Preparation process of special sealant for construction surfaces with water

Abstract

This application discloses a preparation process for a special sealing primer for wet construction surfaces. The process involves capturing product surface images of the sealing primer using a camera, and employing computer vision-based image recognition and analysis technology to perform grid division and HOG feature extraction of the product surface images. Based on the semantic differences in the distribution characteristics of the HOG features of each product surface grid image and the clustered features, the process automatically determines whether the product particles are uneven. This automatic acquisition of product surface images via camera and analysis using computer vision technology significantly improves the speed of quality inspection and reduces the time required for manual inspection. It also enables more accurate identification and analysis of the distribution of particles on the product surface, thereby reducing errors caused by subjective judgment and improving the intelligence level of sealing primer quality inspection.

Description

Technical Field

[0001] This application relates to the field of intelligent preparation, and more specifically, to a preparation process for a special sealing primer for water-bearing construction surfaces. Background Technology

[0002] In the construction and decoration industry, sealing primers are an important type of coating, widely used in the surface treatment of walls, floors, and other surfaces to improve the adhesion and moisture resistance of the coated surface. Especially in humid or high-humidity environments, sealing primers can effectively prevent moisture penetration, protect buildings from damage, and improve the durability and aesthetics of the entire coating system.

[0003] However, traditional sealing primers often fail to perform optimally under humid conditions. When the surface contains moisture or is in a high-humidity environment, these primers may experience problems such as poor adhesion, blistering, and peeling. These issues not only affect the coating finish but can also lead to premature coating failure, increasing maintenance costs. Furthermore, slow drying times extend the application period, increasing costs, and insufficient adhesion can cause the coating to peel off, affecting the final quality of the application.

[0004] Therefore, a process for preparing a sealing primer specifically for wet construction surfaces is desired. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. An embodiment of this application provides a preparation process for a special sealing primer for wet construction surfaces. This process involves capturing product surface images of the special sealing primer for wet construction surfaces using a camera, and employing computer vision-based image recognition and analysis technology to perform grid division and HOG feature extraction of the product surface images. Based on the semantic difference distribution characteristics between the HOG features of each product surface grid image and the clustered features of the HOG features of each product surface grid image, the process automatically determines whether the product particles are uneven. Thus, by automatically capturing product surface images through a camera and analyzing them using computer vision technology, the speed of quality inspection can be greatly improved, reducing the time required for manual inspection. Simultaneously, it can more accurately identify and analyze the distribution of particles on the product surface, thereby reducing errors caused by subjective judgment and improving the intelligent level of sealing primer quality inspection.

[0006] According to one aspect of this application, a process for preparing a special sealing primer for wet construction surfaces is provided, comprising:

[0007] Heat the epoxy resin to 50℃-60℃;

[0008] A plasticizer is added to the heated epoxy resin and stirred until homogeneous to obtain a first mixed material;

[0009] A filler is added during the stirring of the first mixture to obtain a second mixture, wherein the filler is talc powder;

[0010] An additive is added during the stirring of the second mixture to obtain a third mixture, wherein the additive is any one or a combination of several of the following: defoamer, leveling agent, and lubricant;

[0011] A solvent is added to the third mixture to obtain a fourth mixture, wherein the solvent is ethanol or acetone;

[0012] Add a curing agent to the fourth mixture and stir evenly to obtain a special sealing primer for wet construction surfaces;

[0013] The quality inspection of the special sealing primer for the water-bearing construction surface was carried out to obtain the quality inspection results.

[0014] Compared with existing technologies, this application provides a preparation process for a special sealing primer for wet construction surfaces. It uses a camera to capture product surface images of the special sealing primer and employs computer vision-based image recognition and analysis technology to perform grid division and HOG feature extraction of the product surface images. Based on the semantic differences in the distribution characteristics of the HOG features of each product surface grid image and the clustering of these HOG features, it automatically determines whether the product particles are uneven. This automatic acquisition of product surface images via camera and analysis using computer vision technology significantly improves the speed of quality inspection and reduces the time required for manual inspection. Simultaneously, it enables more accurate identification and analysis of the distribution of particles on the product surface, thereby reducing errors caused by subjective judgment and improving the intelligence level of sealing primer quality inspection. Attached Figure Description

[0015] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 This is a flowchart illustrating the preparation process of a special sealing primer for wet construction surfaces according to an embodiment of this application;

[0017] Figure 2 This is a data flow diagram illustrating the preparation process of a special sealing primer for wet construction surfaces according to an embodiment of this application;

[0018] Figure 3This is a flowchart of sub-step S7 of the preparation process of a special sealing primer for wet construction surfaces according to an embodiment of this application. Detailed Implementation

[0019] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0020] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0021] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0022] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0023] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0024] Traditional sealing primers often fail to perform optimally under damp conditions. When the surface being applied contains moisture or is in a high-humidity environment, these primers may exhibit problems such as poor adhesion, blistering, and peeling. These issues not only affect the coating finish but can also lead to premature coating failure, increasing maintenance costs. Furthermore, slow drying times extend the application cycle, increasing costs, and insufficient adhesion can cause the coating to peel off, affecting the final application quality. Therefore, a specialized sealing primer preparation process for wet surfaces is desired.

[0025] The technical solution of this application proposes a preparation process for a special sealing primer for construction surfaces with water. Figure 1This is a flowchart illustrating the preparation process of a special sealing primer for wet construction surfaces according to an embodiment of this application. Figure 2 This is a data flow diagram illustrating the preparation process of a special sealing primer for wet construction surfaces according to an embodiment of this application. Figure 1 and Figure 2 As shown, the preparation process of the special sealing primer for wet construction surfaces according to an embodiment of this application includes the following steps: S1, heating epoxy resin to 50℃-60℃; S2, adding a plasticizer to the heated epoxy resin and stirring evenly to obtain a first mixed material; S3, adding a filler during stirring of the first mixed material to obtain a second mixed material, wherein the filler is talc powder; S4, adding an additive during stirring of the second mixed material to obtain a third mixed material, wherein the additive is any one or a combination of several of defoamers, leveling agents, and lubricants; S5, adding a solvent to the third mixed material to obtain a fourth mixed material, wherein the solvent is ethanol or acetone; S6, adding a curing agent to the fourth mixed material and stirring evenly to obtain a special sealing primer for wet construction surfaces; S7, performing quality inspection on the special sealing primer for wet construction surfaces to obtain quality inspection results.

[0026] Specifically, in S1 and S2, the epoxy resin is heated to 50°C-60°C; and a plasticizer is added to the heated epoxy resin and stirred evenly to obtain a first mixed material. It should be understood that heating the epoxy resin reduces its viscosity, making it easier to mix with the plasticizer. The plasticizer is typically a low molecular weight liquid, which reduces the overall viscosity of the epoxy resin, making it easier to flow and apply. Furthermore, the plasticizer improves the wettability of the epoxy resin, allowing it to adhere better to the substrate surface. This enhances the adhesion of the coating and reduces the formation of bubbles and voids.

[0027] Specifically, in step S3, a filler is added during the mixing of the first mixture to obtain a second mixture, wherein the filler is talc. Talc is an inert, low-density mineral that can increase the volume and weight of the sealing primer without significantly altering its other properties. This is useful in applications requiring increased primer quantity or weight, such as in architectural and industrial coatings. It is worth noting that talc can combine with other components in the primer to enhance adhesion between the primer and the substrate. This is crucial for ensuring strong adhesion of the primer to various surfaces.

[0028] Specifically, in step S4, an additive is added during the stirring of the second mixture to obtain a third mixture. The additive is any one or a combination of several of defoamers, leveling agents, and lubricants. The defoamer removes air bubbles, while the leveling agent improves surface finish; together, they produce a bubble-free and smooth mixture. The lubricant reduces friction, and the leveling agent further improves surface finish; together, they produce a smooth mixture that is easy to apply or mold, and easy to flow and shape.

[0029] Specifically, in step S5, a solvent is added to the third mixture to obtain a fourth mixture, wherein the solvent is ethanol or acetone. It should be understood that both ethanol and acetone are volatile organic solvents that can reduce the viscosity of the mixture, making it easier to flow and apply. Furthermore, both ethanol and acetone are volatile solvents that can evaporate rapidly, thereby accelerating the drying process of the mixture.

[0030] Specifically, in step S6, a curing agent is added to the fourth mixed material and stirred evenly to obtain a special sealing primer for wet construction surfaces. The curing agent is a chemical substance that reacts with the resin in the primer, causing it to cure into a hard protective layer.

[0031] Specifically, in step S7, the special sealant for wet construction surfaces is subjected to quality inspection to obtain inspection results. However, in the process of obtaining quality inspection results for the special sealant for wet construction surfaces, traditional quality inspection of sealants often relies on the experience and subjective judgment of quality inspectors. This may lead to inconsistencies in results, make it difficult to observe some minor defects, and may also increase the risk of missed detections or misjudgments. In addition, manually checking the surface quality of each sample is time-consuming, especially in the case of large-scale production, where manual quality inspection is difficult to keep up with the production pace, thus affecting production efficiency. Based on this, in the above-mentioned quality inspection of the special sealant for wet construction surfaces to obtain inspection results, the technical concept of this application is to acquire product surface images of the special sealant for wet construction surfaces using a camera, and to use computer vision-based image recognition and analysis technology to perform grid division and HOG feature extraction of the product surface images. In this way, based on the semantic difference distribution characteristics between the HOG features of each product surface grid image and the features after clustering the HOG features of each product surface grid image, the unevenness of product particles can be automatically determined. In this way, automatically capturing product surface images via a camera and analyzing them using computer vision technology can significantly improve the speed of quality inspection and reduce the time required for manual inspection. Simultaneously, it can more accurately identify and analyze the distribution of particles on the product surface, thereby reducing errors caused by subjective judgment and improving the level of intelligence in the quality inspection of sealing adhesives. Specifically, in a specific example of this application, such as... Figure 3As shown, step S7 includes: S71, acquiring a product surface image of the special sealant for water-bearing construction surfaces captured by a camera; S72, dividing the product surface image of the special sealant for water-bearing construction surfaces into a grid and extracting HOG features to obtain a set of HOG feature vectors of product surface grid granularity; S73, performing cluster analysis on the set of HOG feature vectors of product surface grid granularity to obtain cluster center vectors of product surface grid granularity texture features; S74, calculating the texture semantic difference coefficient between each product surface grid granularity HOG feature vector in the set of product surface grid granularity HOG feature vectors and the cluster center vectors of product surface grid granularity texture features to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients; S75, obtaining a quality inspection result based on the texture semantic difference distribution topology matrix, the quality inspection result being used to indicate whether the product particles are uneven.

[0032] Specifically, steps S71 and S72 involve acquiring product surface images of the water-based sealing primer captured by a camera; and then extracting HOG features from the product surface images of the water-based sealing primer after meshing to obtain a set of HOG feature vectors at the product surface mesh granularity. Considering that different local regions in the product surface image contain varying amounts of product surface feature information, in order to analyze and process the product feature information of each local region more meticulously, the technical solution of this application meshes the product surface image of the water-based sealing primer to obtain a set of product surface mesh granular images. This divides the product surface into multiple small regions, facilitating individual analysis of the image features of each local region, thereby more precisely identifying surface inhomogeneities or defects. Furthermore, considering that each product surface mesh granular image in the set of product surface mesh granular images contains texture and edge feature information about the product surface, and that HOG features can effectively characterize local texture information in the image, this is very useful for analyzing particle inhomogeneity or other surface characteristics of the product surface. Based on this, in the technical solution of this application, the HOG features of each product surface mesh grain image in the set of product surface mesh grain images are extracted to obtain a set of product surface mesh grain HOG feature vectors. That is, the HOG features contain descriptions of the edges and shapes in the product surface mesh grain images, which helps to more accurately identify and distinguish different surface defect patterns.

[0033] Specifically, in step S73, cluster analysis is performed on the set of HOG feature vectors of the product surface mesh granularity to obtain the cluster center vector of the product surface mesh granularity texture features. Considering that each HOG feature of the product surface mesh granularity in the set of HOG feature vectors expresses the texture information of the product surface mesh granularity contained within a local mesh, and in order to better understand the central trend of the texture features of the entire product surface image, thereby identifying and understanding the typical texture patterns and common features of the product surface, in the technical solution of this application, cluster analysis is performed on the set of HOG feature vectors of the product surface mesh granularity to obtain the cluster center vector of the product surface mesh granularity texture features. In particular, in a specific embodiment of this application, the cluster analysis can use the mean vector of the set of HOG feature vectors of the product surface mesh granularity as the cluster center vector of the product surface mesh granularity texture features.

[0034] Specifically, in step S74, the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity feature vectors and the product surface mesh granularity texture feature cluster center vector is calculated to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients. Considering that each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity feature vectors expresses the local texture and shape information of the product surface in different mesh regions, reflecting the detailed features of each local mesh surface, such as graininess, scratches, unevenness, or other surface defects. The product surface mesh granularity texture feature cluster center vector is calculated from multiple product surface mesh granularity HOG feature vectors, representing the common features of a group of mesh regions with similar texture characteristics. Therefore, in order to analyze and compare defect information within each grid region in a more granular manner, the technical solution of this application calculates the texture semantic difference coefficient between each product surface grid granularity HOG feature vector in the set of product surface grid granularity HOG feature vectors and the product surface grid granularity texture feature cluster center vector to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients. Specifically, firstly, by calculating the difference vector and covariance matrix between each product surface grid granularity HOG feature vector and the product surface grid granularity texture feature cluster center vector, the difference between each local product surface grid granularity HOG feature vector and the overall texture feature cluster center is quantified and measured to identify regions that are significantly different from typical texture patterns. These regions may indicate the presence of defects or anomalies.

[0035] In embodiments of this application, calculating the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients includes: calculating the positional difference between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain a set of product surface mesh granularity HOG difference feature vectors; calculating the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain a set of product surface mesh granularity HOG difference feature vectors; and calculating the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients. The covariance matrix between the cluster center vectors of the texture features of the product surface mesh granularity is calculated, and the reciprocal of the covariance matrix is ​​calculated to obtain a set of HOG inverse covariance matrices of the product surface mesh granularity; based on the matching value between the corresponding product surface mesh granularity HOG difference feature vectors and the product surface mesh granularity HOG inverse covariance matrix in the set of product surface mesh granularity HOG difference feature vectors and the product surface mesh granularity HOG inverse covariance matrix, a set of product surface mesh granularity HOG matching values ​​is obtained; the square root of each product surface mesh granularity HOG matching value in the set of product surface mesh granularity HOG matching values ​​is calculated to obtain the texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients.

[0036] Specifically, the set of matching values ​​for product surface mesh granularity HOG matching values ​​is obtained based on the matching values ​​between the corresponding product surface mesh granularity HOG differential feature vectors and the product surface mesh granularity HOG inverse covariance matrix in the set of product surface mesh granularity HOG differential feature vectors and the product surface mesh granularity HOG inverse covariance matrix. This includes calculating the transpose of the product surface mesh granularity HOG differential feature vectors and the product surface mesh granularity HOG inverse covariance matrix and the product surface mesh granularity HOG differential feature vectors to obtain the product surface mesh granularity HOG matching value.

[0037] In summary, in the above embodiments, calculating the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients includes: calculating the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector using the following semantic difference calculation formula to obtain multiple texture semantic difference coefficients; wherein, the semantic difference calculation formula is:

[0038]

[0039] Among them, x i It is the i-th HOG feature vector of the product surface mesh granularity in the set of product surface mesh granularity feature vectors, v c It is the cluster center vector of the surface mesh granularity texture features of the product, (·) T S is the transpose of the vector. i Let D(x) be the covariance matrix between the HOG feature vector of the i-th product surface mesh granularity and the cluster center vector of the texture feature of the product surface mesh granularity. i ,v c ) is the texture semantic difference coefficient between the i-th product surface mesh granularity HOG feature vector and the product surface mesh granularity texture feature cluster center vector.

[0040] Specifically, in step S75, a quality inspection result is obtained based on the texture semantic difference distribution topology matrix. This result indicates whether the product particles are uneven. In a specific example of this application, the texture semantic difference distribution topology matrix is ​​input into a classifier-based quality inspection module to obtain the quality inspection result, which indicates whether the product particles are uneven. That is, the texture semantic difference distribution topology feature, obtained by semantic difference calculation between each product surface grid granularity HOG feature vector in the set of product surface grid granularity texture feature vectors and the product surface grid granularity texture feature cluster center vector, is used for classification processing to automatically determine whether the product particles are uneven. In this way, by automatically acquiring product surface images through a camera and analyzing the product surface images using computer vision technology, the speed of quality inspection can be greatly improved, reducing the time required for manual inspection. At the same time, it can more accurately identify and analyze the distribution of product surface particles, thereby reducing errors caused by subjective judgment and improving the intelligence level of the sealing adhesive quality inspection. More specifically, the texture semantic difference distribution topology matrix is ​​input into a classifier-based quality inspection module to obtain the quality inspection result, which is used to indicate whether the product particles are uneven. This includes: expanding the texture semantic difference distribution topology matrix into a classification feature vector based on row vectors or column vectors; using multiple fully connected layers of the classifier to fully connect and encode the classification feature vector to obtain an encoded classification feature vector; and passing the encoded classification feature vector through the classifier's Softmax classification function to obtain the quality inspection result.

[0041] In a preferred embodiment of this application, considering that the set of HOG feature vectors of the product surface mesh granularity respectively expresses the HOG features of each product surface mesh granularity image, when performing cluster analysis and calculating the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector and the texture feature cluster center vector of the product surface mesh granularity, considering the texture semantic difference feature shift caused by the uneven mapping of each product surface mesh granularity HOG feature vector to the cluster center, it is expected to further improve the feature difference aggregation regression expression effect of the texture semantic difference distribution topology vector, thereby improving the accuracy of the quality inspection results obtained by the quality inspection module based on the classifier.

[0042] Therefore, when the texture semantic difference distribution topology vector is input into the classifier-based quality inspection module, this application optimizes the texture semantic difference distribution topology vector. The optimization includes the following steps:

[0043] Arrange the feature values ​​of the texture semantic difference distribution topology vector in ascending order to form the texture semantic difference distribution ordered topology vector;

[0044] In response to the fact that the absolute value of the difference between the i-th eigenvalue and the (i+1)-th eigenvalue of the texture semantic difference distribution order topology vector is less than or equal to the distance difference hyperparameter ε, the weighted sum between the i-th eigenvalue and the (i+1)-th eigenvalue is calculated as the optimized (i+1)-th eigenvalue, i.e.:

[0045] v′ i+1 =α×v i +β×v i+1

[0046] Among them, v i v i+1 v′ represents the i-th and (i+1)-th eigenvalues ​​of the texture semantic difference distribution order topological vector, respectively. i+1 For the (i+1)th feature value to be optimized, α and β are different weight parameters;

[0047] In response to the absolute value of the difference between the i-th eigenvalue and the (i+1)-th eigenvalue of the texture semantic difference distribution topological vector being greater than the distance difference hyperparameter ε, the square root of the sum of squares of all eigenvalues ​​of the texture semantic difference distribution topological vector is calculated, and the square root is multiplied by 2 and then divided by the square of the length of the texture semantic difference distribution topological vector to obtain the texture semantic difference distribution spatial primitive value γ.

[0048] After multiplying the spatial primitive value of the texture semantic difference distribution by the i-th feature value, the weighted sum between the product and the (i+1)-th feature value is calculated to obtain the optimized (i+1)-th feature value, that is:

[0049] v′ i+1 =δ×γ×v i -θ×v i+1

[0050] Where δ and θ represent different weight parameters;

[0051] The optimized (i+1)th eigenvalue of the texture semantic difference distribution order topology vector is combined to obtain the optimized texture semantic difference distribution topology vector, wherein the first eigenvalue of the texture semantic difference distribution order topology vector remains unchanged.

[0052] Thus, addressing the problem of insufficient global feature difference aggregation and regression representation capability caused by long distances exceeding a predetermined local distribution interval threshold under a predetermined feature value order distribution of the feature set of the texture semantic difference distribution topology vector, the complex structure of its feature value global network interaction is captured by using a high-dimensional feature space primitive representation based on self-inner product fusion of the texture semantic difference distribution topology vector. This is achieved by simulating scale-based high-dimensional feature space latent primitives to reconstruct the difference aggregation and regression relationship between feature values ​​of the texture semantic difference distribution topology vector, thereby realizing the encoding and reconstruction of the true sequence distribution behavior of the texture semantic difference distribution topology vector over long distances. This improves the aggregation expression effect of the texture semantic difference distribution topology vector and enhances the accuracy of the quality inspection results obtained by the classifier-based quality inspection module.

[0053] Therefore, for the high-dimensional feature manifold of the texture semantic difference distribution topological vector, the eigenvalues ​​of the feature set are used as the vector field representation of the aggregation dimension. The superposition values ​​of the vector field of the texture semantic difference distribution topological vector at isolated zero positions are used as order information to fix the local positions of the eigenvalues ​​of its feature set. A bias is added as a reward for the invertibility of the feature regression distribution field of the texture semantic difference distribution topological vector, so as to achieve target tracking of the mapping of the regression distribution of the texture semantic difference distribution topological vector to the eigenvalue positions. This allows the feature set of the texture semantic difference distribution topological vector to perceive the mapping migration to the aggregation distribution, thereby improving the interpretability of the feature difference aggregation regression of the texture semantic difference distribution topological vector and improving the accuracy of the quality inspection results obtained by the quality inspection module based on the classifier. In this way, automatically acquiring product surface images through a camera and analyzing them using computer vision technology can greatly improve the speed of quality inspection and reduce the time required for manual inspection. At the same time, it can more accurately identify and analyze the distribution of particles on the product surface, thereby reducing errors caused by subjective judgment and improving the intelligence level of the quality inspection of the sealing adhesive.

[0054] In summary, the preparation process of the special sealing primer for wet construction surfaces according to the embodiments of this application is explained. It involves acquiring product surface images of the special sealing primer for wet construction surfaces using a camera, and employing computer vision-based image recognition and analysis technology to perform grid division and HOG feature extraction of the product surface images. Based on the semantic difference distribution characteristics between the HOG features of each product surface grid image and the clustered features of the HOG features of each product surface grid image, the process automatically determines whether the product particles are uneven. Thus, by automatically acquiring product surface images through a camera and analyzing them using computer vision technology, the speed of quality inspection can be greatly improved, reducing the time required for manual inspection. Simultaneously, it can more accurately identify and analyze the distribution of particles on the product surface, thereby reducing errors caused by subjective judgment and improving the intelligent level of sealing primer quality inspection.

[0055] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A preparation process for a special sealing primer for wet construction surfaces, characterized in that, include: Heat the epoxy resin to 50℃-60℃; A plasticizer is added to the heated epoxy resin and stirred until homogeneous to obtain a first mixed material; A filler is added during the stirring of the first mixture to obtain a second mixture, wherein the filler is talc powder; An additive is added during the stirring of the second mixture to obtain a third mixture, wherein the additive is any one or a combination of several of the following: defoamer, leveling agent, and lubricant; A solvent is added to the third mixture to obtain a fourth mixture, wherein the solvent is ethanol or acetone; Add a curing agent to the fourth mixture and stir evenly to obtain a special sealing primer for wet construction surfaces; The quality inspection results were obtained by performing a quality inspection on the special sealing primer for the water-bearing construction surface. The quality inspection of the special sealing primer for the wet construction surface is carried out to obtain the inspection results, including: Acquire product surface images of the special sealing primer for water-bearing construction surfaces captured by a camera; After dividing the product surface image of the special sealing primer for water-conducting surfaces into a grid, HOG features are extracted to obtain a set of HOG feature vectors of the product surface grid granularity; Cluster analysis is performed on the set of HOG feature vectors of the product surface mesh granularity to obtain the cluster center vector of the product surface mesh granularity texture feature; The texture semantic difference coefficients are calculated between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients. The calculation of the texture semantic difference coefficients includes: calculating the product surface mesh granularity HOG difference feature vector and the product surface mesh granularity HOG inverse covariance matrix between each product surface mesh granularity HOG feature vector and the product surface mesh granularity texture feature cluster center vector, and obtaining the texture semantic difference coefficients based on the matching value between the product surface mesh granularity HOG difference feature vector and the product surface mesh granularity HOG inverse covariance matrix. Based on the texture semantic difference distribution topology matrix, the quality inspection results are obtained, which are used to indicate whether the product particles are uneven.

2. The preparation process of the special sealing primer for water-bearing construction surfaces according to claim 1, characterized in that, The curing agent is a polyamide curing agent or an amine curing agent.

3. The preparation process of the special sealing primer for water-bearing construction surfaces according to claim 2, characterized in that, After dividing the product surface image of the special sealing primer for water-bearing construction surfaces into a grid, HOG features are extracted to obtain a set of HOG feature vectors of the product surface grid granularity, including: The surface image of the special sealing primer for water-contaminated construction surfaces is divided into a grid to obtain a set of product surface grid particle size images; Extract the HOG features of each product surface mesh grain image from the set of product surface mesh grain images to obtain the set of product surface mesh grain HOG feature vectors.

4. The preparation process of the special sealing primer for water-bearing construction surfaces according to claim 3, characterized in that, Cluster analysis is performed on the set of HOG feature vectors of the product surface mesh granularity to obtain the cluster center vector of the product surface mesh granularity texture feature, including: calculating the mean vector of the set of HOG feature vectors of the product surface mesh granularity as the cluster center vector of the product surface mesh granularity texture feature.

5. The preparation process of the special sealing primer for water-bearing construction surfaces according to claim 4, characterized in that, Calculate the texture semantic difference coefficient between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity texture feature cluster center vectors to obtain a texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients, including: Calculate the positional difference between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector to obtain the set of product surface mesh granularity HOG difference feature vectors; Calculate the covariance matrix between each product surface mesh granularity HOG feature vector in the set of product surface mesh granularity HOG feature vectors and the product surface mesh granularity texture feature cluster center vector, and calculate the reciprocal of the covariance matrix to obtain the set of product surface mesh granularity HOG inverse covariance matrices; Based on the matching values ​​between the HOG difference feature vectors of the product surface mesh granularity and the HOG inverse covariance matrix of the product surface mesh granularity in each set of the product surface mesh granularity HOG difference feature vectors and the product surface mesh granularity HOG inverse covariance matrix, a set of HOG matching values ​​of the product surface mesh granularity is obtained. Calculate the square root of the HOG matching value of each product surface mesh granularity in the set of product surface mesh granularity values ​​to obtain the texture semantic difference distribution topology matrix composed of multiple texture semantic difference coefficients.

6. The preparation process of the special sealing primer for water-bearing construction surfaces according to claim 5, characterized in that, Based on the set of HOG difference feature vectors of the product surface mesh granularity and the set of HOG inverse covariance matrices of the product surface mesh granularity, a set of HOG matching values ​​of the product surface mesh granularity is obtained, including: calculating the transpose of the HOG difference feature vector of the product surface mesh granularity and the product surface mesh granularity HOG inverse covariance matrix of the product surface mesh granularity multiplied by the HOG difference feature vector of the product surface mesh granularity to obtain the HOG matching value of the product surface mesh granularity.

7. The preparation process of the special sealing primer for wet construction surfaces according to claim 6, characterized in that, Based on the texture semantic difference distribution topology matrix, a quality inspection result is obtained. The quality inspection result is used to indicate whether the product particles are uneven. This includes: inputting the texture semantic difference distribution topology matrix into a classifier-based quality inspection module to obtain the quality inspection result, which is used to indicate whether the product particles are uneven.

8. The preparation process of the special sealing primer for water-bearing construction surfaces according to claim 7, characterized in that, The texture semantic difference distribution topology matrix is ​​input into a classifier-based quality inspection module to obtain the quality inspection result, which indicates whether the product particles are uneven, including: The texture semantic difference distribution topology matrix is ​​expanded into a classification feature vector based on row vectors or column vectors; The classification feature vector is fully encoded using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and The encoded classification feature vector is passed through the Softmax classification function of the classifier to obtain the quality inspection result.

Images