Glass surface nano coating thickness control method and system

Abstract

The invention provides a glass surface nano coating thickness control method and system, belongs to the technical field of detection control, and is used for realizing low-cost and high-precision nano coating thickness detection control. The control equipment acquires a target image, the surface of the glass panel is coated with a nano coating, and the target image is obtained by shooting the nano coating on the surface of the glass panel under the condition that the surface of the nano coating is atomized; the control device discretely segments the target image into K sub-image sets; the control equipment analyzes the K subimage sets through an AI / ML model to obtain corresponding K analysis results, and fuses the K analysis results to obtain an analysis result of the target image, and the analysis result of the target image is an image of a region marked with the non-uniform nano coating on the glass panel; and the control equipment instructs the coating equipment to correct the nano coating of the non-uniform area.

Description

Technical Field

[0001] This application relates to the technical field of detection and control, and particularly to a method and system for controlling the thickness of a nano-coating on the surface of glass. Background Art

[0002] With the rapid development of materials science and surface engineering technology, the application of nano-coatings on the surface of glass has become increasingly widespread. For example, low-emissivity (Low-E) coatings in architectural glass, anti-fog / anti-ice coatings on automotive glass, and anti-reflection and hydrophobic coatings on the screens of electronic devices all rely on functional thin films with nano-scale thicknesses to achieve optimization of optical, electrical, or mechanical properties. The thickness of the nano-coating (usually in the range of 1 - 500 nm) directly affects key parameters such as its light transmittance, conductivity, weather resistance, and interfacial bonding strength. Therefore, the development of high-precision, high-efficiency, and non-destructive thickness detection technologies has become the core requirement for improving the quality of coating processes and the reliability of industrial applications.

[0003] Current detection methods include the following:

[0004] Spectroscopic Ellipsometry, which calculates the film thickness by analyzing the reflection characteristics of polarized light on the surface of the coating, is suitable for non-contact detection of transparent or semi-transparent thin films. However, its accuracy is limited by the accuracy of the optical model, it has poor adaptability to multi-layer composite coatings or non-uniform surface structures, and requires complex mathematical inversion algorithms, making it difficult to meet the requirements of on-line real-time detection.

[0005] Atomic Force Microscope (AFM) and Scanning Electron Microscope (SEM), which directly measure the film thickness through surface topography scanning or cross-section observation, can achieve a resolution of up to sub-nanometer level. However, such methods require destructive sample preparation (such as SEM requires cutting the cross-section and plating a conductive layer), with low detection efficiency and high cost, and cannot be used for quality control during the production process.

[0006] Therefore, how to achieve low-cost and high-precision detection and control of the thickness of nano-coatings is a current research issue. Summary of the Invention

[0007] Embodiments of this application provide a method and system for controlling the thickness of a nano-coating on the surface of glass to achieve low-cost and high-precision detection and control of the thickness of the nano-coating.

[0008] To achieve the above object, this application adopts the following technical solutions:

[0009] In a first aspect, an embodiment of the present application provides a method for controlling the thickness of a nano - coating on a glass surface. This method is applied to a control device and includes: The control device acquires a target image. The glass panel surface is coated with a nano - coating, and the target image is obtained by photographing the nano - coating on the glass panel surface under the condition that the surface of the nano - coating is atomized. The control device discretely divides the target image into K sub - image sets. Any two sub - image sets among the K sub - image sets contain different sub - images. K is an integer greater than 1. For the i - th sub - image set among the K sub - image sets, any two sub - images included therein are not adjacent in the position of the target image, and any two sub - images included in the i - th sub - image set have the same size. i is an integer taking values from 1 to K. The control device analyzes the K sub - image sets respectively through an AI / ML model to obtain corresponding K analysis results, and fuses the K analysis results to obtain an analysis result of the target image. The analysis result of the target image is an image marking the non - uniform areas of the nano - coating on the glass panel. The control device instructs the coating device to correct the nano - coating in the non - uniform areas.

[0010] Optionally, the control device discretely divides the target image into K sub - image sets, including: The control device discretely divides the target image into two sub - image sets according to the size of the target image being less than a threshold size.

[0011] Optionally, the control device discretely divides the target image into two sub - image sets according to the size of the target image being less than a threshold size, including: The control device divides the target image into M1*N1 grids through rasterization processing according to the size of the target image being less than a threshold size. M1 is the number of grids included in one row of the target image, N1 is the number of grids included in one column of the target image, and M1 and N1 are integers greater than or equal to 2. The control device determines the grids with odd column indices in each row with an odd row index among the M1*N1 grids as a sub - image belonging to the first sub - image set, and the grids with even column indices in each row with an even row index among the M1*N1 grids as a sub - image belonging to the first sub - image set. The control device determines the grids with even column indices in each row with an odd row index among the M1*N1 grids as a sub - image belonging to the second sub - image set, and the grids with odd column indices in each row with an even row index among the M1*N1 grids as a sub - image belonging to the second sub - image set. Among them, the two sub - image sets include the first sub - image set and the second sub - image set.

[0012] Optionally, the control device discretely divides the target image into K sub - image sets, including: The control device discretely divides the target image into three sub - image sets according to the size of the target image being greater than or equal to the threshold size.

[0013] Optionally, the control device discretely divides the target image into three sub-image sets according to the size of the target image being greater than or equal to a threshold size, including: the control device divides the target image into M2*N2 grids through rasterization processing according to the size of the target image being greater than or equal to the threshold size, where M2 is the number of grids included in one row of the target image, N2 is the number of grids included in one column of the target image, and M2 and N2 are integers greater than or equal to 3; the control device determines the grids with a column index starting from 1 and a column index step of 3 in the row with row index 1 among the M2*N2 grids, the grids with a column index starting from 2 and a column index step of 3 in the row with row index 2 among the M2*N2 grids, the grids with a column index starting from 3 and a column index step of 3 in the row with row index 3 among the M2*N2 grids, and the grids with a column index starting from 1 and a column index step of 3 in the row with row index 4 among the M2*N2 grids as a sub-image belonging to the first sub-image set, and then, and so on, until traversing to the row with row index N2 among the M2*N2 grids; the control device determines the grids with a column index starting from 2 and a column index step of 3 in the row with row index 1 among the M2*N2 grids, the grids with a column index starting from 3 and a column index step of 3 in the row with row index 2 among the M2*N2 grids, the grids with a column index starting from 4 and a column index step of 3 in the row with row index 3 among the M2*N2 grids, and the grids with a column index starting from 2 and a column index step of 3 in the row with row index 4 among the M2*N2 grids as a sub-image belonging to the second sub-image set, and then, and so on, until traversing to the row with row index N2 among the M2*N2 grids; the control device determines the grids with a column index starting from 3 and a column index step of 3 in the row with row index 1 among the M2*N2 grids, the grids with a column index starting from 4 and a column index step of 3 in the row with row index 2 among the M2*N2 grids, the grids with a column index starting from 5 and a column index step of 3 in the row with row index 3 among the M2*N2 grids, and the grids with a column index starting from 3 and a column index step of 3 in the row with row index 4 among the M2*N2 grids as a sub-image belonging to the second sub-image set, and then, and so on, until traversing to the row with row index N2 among the M2*N2 grids; where the three sub-image sets include the first sub-image set, the second sub-image set, and the third sub-image set.

[0014] Optionally, the control device analyzes the K sub-image sets through the AI / ML model respectively to obtain the corresponding K analysis results, and fuses the K analysis results to obtain the analysis result of the target image, including: for the i-th sub-image set, where i ranges from 1 to K, the control device randomly combines the sub-images in the i-th sub-image set in pairs to obtain P sub-image groups, and analyzes the j-th sub-image group in the P sub-image groups through the AI / ML model to obtain the analysis result of the j-th sub-image group, where j is an integer ranging from 1 to P, and the analysis result of the j-th sub-image group is two sub-images that label or do not label the non-uniform area of the nano-coating; the i-th analysis result among the K analysis results includes the analysis result of the j-th sub-image group; the control device splices the K analysis results in the order of discrete segmentation to obtain the analysis result of the target image.

[0015] Optionally, the control device randomly combines the sub-images in the i-th sub-image set in pairs to obtain P sub-image groups, and analyzes the j-th sub-image group in the P sub-image groups through the AI / ML model to obtain the analysis result of the j-th sub-image group, including: for the i-th sub-image set, where i ranges from 1 to K: the control device randomly combines the sub-images in the i-th sub-image set in pairs to obtain P sub-image groups; the control device inputs the j-th sub-image group into the feature extraction layer of the AI / ML model to obtain two feature sequences corresponding one by one to the two sub-images in the j-th sub-image group output by the feature extraction layer of the AI / ML model; the control device analyzes the two feature sequences through the feature processing layer of the AI / ML model to obtain the analysis result of the j-th sub-image group.

[0016] Optionally, the control device analyzes the two feature sequences through the feature processing layer of the AI / ML model to obtain the analysis result of the j-th sub-image group, including: the control device calculates the difference between each two feature elements at the corresponding positions in the two feature sequences, and sets the values of each two feature elements with a difference less than a preset threshold to default values in the two feature sequences to obtain two corrected feature sequences; the control device interleaves the feature elements of the two corrected feature sequences according to the energy level to obtain two interleaved feature sequences, and inputs the two interleaved feature sequences into the feature processing layer of the AI / ML model to obtain the analysis result of the j-th sub-image group output by the feature processing layer of the AI / ML model.

[0017] Optionally, the control device interleaves the feature elements of each of the two corrected feature sequences according to the energy level to obtain two interleaved feature sequences, including: for each of the two corrected feature sequences, if the value of a feature element in the corrected feature sequence is larger, the energy of the feature element is higher. On this basis, the control device interleaves the feature elements with the same energy level in the corrected feature sequence together in the order from high to low energy level to obtain two interleaved feature sequences.

[0018] In a second aspect, an embodiment of the present application provides a glass surface nano-coating thickness control system, which includes a control device and a coating device. The control device is configured to: The control device acquires a target image. The glass panel surface is coated with a nano-coating. The target image is obtained by photographing the nano-coating on the glass panel surface under the condition that the nano-coating surface is atomized. The control device discretely divides the target image into K sub-image sets. Any two sub-image sets in the K sub-image sets contain different sub-images. K is an integer greater than 1. For the i-th sub-image set in the K sub-image sets, any two sub-images included in the i-th sub-image set are not adjacent in the position of the target image, and any two sub-images included in the i-th sub-image set have the same size. i is an integer taking values from 1 to K. The control device analyzes at least two sub-image sets through an AI / ML model respectively to obtain corresponding at least two analysis results, and fuses the at least two analysis results to obtain an analysis result of the target image. The analysis result of the target image indicates the non-uniform area of the nano-coating in the glass panel. The control device instructs the coating device to correct the nano-coating in the non-uniform area.

[0019] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, on which program code is stored. When the program code is run by the computer, it executes the method described in the first aspect.

[0020] In summary, the above method and device have the following technical effects:

[0021] By atomizing the surface of the nano - coating, atomized water droplets can be distributed on the surface of the nano - coating. At this time, if the thickness of the nano - coating is uneven, it may lead to uneven distribution of the atomized water droplets, and there may be concentrated or sparse situations. In view of this characteristic, when atomizing the surface of the nano - coating, a target image can be obtained by photographing the nano - coating on the glass panel surface. Then, the device is controlled to discretely segment the target image into K sub - image sets, and then the AI / ML model is used to analyze the K sub - image sets respectively to obtain the corresponding K analysis results, and the K analysis results are fused to obtain the analysis result of the target image. In this way, the amount of data to be processed can be reduced, and processing with a single sub - image as the granularity can also improve the processing accuracy, so as to accurately indicate the area where the nano - coating on the glass panel is non - uniform. Finally, the control device instructs the coating device to correct the nano - coating in the non - uniform area, thereby realizing low - cost and high - precision detection and control of the nano - coating thickness. Description of the Drawings

[0022] Figure 1 Schematic diagram of the architecture of a detection and control system provided by an embodiment of the present application;

[0023] Figure 2 Flowchart of a method for controlling the thickness of a nano - coating on the glass surface provided by an embodiment of the present application;

[0024] Figure 3 Schematic diagram of a scenario of a method for controlling the thickness of a nano - coating on the glass surface provided by an embodiment of the present application;

[0025] Figure 4 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Detailed Embodiments

[0026] 1. Deep Neural Network (DNN):

[0027] DNN is a specific implementation form of machine learning. According to the universal approximation theorem, a neural network can theoretically approximate any continuous function, so that the neural network has the ability to learn any mapping. According to the construction method of the network, DNN can be divided into feed - forward neural network (FNN), convolutional neural network (CNN), and recurrent neural network (RNN).

[0028] The FNN network is characterized in that neurons in adjacent layers are fully connected pairwise, which usually requires a large amount of storage space and leads to a high computational complexity for the FNN.

[0029] A CNN is a neural network specifically designed to process data with a similar grid structure. For example, time series data (discretely sampled on the time axis) and image data (two-dimensionally discretely sampled) can both be considered data with a similar grid structure. Instead of using all the input information for calculation at once, the CNN uses a window of a fixed size to intercept part of the information for convolution operations, which greatly reduces the computational amount of model parameters. Additionally, depending on the type of information intercepted by the window (such as people and objects in the same picture being different types of information), different convolution kernels can be used for each window, enabling the CNN to better extract the features of the input data.

[0030] An RNN is a type of DNN network that utilizes feedback time series information. Its input includes the new input value at the current moment and its own output value at the previous moment. The RNN is suitable for obtaining sequence features that are relevant in time and is particularly applicable to applications such as speech recognition and channel coding and decoding.

[0031] In the embodiments of the present invention, "indication" may include direct indication and indirect indication, and may also include explicit indication and implicit indication. If the information indicated by a certain piece of information is called the information to be indicated, then in the specific implementation process, there are many ways to indicate the information to be indicated. For example, but not limited to, the information to be indicated can be directly indicated, such as the information to be indicated itself or the index of the information to be indicated, etc. It is also possible to indirectly indicate the information to be indicated by indicating other information, where there is an association relationship between the other information and the information to be indicated. It is also possible to only indicate a part of the information to be indicated, while the other parts of the information to be indicated are known or pre-agreed. For example, it is also possible to use the pre-agreed (such as protocol-specified) arrangement order of each piece of information to achieve the indication of specific information, thereby reducing the indication overhead to a certain extent. At the same time, the common parts of each piece of information can be identified and indicated uniformly to reduce the indication overhead caused by separately indicating the same information.

[0032] In addition, the specific indication method can also be various existing indication methods, such as but not limited to, the above-mentioned indication methods and their various combinations, etc. The specific details of various indication methods can refer to the prior art and will not be elaborated herein. As can be seen from the above, for example, when it is necessary to indicate multiple pieces of information of the same type, there may be a situation where the indication methods for different pieces of information are different. In the specific implementation process, the required indication method can be selected according to specific needs. The embodiments of the present invention do not limit the selected indication method. In this way, the indication methods involved in the embodiments of the present invention should be understood to cover various methods that can enable the party to be indicated to obtain the information to be indicated.

[0033] It should be understood that the information to be indicated can be sent as a whole or divided into multiple sub-information and sent separately. Moreover, the sending periods and / or sending timings of these sub-information can be the same or different. The specific sending method is not limited in the embodiments of the present invention. Among them, the sending periods and / or sending timings of these sub-information can be predefined, for example, predefined according to a protocol, or can be configured by the sending device by sending configuration information to the receiving device.

[0034] "Predefined" or "pre-configured" can be achieved by pre-saving corresponding codes, tables or other ways that can be used to indicate relevant information in the device. The embodiments of the present invention do not limit its specific implementation method. Among them, "saving" can mean saving in one or more memories. The one or more memories can be set separately, or integrated in an encoder or decoder, a processor, or an electronic device. The one or more memories can also be partly set separately and partly integrated in a decoder, a processor, or an electronic device. The type of the memory can be any form of storage medium, which is not limited in the embodiments of the present invention.

[0035] The "protocol" involved in the embodiments of the present invention can refer to a protocol family in the communication field, a standard protocol with a frame structure similar to that of a protocol family, or a relevant protocol in a reliable access method system applied to future Internet of Things devices. The embodiments of the present invention do not make specific limitations on this.

[0036] In the embodiments of the present invention, descriptions such as "when...", "in the case of...", "if", and "when" all refer to that the device will make corresponding processing under a certain objective situation, which does not limit time, and does not require the device to have a judgment action when implemented, nor does it mean the existence of other limitations.

[0037] In the description of the embodiments of the present invention, unless otherwise specified, " / " indicates that the objects associated before and after are in an "or" relationship. For example, A / B may represent A or B. The "and / or" in the embodiments of the present invention is merely a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. Also, in the description of the embodiments of the present invention, unless otherwise specified, "a plurality of" means two or more than two. "At least one (item)" or its similar expression refers to any combination of these items, including any combination of a single item or plural items. For example, at least one (item) of a, b, or c may represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple. Additionally, for the convenience of clearly describing the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, terms such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and roles. Those skilled in the art can understand that terms such as "first" and "second" do not limit the quantity and execution order, and "first" and "second" do not necessarily mean different. At the same time, in the embodiments of the present invention, words such as "exemplary" or "for example" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly, using words such as "exemplary" or "for example" aims to present relevant concepts in a specific way for easy understanding.

[0038] The network architecture and service scenarios described in the embodiments of the present invention are for more clearly illustrating the technical solutions of the embodiments of the present invention, and do not constitute a limitation on the technical solutions provided by the embodiments of the present invention. Those of ordinary skill in the art know that with the evolution of the network architecture and the emergence of new service scenarios, the technical solutions provided by the embodiments of the present invention are equally applicable to similar technical problems.

[0039] The technical solutions in the present application will be described below with reference to the accompanying drawings.

[0040] Please refer to Figure 1 , the embodiments of the present application provide a detection control system, and the detection control system may include: a control device and a coating device.

[0041] Among them, both the control device and the coating device can be understood as terminals. The terminal can be a terminal with communication functions, or can be a chip or chip system disposed in the terminal. This terminal device can also be referred to as a user equipment (UE), access terminal, user unit, user station, mobile station, mobile platform, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device. The terminal device in the embodiments of the present application can be a mobile phone, a tablet computer (Pad), or a computer with wireless transceiver functions. Alternatively, it can also be a server, a server cluster, a virtual server, or a virtual server cluster.

[0042] This will be described below in conjunction with the method.

[0043] Please refer to Figure 2 , the embodiments of the present application provide a method for controlling the thickness of a nano-coating on the glass surface. This method is applied to the above detection and control system, and the process of this method includes:

[0044] S201, the control device acquires a target image.

[0045] The glass panel is coated with a nano-coating. The glass panel can be a rectangular panel, or the present application takes a rectangle as an example for introduction.

[0046] Depending on the application scenario, the specific implementation of the nano-coating may also be different. For example, it can be metal oxide, polymer, carbon-based composite material, layer composite structure (such as gradient refractive index coating), etc., without specific limitation. The nano-coating in the embodiments of the present application has low adhesion and hydrophobicity, that is, the chemical composition on the coating surface (such as fluorine-containing or silane groups) makes it have moderate hydrophobicity (contact angle about 90° - 120°), and the adhesion force between the water droplets and the surface is relatively low, but it is not sufficient to form the super-hydrophobic "lotus effect" (contact angle > 150°). Therefore, if the nano-coating surface of the glass panel is atomized, the water droplets will not quickly coalesce and roll off, but will remain in a dispersed state. At this time, the uniformity of the thickness of the nano-coating on the glass panel will affect the distribution of the water droplets. For example, if the thickness of the nano-coating is uneven, for the non-uniform area, that is, the area where the thickness does not basically remain the same, but the change in thickness is relatively large and discrete, then the distribution of the atomized water droplets in this area is also not uniform, and there may be a concentrated or sparse situation.

[0047] Based on the above characteristics, the target image can be obtained by photographing the nano-coating on the glass panel surface under the condition of atomizing the nano-coating surface. For example, set the camera position perpendicular to the surface of the glass panel coated with the nano-coating, and then photograph it to obtain the target image.

[0048] S202, the control device discretely divides the target image into K sub-image sets.

[0049] Any two sub-image sets among the K sub-image sets contain different sub-images. K is an integer greater than 1. For any two sub-images in the i-th sub-image set among the K sub-image sets, their positions in the target image are not adjacent, and any two sub-images in the i-th sub-image set have the same size. i is an integer ranging from 1 to K. The purpose of this is that subsequent artificial intelligence (AI) / machine learning (ML) models can process at the granularity of sub-images. On the one hand, it can reduce the performance requirements for AI / ML models, and on the other hand, it can also improve the accuracy. Additionally, since the target image is discretely divided, the discreteness of the division makes it highly probable that regions with continuous changes in water mist distribution are divided into different sub-image sets and thus separately processed by AI / ML models, avoiding the situation where regions with continuous changes in water mist distribution affect the AI / ML model and cause misprocessing, thereby also improving the accuracy. Specifically, it can include the following two situations:

[0050] Situation 1:

[0051] When the size of the target image is less than the threshold size, the control device discretely divides the target image into two sub-image sets.

[0052] For example, when the size of the target image is less than the threshold size, the control device divides the target image into M1*N1 grids through rasterization processing. M1 is the number of grids in one row of the target image, N1 is the number of grids in one column of the target image, and M1 and N1 are integers greater than or equal to 2. On this basis, the control device determines the grids with odd column indices in each row with an odd row index among the M1*N1 grids as a sub-image belonging to the first sub-image set, and the grids with even column indices in each row with an even row index among the M1*N1 grids as a sub-image belonging to the first sub-image set; the control device determines the grids with even column indices in each row with an odd row index among the M1*N1 grids as a sub-image belonging to the second sub-image set, and the grids with odd column indices in each row with an even row index among the M1*N1 grids as a sub-image belonging to the second sub-image set; where the two sub-image sets include the first sub-image set and the second sub-image set.

[0053] For easy understanding, an example of Situation 1 can be as Figure 3As shown in (a) therein, at this time, the first sub-image set includes: 3 sub-images corresponding to the gratings in the 1st, 3rd, and 5th columns of the 1st row, 3 sub-images corresponding to the gratings in the 2nd, 4th, and 6th columns of the 2nd row, 3 sub-images corresponding to the gratings in the 1st, 3rd, and 5th columns of the 3rd row, 3 sub-images corresponding to the gratings in the 2nd, 4th, and 6th columns of the 4th row, 3 sub-images corresponding to the gratings in the 1st, 3rd, and 5th columns of the 5th row, and 3 sub-images corresponding to the gratings in the 2nd, 4th, and 6th columns of the 6th row. The second sub-image set includes: 3 sub-images corresponding to the gratings in the 2nd, 4th, and 6th columns of the 1st row, 3 sub-images corresponding to the gratings in the 1st, 3rd, and 5th columns of the 2nd row, 3 sub-images corresponding to the gratings in the 2nd, 4th, and 6th columns of the 3rd row, 3 sub-images corresponding to the gratings in the 1st, 3rd, and 5th columns of the 4th row, 3 sub-images corresponding to the gratings in the 2nd, 4th, and 6th columns of the 5th row, and 3 sub-images corresponding to the gratings in the 1st, 3rd, and 5th columns of the 6th row. Any two sub-images in the first sub-image set or the second sub-image set are not adjacent, that is, there is no shared edge, so they are considered discrete. The segmentation method in the above case 1 can also be understood as a pseudo-random discrete segmentation.

[0054] Case 2:

[0055] The control device discretely divides the target image into three sub-image sets according to the size of the target image being greater than or equal to the threshold size.

[0056] For example, the control device, according to the size of the target image being greater than or equal to the threshold size, divides the target image into M2*N2 gratings through rasterization processing, where M2 is the number of gratings included in one row of the target image, N2 is the number of gratings included in one column of the target image, and M2 and N2 are integers greater than or equal to 3.

[0057] On this basis, the control device determines the gratings with the starting column index of 1 and the column index step of 3 in the row with row index 1 of the M2*N2 gratings, the gratings with the starting column index of 2 and the column index step of 3 in the row with row index 2 of the M2*N2 gratings, the gratings with the starting column index of 3 and the column index step of 3 in the row with row index 3 of the M2*N2 gratings, and the gratings with the starting column index of 1 and the column index step of 3 in the row with row index 4 of the M2*N2 gratings as a sub-image belonging to the first sub-image set, and then, by analogy, until traversing to the row with row index N2 of the M2*N2 gratings.

[0058] Similarly, the control device determines the grilles with a starting column index of 2 and a column index step size of 3 in the row with a row index of 1 among the M2 * N2 grilles, the grilles with a starting column index of 3 and a column index step size of 3 in the row with a row index of 2 among the M2 * N2 grilles, the grilles with a starting column index of 4 and a column index step size of 3 in the row with a row index of 3 among the M2 * N2 grilles, and the grilles with a starting column index of 2 and a column index step size of 3 in the row with a row index of 4 among the M2 * N2 grilles as a sub-image belonging to the second sub-image set. Then, and so on, until the row with a row index of N2 among the M2 * N2 grilles is traversed;

[0059] Similarly, the control device determines the grilles with a starting column index of 3 and a column index step size of 3 in the row with a row index of 1 among the M2 * N2 grilles, the grilles with a starting column index of 4 and a column index step size of 3 in the row with a row index of 2 among the M2 * N2 grilles, the grilles with a starting column index of 5 and a column index step size of 3 in the row with a row index of 3 among the M2 * N2 grilles, and the grilles with a starting column index of 3 and a column index step size of 3 in the row with a row index of 4 among the M2 * N2 grilles as a sub-image belonging to the second sub-image set. Then, and so on, until the row with a row index of N2 among the M2 * N2 grilles is traversed.

[0060] Among them, the three sub-image sets include a first sub-image set, a second sub-image set, and a third sub-image set.

[0061] For easy understanding, an example of Case 2 can be as Figure 3As shown in (b) therein, at this time, the first sub-image set includes: 2 sub-images corresponding to the gratings in the 1st and 4th columns of the 1st row, 2 sub-images corresponding to the gratings in the 2nd and 5th columns of the 2nd row, 2 sub-images corresponding to the gratings in the 1st and 4th columns of the 3rd row, 2 sub-images corresponding to the gratings in the 2nd and 5th columns of the 4th row, 2 sub-images corresponding to the gratings in the 1st and 4th columns of the 5th row, and 2 sub-images corresponding to the gratings in the 2nd and 5th columns of the 6th row. The second sub-image set includes: 2 sub-images corresponding to the gratings in the 2nd and 5th columns of the 1st row, 2 sub-images corresponding to the gratings in the 3rd and 6th columns of the 2nd row, 2 sub-images corresponding to the gratings in the 2nd and 5th columns of the 3rd row, 2 sub-images corresponding to the gratings in the 3rd and 6th columns of the 4th row, 2 sub-images corresponding to the gratings in the 2nd and 5th columns of the 5th row, and 2 sub-images corresponding to the gratings in the 3rd and 6th columns of the 6th row. The third sub-image set includes: 2 sub-images corresponding to the gratings in the 3rd and 6th columns of the 1st row, 2 sub-images corresponding to the gratings in the 1st and 4th columns of the 2nd row, 2 sub-images corresponding to the gratings in the 3rd and 6th columns of the 3rd row, 2 sub-images corresponding to the gratings in the 1st and 4th columns of the 4th row, 2 sub-images corresponding to the gratings in the 3rd and 6th columns of the 5th row, and 2 sub-images corresponding to the gratings in the 1st and 4th columns of the 6th row. That is, any two sub-images in the first sub-image set, the second sub-image set, or the third sub-image set are not adjacent, that is, there is no shared side, so they are considered discrete. The segmentation method in the above case 2 can also be understood as a pseudo-random discrete segmentation.

[0062] S203. The control device analyzes the K sub-image sets through the AI / ML model respectively to obtain the corresponding K analysis results, and fuses the K analysis results to obtain the analysis result of the target image.

[0063] The AI / ML model can be the above-mentioned DNN, specifically, it can be a CNN.

[0064] The analysis result of the target image is an image marking the non-uniform area of the nano-coating on the glass panel, which can be understood as marking the non-uniform area of the nano-coating on the target image.

[0065] For example, for the i-th sub-image set, where i traverses from 1 to K, the control device can randomly combine the sub-images in the i-th sub-image set in pairs to obtain P sub-image groups, and analyze the j-th sub-image group in the P sub-image groups through the AI / ML model to obtain the analysis result of the j-th sub-image group, where j is an integer traversing from 1 to P. The analysis result of the j-th sub-image group is two sub-images marking or not marking the non-uniform area of the nano-coating; the i-th analysis result among the K analysis results includes the analysis result of the j-th sub-image group.

[0066] Specifically, for the i-th sub-image set, where i traverses from 1 to K:

[0067] The control device can randomly combine the sub-images in the i-th set of sub-images in pairs to obtain P groups of sub-images; the control device can input the j-th group of sub-images into the feature extraction layer of the AI / ML model to obtain two feature sequences corresponding one-to-one to the two sub-images in the j-th group of sub-images output by the feature extraction layer of the AI / ML model. Among them, the feature extraction layer can include a convolutional layer and a pooling layer. The control device analyzes the two feature sequences through the feature processing layer of the AI / ML model to obtain the analysis result of the j-th group of sub-images. For example, the control device can calculate the difference between each two feature elements at the corresponding positions in the two feature sequences (such as the difference between the two feature elements with index 0, the difference between the two feature elements with index 1, the difference between the two feature elements with index 2, and so on), and set (or replace) the values of each two feature elements with a difference less than the preset threshold in the two feature sequences to a default value, such as 0, to obtain two corrected feature sequences. It should be understood that since the feature extraction layer performs convolution and pooling processing on the two sub-images in the j-th group of sub-images with the same scale, the sizes of these two feature sequences are the same, that is, they contain the same number of feature elements, and each feature element can be a vector, which can be represented as e j(xπ+μ)In this case, the difference value can be converted into an absolute value, such as a positive number. The control device can interleave the feature elements of each of the two corrected feature sequences according to the energy level to obtain two interleaved feature sequences. For example, for each of the two corrected feature sequences, if the value of a feature element (after being converted) in the corrected feature sequence is larger, then the energy of this feature element is higher. On this basis, the control device interleaves the feature elements with the same energy level in the corrected feature sequence together in the order from high to low energy level to obtain two interleaved feature sequences. The default value is the lowest energy level. For example, a corrected feature sequence contains 100 feature elements and a total of 4 energy levels. The traversal order of the interleaved feature sequence from front to back is as follows: first, 36 feature elements sorted in descending order of energy in the highest energy level #1, then 22 feature elements sorted in descending order of energy in the second highest energy level #2, then 20 feature elements sorted in descending order of energy in the second lowest energy level #3, and finally 22 feature elements in the lowest energy level #4, and these 22 feature elements are all default values. The advantage of this is that it can remove the randomness and discreteness of the feature sequence, making the feature sequence an ordered process for AI / ML, and sorting by energy from high to low can improve the processing performance of the AI / ML model, avoid redundant features from being confused, and thus improve the accuracy. Finally, the control device inputs the two interleaved feature sequences into the feature processing layer of the AI / ML model to obtain the analysis result of the j-th sub-image group output by the feature processing layer of the AI / ML model. Thus, the control device stitches together the K analysis results in the order of discrete segmentation, that is, restores the target image with annotations to obtain the analysis result of the target image. For example, an analysis result of the analysis result of the target image can be as shown in Figure 3 as shown in (c) of

[0068] It should be understood that the number of energy levels can be fixed, such as 4. The interval of each energy level can be evenly divided into 4 intervals according to the difference between the feature element with the maximum energy and the feature element with the minimum energy (default value) in a corrected feature sequence, and each interval corresponds to an energy level.

[0069] It should also be understood that since the control device interleaves the characteristic elements of each of the two corrected characteristic sequences according to the energy level to obtain two interleaved characteristic sequences, and inputs the two interleaved characteristic sequences into the characteristic processing layer of the AI / ML model, the analysis result of the j-th sub-image group output by the characteristic processing layer of the AI / ML model is obtained. In the sub-image, the values of the characteristic elements extracted from the region where the nano-coating is uniform are also relatively stable. However, in the region where the nano-coating is non-uniform, due to the variation in the thickness of the nano-coating, the values of the extracted characteristic elements also vary greatly. In addition, the regions where the nano-coating is non-uniform usually have randomness, and it is unlikely to have two non-uniform regions of the nano-coating with the same shape. Therefore, through the above method, the regions where the nano-coating is uniform at the same position in the two sub-images can be directly excluded, and then only the possible regions where the nano-coating is non-uniform are analyzed subsequently, which can effectively reduce the amount of information.

[0070] S204, the control device instructs the coating device to correct the nano-coating in the non-uniform region.

[0071] For example, the control device can inform the coating device of the information about the non-uniform region, so that the coating device can clean the non-uniform region and re-coat the nano-coating.

[0072] In summary, by atomizing the surface of the nano-coating, the atomized water droplets can be distributed on the surface of the nano-coating. At this time, if the thickness of the nano-coating is uneven, it may cause the distribution of the atomized water droplets to be uneven, and there may be a concentrated or sparse situation. For this characteristic, when the surface of the nano-coating is atomized, a target image can be taken of the nano-coating on the glass panel surface. Then, the control device discretely divides the target image into K sub-image sets, and then the AI / ML model analyzes the K sub-image sets respectively to obtain the corresponding K analysis results, and fuses the K analysis results to obtain the analysis result of the target image. In this way, the amount of data to be processed can be reduced, and processing at the granularity of a single sub-image can also improve the processing accuracy, so as to accurately indicate the region where the nano-coating is non-uniform on the glass panel. Finally, the control device instructs the coating device to correct the nano-coating in the non-uniform region, thereby realizing low-cost and high-precision detection and control of the nano-coating thickness.

[0073] As described above in combination with Figure 3The method provided by the embodiments of the present application is described in detail. The following introduces a glass surface nano - coating thickness control system for implementing the method provided by the embodiments of the present application. The system includes a control device and a coating device. The control device is configured to: The control device acquires a target image. The surface of the glass panel is coated with a nano - coating. The target image is obtained by photographing the nano - coating on the surface of the glass panel when the surface of the nano - coating is atomized. The control device discretely divides the target image into K sub - image sets. Any two sub - image sets among the K sub - image sets contain different sub - images. K is an integer greater than 1. For the i - th sub - image set among the K sub - image sets, any two sub - images included in it are not adjacent in the position in the target image, and any two sub - images included in the i - th sub - image set have the same size. i is an integer taking values from 1 to K. The control device analyzes at least two sub - image sets through an AI / ML model respectively, obtains the corresponding at least two analysis results, and fuses the at least two analysis results to obtain the analysis result of the target image. The analysis result of the target image indicates the non - uniform area of the nano - coating in the glass panel. The control device instructs the coating device to correct the nano - coating in the non - uniform area.

[0074] For the specific details of this system, reference can also be made to the relevant introduction of the above - mentioned method, which will not be elaborated here.

[0075] Next, in combination with Figure 4 each component of the electronic device 500 will be specifically introduced:

[0076] Among them, the processor 501 is the control center of the electronic device 500, which can be a single processor or a collective term for multiple processing elements. For example, the processor 501 is one or more central processing units (CPUs), or can be an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. For example: one or more digital signal processors (DSPs), or one or more field - programmable gate arrays (FPGAs).

[0077] Optionally, the processor 501 can execute various functions of the electronic device 500 by running or executing software programs stored in the memory 502 and calling data stored in the memory 502, such as the functions in the method Figure 2 shown above.

[0078] In a specific implementation, as an embodiment, the processor 501 may include one or more CPUs. For exampleFigure 4 CPU0 and CPU1 shown therein.

[0079] In a specific implementation, as an example, the electronic device 500 may also include multiple processors. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0080] Among them, the memory 502 is used to store the software program for executing the solution of this application and is controlled by the processor 501 for execution. The specific implementation method can refer to the above method embodiment and will not be elaborated here.

[0081] Optionally, the memory 502 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 502 may be integrated with the processor 501 or may exist independently, and is coupled to the processor 501 through the interface circuit of the electronic device 500 ( Figure 4 not shown). This application embodiment does not make specific limitations on this.

[0082] The transceiver 503 is used for communication with other devices. For example, when the multi-beam based positioning device is a terminal, the transceiver 503 can be used for communication with a network device or with another terminal.

[0083] Optionally, the transceiver 503 may include a receiver and a transmitter ( Figure 4 not shown separately). Among them, the receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.

[0084] Optionally, the transceiver 503 may be integrated with the processor 501 or may exist independently, and through the interface circuit of the electronic device 500 (Figure 4 is not shown) is coupled to the processor 501, and the embodiments of the present application do not make specific limitations thereon.

[0085] It should be noted that Figure 4 the structure of the electronic device 500 shown in does not constitute a limitation on the device. The actual electronic device 500 may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0086] In addition, for the technical effects of the electronic device 500, reference may be made to the technical effects of the methods in the above method embodiments, which will not be elaborated herein.

[0087] It should be understood that the processor in the embodiments of the present application may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0088] It should also be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM).

[0089] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions according to the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0090] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context.

[0091] In the present application, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following" or its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0092] It should be understood that in various embodiments of the present application, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0093] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0094] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0095] In several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some feature fields can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be electrical, mechanical, or other forms.

[0096] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0097] In addition, the functional units in each embodiment of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0098] If a function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0099] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.

Claims

1. A method for controlling the thickness of a nano coating on a glass surface, characterized in that: The method is applied to a control device, and the method comprises: The control device acquires a target image, the surface of the glass panel is coated with a nano coating, and the target image is obtained by photographing the nano coating on the surface of the glass panel when the surface of the nano coating is atomized; The control device discretely divides the target image into K sub-image sets, any two of the K sub-image sets contain different sub-images, K is an integer greater than 1, any two sub-images contained in the i-th sub-image set of the K sub-image sets are not adjacent in position in the target image, any two sub-images contained in the i-th sub-image set are the same in size, and i is an integer ranging from 1 to K; The control device analyzes the K sub-image sets respectively through the AI / ML model to obtain corresponding K analysis results, and fuses the K analysis results to obtain the analysis result of the target image, where the analysis result of the target image is an image marking the area of ​​the nano-coating non-uniformity on the glass panel; The control device instructs the coating device to correct the nano-coating in the non-uniform area.

2. The method according to claim 1, characterized in that The control device discretely divides the target image into K sub-image sets, including: The control device discretely divides the target image into two sub-image sets according to the size of the target image being smaller than a threshold size.

3. The method according to claim 2, characterized in that The control device discretely divides the target image into two sub-image sets according to the size of the target image being smaller than a threshold size, including: The control device divides the target image into M1*N1 grids by gridding processing according to the size of the target image being smaller than the threshold size, where M1 is the number of grids contained in a row of the target image, and N1 is the number of grids contained in a column of the target image, and M1 and N1 are integers greater than or equal to 2; The control device determines the grid with an odd column index in each row with an odd row index in the M1*N1 grids as a sub-image in the first sub-image set, and determines the grid with an even column index in each row with an even row index in the M1*N1 grids as a sub-image in the first sub-image set; The control device determines the grid with an even column index in each row with an odd row index in the M1*N1 grids as a sub-image in the second sub-image set, and determines the grid with an odd column index in each row with an even row index in the M1*N1 grids as a sub-image in the second sub-image set; The two sub-image sets include the first sub-image set and the second sub-image set.

4. The method according to claim 1, characterized in that: The control device discretely divides the target image into K sub-image sets, including: The control device discretely divides the target image into three sub-image sets according to the size of the target image being greater than or equal to a threshold size.

5. The method according to claim 4, characterized in that The control device discretely divides the target image into three sub-image sets according to the size of the target image being greater than or equal to a threshold size, including: The control device divides the target image into M2*N2 grids by gridding processing according to the size of the target image being greater than or equal to the threshold size, where M2 is the number of grids contained in a row of the target image, N2 is the number of grids contained in a column of the target image, and M2 and N2 are integers greater than or equal to 3; The control device determines the grid whose column index starts at 1 and whose step length of the column index is 3 in a row of grids with a row index of 1 among the M2*N2 grids, the grid whose column index starts at 2 and whose step length of the column index is 3 in a row of grids with a row index of 2 among the M2*N2 grids, the grid whose column index starts at 3 and whose step length of the column index is 3 in a row of grids with a row index of 3 among the M2*N2 grids, and the grid whose column index starts at 1 and whose step length of the column index is 3 in a row of grids with a row index of 4 among the M2*N2 grids as a sub-image belonging to the first sub-image set, and then, and so on, until traversing to a row of grids with a row index of N2 among the M2*N2 grids; The control device determines a grid whose column index starts at 2 and whose step length of the column index is 3 in a row of grids with a row index of 1 among the M2*N2 grids, a grid whose column index starts at 3 and whose step length of the column index is 3 in a row of grids with a row index of 2 among the M2*N2 grids, a grid whose column index starts at 4 and whose step length of the column index is 3 in a row of grids with a row index of 3 among the M2*N2 grids, and a grid whose column index starts at 2 and whose step length of the column index is 3 in a row of grids with a row index of 4 among the M2*N2 grids, as a sub-image belonging to the second sub-image set, and then, and so on, until traversing to a row of grids with a row index of N2 among the M2*N2 grids; The control device determines a grid whose column index starts at 3 and whose step length of the column index is 3 in a row of grids with a row index of 1 among the M2*N2 grids, a grid whose column index starts at 4 and whose step length of the column index is 3 in a row of grids with a row index of 2 among the M2*N2 grids, a grid whose column index starts at 5 and whose step length of the column index is 3 in a row of grids with a row index of 3 among the M2*N2 grids, and a grid whose column index starts at 3 and whose step length of the column index is 3 in a row of grids with a row index of 4 among the M2*N2 grids as a sub-image belonging to the second sub-image set, and then, and so on, until traversing to a row of grids with a row index of N2 among the M2*N2 grids; The three sub-image sets include the first sub-image set, the second sub-image set and the third sub-image set.

6. The method according to any one of claims 1 to 5, characterized in that: The control device analyzes the K sub-image sets respectively through the AI / ML model to obtain corresponding K analysis results, and fuses the K analysis results to obtain the analysis result of the target image, including: For the i-th sub-image set, i traverses from 1 to K, the control device randomly combines the sub-images in the i-th sub-image set in pairs to obtain P sub-image groups, and analyzes the j-th sub-image group in the P sub-image groups through the AI / ML model to obtain an analysis result of the j-th sub-image group, where j is an integer traversing from 1 to P, and the analysis result of the j-th sub-image group is two sub-images with or without annotated areas of the nano-coating non-uniformity; the i-th analysis result in the K analysis results includes the analysis result of the j-th sub-image group; The control device splices the K analysis results in the order of the discrete segmentation to obtain the analysis result of the target image.

7. The method according to claim 6, characterized in that The control device randomly combines sub-images in the i-th sub-image set in pairs to obtain P sub-image groups, and analyzes the j-th sub-image group in the P sub-image groups by using the AI / ML model to obtain an analysis result of the j-th sub-image group, including: For the i-th sub-image set, i traverses from 1 to K: The control device randomly combines the sub-images in the i-th sub-image set in pairs to obtain the P sub-image groups; the control device inputs the j-th sub-image group into the feature extraction layer of the AI / ML model to obtain two feature sequences output by the feature extraction layer of the AI / ML model, which correspond one to one to two sub-images in the j-th sub-image group; the control device analyzes the two feature sequences through the feature processing layer of the AI / ML model to obtain the analysis result of the j-th sub-image group.

8. The method according to claim 7, characterized in that The control device analyzes the two feature sequences through the feature processing layer of the AI / ML model to obtain an analysis result of the j-th sub-image group, including: The control device calculates the difference between each two feature elements at corresponding positions in the two feature sequences, and sets the value of each two feature elements whose difference is less than a preset threshold in the two feature sequences as a default value, thereby obtaining two modified feature sequences; The control device interleaves the characteristic elements of each of the two modified feature sequences according to the energy level to obtain two interleaved feature sequences, and inputs the two interleaved feature sequences into the feature processing layer of the AI / ML model to obtain the analysis result of the j-th sub-image group output by the feature processing layer of the AI / ML model.

9. The method according to claim 8, characterized in that The control device interleaves the characteristic elements of the two modified characteristic sequences according to the energy level to obtain two interleaved characteristic sequences, including: For each of the two corrected feature sequences, if the value of a feature element in the corrected feature sequence is larger, the energy of the feature element is higher. On this basis, the control device interweaves the feature elements of the same energy level in the corrected feature sequence in order from high to low energy levels to obtain the two interleaved feature sequences.

10. A glass surface nano coating thickness control system, characterized in that: The system comprises a control device and a coating device, wherein the control device is configured to: The control device acquires a target image, the surface of the glass panel is coated with a nano coating, and the target image is obtained by photographing the nano coating on the surface of the glass panel when the surface of the nano coating is atomized; The control device discretely divides the target image into K sub-image sets, any two of the K sub-image sets contain different sub-images, K is an integer greater than 1, any two sub-images contained in the i-th sub-image set of the K sub-image sets are not adjacent in position in the target image, any two sub-images contained in the i-th sub-image set are the same in size, and i is an integer ranging from 1 to K; The control device analyzes the at least two sub-image sets respectively through the AI / ML model to obtain at least two corresponding analysis results, and fuses the at least two analysis results to obtain an analysis result of the target image, wherein the analysis result of the target image indicates an area where the nano-coating in the glass panel is non-uniform; The control device instructs the coating device to correct the nano-coating in the non-uniform area.

Images