match with the glass substrate to be processed. After determining the target vector clustering cluster, it is matched again with each historical surface feature vector in the target vector clustering cluster, and finally the historical surface feature vector corresponding to the largest matching degree in the second matching degree list is determined as the target historical surface feature vector. The target historical surface feature vector determined in this way is more compatible with the surface vector of the glass substrate to be processed, and this embodiment avoids full matching of all historical surface feature vectors through two matchings and one clustering, thereby reducing the amount of data processing and improving the data processing speed.

[0046] In an exemplary embodiment of the present application, the influence fluctuation value of the glass sample corresponding to the preset classification vector is determined according to the following steps: S210, obtaining a qualified pressure curve list set HL=(HL1, HL2, ..., HL x ,…,HL y ); x = 1, 2, ..., y; where y is the number of qualified pressure curve lists corresponding to YF; HL x is the curve identifier of the xth qualified pressure curve list corresponding to YF; the qualified pressure curve list includes each reasonable pressure curve corresponding to the process link requiring pressure starting from the light pressure stage; the reasonable pressure curve is the pressure curve that makes the corresponding glass sample a qualified glass sample after processing; S220, according to HL and each performance parameter, obtain the impact fluctuation value YFB corresponding to YF; wherein YFB meets the following conditions: YFB=(Σ m j=1 ZB j ) / m; Where m is the number of performance parameters; ZBj is the index impact fluctuation value corresponding to the jth performance parameter; j=1, 2, …, m; ZB j =(Σ y x=1 (C j,x- avg(C j ) 2 ) / y; C j is a list of index values ​​corresponding to j performance parameters obtained by processing the qualified glass samples corresponding to YF using each qualified pressure curve list in HL; C j =(C j，1 , C j，2 , …, C j，x , …, C j，y ); C j,x To use HL x The index values ​​of j performance parameters obtained after processing YF; avg() is the preset average value determination function.

[0047] In this embodiment, the preset classification vector YF has a corresponding qualified pressure curve list. Although each qualified pressure curve list can make the corresponding glass sample qualified, the performance parameters of the glass samples finally obtained after being processed by different qualified pressure curve lists are either just at the passing line or excellent. Obtain the influence fluctuation value of each performance parameter corresponding to YF. Here, if ZB j The larger the value, the more sensitive the jth performance parameter is to the changes in different pressure curve lists; conversely, if ZB j The smaller it is, the less sensitive the j-th performance parameter is to the changes in different pressure curve lists, that is, the smaller the difference between the values ​​of the j-th performance parameter for different pressure curves. Thus, the sensitivity of each performance parameter to the changes in different pressure curve lists is obtained, and finally the influence fluctuation value YFB corresponding to YF is obtained, that is, the influence fluctuation value of each performance parameter is integrated. Here, the influence fluctuation value is determined according to the corresponding variance formula, which can reflect the fluctuation between the values ​​of the performance parameters obtained after processing according to different curve lists.

[0048] The impact fluctuation value determined in this embodiment can reflect the sensitivity of the corresponding sampling area to different pressure curve lists, thereby improving the accuracy of subsequent matching and the accuracy of the target pressure curve list finally determined.

[0049] In an exemplary embodiment of the present application, an electronic device capable of implementing the above method is also provided.

[0050] Those skilled in the art will appreciate that various aspects of the present application may be implemented as a system, method or program product. Therefore, various aspects of the present application may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to as "circuit", "module" or "system" herein.

[0051] The electronic device according to this embodiment of the present application is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0052] The electronic device is presented in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: the at least one processor mentioned above, the at least one storage device mentioned above, and a bus connecting different system components (including storage devices and processors).

[0053] The storage stores program codes, which can be executed by the processor, so that the processor executes the steps described in the above “Exemplary Method” section of this specification according to various exemplary embodiments of the present application.

[0054] The memory may include readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory, and may further include read only memory (ROM).

[0055] The storage may also include a program / utility having a set (at least one) of program modules, such program modules including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0056] The bus may represent one or more of several types of bus structures including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures.

[0057] The electronic device may also communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may communicate with one or more devices that enable a user to interact with the electronic device, and / or may communicate with any device (e.g., routers, modems, etc.) that enables the electronic device to communicate with one or more other computing devices. This communication may be performed through an input / output (I / O) interface. In addition, the electronic device may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter. As shown in the figure, the network adapter communicates with other modules of the electronic device through a bus. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0058] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solution according to the implementation method of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation method of the present application.

[0059] In an exemplary embodiment of the present application, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present application can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary implementations of the present application described in the above "Exemplary Method" section of the present specification.

[0060] The program product may adopt any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0061] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0062] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0063] Program code for performing the operations of the present application may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0064] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present application, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0065] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0066] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for preparing a high-precision ultra-thin glass substrate, characterized in that: include: Obtaining a surface feature vector of each sampling area of ​​the glass substrate to be processed; wherein each sampling area has a classification label; the classification label represents the degree of influence of the pressure curve category on the performance parameter of the sampling area; the classification label includes a key label and a non-key label; the glass substrate to be processed is a glass substrate before being processed in the light pressure stage; If the number of sampling areas whose classification labels are key labels is greater than a preset threshold, a preset classification vector with the highest matching degree with the surface feature vector of each sampling area is obtained from a number of preset classification vectors as a target classification vector corresponding to the sampling area; each glass sample corresponding to each preset classification vector is a qualified glass sample; each glass sample corresponding to each preset classification vector has an influence fluctuation value; the influence fluctuation values ​​corresponding to any two preset classification vectors are different; the influence fluctuation value represents the degree to which the performance parameters of the glass sample are affected by the preset pressure curve category; According to the influence fluctuation value of the target classification vector corresponding to each sampling area, the normalized weight of each sampling area is obtained; wherein the influence fluctuation value is proportional to the normalized weight; According to the surface feature vector and normalized weight of each sampling area, a fused feature vector of the glass substrate to be processed is obtained; wherein the fused feature vector is used to determine the target pressure curve list with the highest matching degree for the glass substrate to be processed in the pressure curve list corresponding to several historical qualified glass samples, and the glass substrate to be processed is prepared based on the target pressure curve list.

2. The method for preparing a high-precision ultra-thin glass substrate according to claim 1, characterized in that: After obtaining the surface feature vector corresponding to each sampling area of ​​the glass substrate to be processed, the method further includes: If the number of sampling regions whose classification labels are key labels is equal to or less than a preset threshold, a fused feature vector of the glass substrate to be processed is obtained according to the surface feature vector corresponding to each sampling region.

3. The method for preparing a high-precision ultra-thin glass substrate according to claim 1, characterized in that: The fused feature vector determines the target pressure curve list with the highest matching degree for the glass substrate to be processed from the pressure curve lists corresponding to several historical qualified glass samples according to the following steps: Obtain the historical surface feature vectors corresponding to the historical qualified glass samples before the light pressing stage treatment to obtain the historical surface feature vector list LT = (LT1, LT2, ..., LT i ,…,LT n ), i=1, 2, ..., n, where n is the number of historical qualified glass samples; LT i is the historical surface feature vector corresponding to the i-th historical qualified glass sample before the light pressure stage processing; each historical qualified glass sample has a corresponding historical pressure curve list; the historical pressure curve list includes each pressure curve corresponding to each process link requiring pressure starting from the light pressure stage; Cluster LT to obtain the vector cluster list YJ=(YJ1, YJ2, …, YJ p , …, YJ q ), p = 1, 2, ..., q, where q is the number of vector clusters; YJ p is the pth vector cluster; the matching degree between any two historical surface feature vectors in the same vector cluster is greater than the preset matching degree threshold; YJ p =(YJ p，1 , Y.J. p，2 , …, YJ p，a , …, YJ p，f（p） ); a = 1, 2, ..., f (p); f (p) is the number of historical surface feature vectors contained in the p-th vector cluster; YJ p，a is the ath historical surface feature vector contained in the pth vector cluster; each vector cluster has a corresponding sign vector; the sign vector is the center vector of the corresponding vector cluster; each vector cluster has a corresponding key historical pressure curve list; the key historical pressure curve list is a historical pressure curve list corresponding to the historical qualified glass sample with the smallest comprehensive difference between the standard performance parameters and the corresponding vector cluster; According to DT and YJ, the first matching list YP=(YP1, YP2, ..., YP p , …, YP q ); among them, YP p BYT is the symbol vector of the p-th vector cluster p The matching degree with DT; DT is the fusion feature vector of the glass substrate to be processed; Determine MAX(YP) as the target vector cluster; The key historical pressure curve list corresponding to the target vector cluster is determined as the target pressure curve list.

4. The method for preparing a high-precision ultra-thin glass substrate according to claim 3, characterized in that: YP p Meet the following characteristics: YP p =(BYT p ·DT) / (|BYT p |×|DT|)。 5. The method for preparing a high-precision ultra-thin glass substrate according to claim 3, characterized in that: After determining MAX(YP) as the target vector cluster, the method further includes: Obtaining the matching degree between DT and each historical surface feature vector in the target vector clustering cluster; to obtain a second matching degree list; Determine the historical surface feature vector corresponding to the largest matching degree in the second matching degree list as the target historical surface feature vector; The historical pressure curve list corresponding to the target historical surface feature vector is determined as the target pressure curve list.

6. The method for preparing a high-precision ultra-thin glass substrate according to claim 1, characterized in that: The influence fluctuation value of the glass sample corresponding to the preset classification vector is determined according to the following steps: Get the qualified pressure curve list set HL=(HL1, HL2, ..., HL x ,…,HL y ); x = 1, 2, ..., y; where y is the number of qualified pressure curve lists corresponding to YF; HL x is the curve identifier of the xth qualified pressure curve list corresponding to YF; the qualified pressure curve list includes each reasonable pressure curve corresponding to the process link requiring pressure starting from the light pressure stage; the reasonable pressure curve is the pressure curve that makes the corresponding glass sample a qualified glass sample after processing; According to HL and each performance parameter, the impact fluctuation value YFB corresponding to YF is obtained.

7. The method for preparing a high-precision ultra-thin glass substrate according to claim 6, characterized in that: YFB meets the following conditions: YFB=(Σ m j=1 ZB j ) / m; Where m is the number of performance parameters; ZB j is the index impact fluctuation value corresponding to the jth performance parameter; j=1, 2, …, m; ZB j =(Σ y x=1 (C j,x- avg(C j ) 2 ) / y; C j is a list of index values ​​corresponding to j performance parameters obtained by processing the qualified glass samples corresponding to YF using each qualified pressure curve list in HL; C j =(C j，1 , C j，2 , …, C j，x , …, C j，y ); C j,x To use HL x The index values ​​of j performance parameters obtained after processing YF; avg() is the preset average value determination function.

8. The method for preparing a high-precision ultra-thin glass substrate according to claim 6, characterized in that: Performance parameters include substrate parallelism, flatness, smoothness and roughness.

Images