Production method of lightweight porous daily-use ceramics

Abstract

The present application discloses a method for producing lightweight porous daily-use ceramics, which uses a camera to collect a video of the stirring state of the mud, and introduces a video processing and analysis algorithm at the back end to analyze the stirring state video, thereby performing automated, precise and intelligent monitoring and control of the mud stirring state, thereby improving the accuracy and efficiency of stirring state monitoring. In this way, the degree of automation and production efficiency of the ceramic preparation process can be improved, the subjectivity and uncertainty of human judgment can be reduced, and the quality and consistency of lightweight porous daily-use ceramic products can be improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent production, and more specifically, to a method for producing lightweight porous daily-use ceramics. Background Art

[0002] Lightweight porous daily-use ceramics are a type of ceramic material with a porous structure. They have good thermal insulation, water absorption and aesthetic properties and are widely used in tableware, flower pots, decorations and other fields in daily life.

[0003] In the production process of lightweight porous daily-use ceramics, mixing and stirring the raw materials is a key process link, which affects the physical properties of the mixed raw material slurry such as uniformity, fluidity and viscosity, and further affects the quality and performance of the ceramic body. Therefore, it is necessary to monitor and control the stirring state in real time to ensure the effect and efficiency of stirring. However, the traditional stirring state monitoring method mainly relies on manual experience observation and judgment or time control to determine whether to stop stirring. This method is subjective and uncertain, and has large errors and low efficiency, which can easily lead to too long or too short stirring time, thus affecting the quality and performance of the product.

[0004] Therefore, an optimized production solution for lightweight porous daily-use ceramics is desired. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiment of the present application provides a production method of lightweight porous daily-use ceramics, which collects the stirring state video of the mud through a camera, and introduces a video processing and analysis algorithm at the back end to analyze the stirring state video, so as to carry out automatic, precise and intelligent monitoring and control of the stirring state of the mud, and improve the accuracy and efficiency of stirring state monitoring. In this way, the degree of automation and production efficiency of the ceramic preparation process can be improved, and the subjectivity and uncertainty of human judgment can be reduced, thereby improving the quality and consistency of lightweight porous daily-use ceramic products.

[0006] According to one aspect of the present application, a method for producing a lightweight porous daily-use ceramic is provided, comprising:

[0007] Clay, quartz, feldspar and calcium carbonate are mixed and stirred with water to form a slurry;

[0008] Injecting the slurry into a mold for vibration, pressure and vacuum treatment, so that the slurry fully fills the mold, and exhausts excess water and bubbles to obtain a pre-treated mold;

[0009] Putting the pretreated mold into a drying chamber for drying, so that the slurry in the mold is dried and formed to obtain a dried ceramic body;

[0010] After the dried ceramic body is demolded, grinding, trimming and cleaning are performed to remove surface impurities of the dried ceramic body to obtain a pretreated ceramic body;

[0011] placing the pretreated ceramic body into a kiln for firing to obtain fired ceramic;

[0012] The fired ceramic is taken out, cooled, inspected and packaged to obtain a lightweight porous daily-use ceramic.

[0013] Compared with the prior art, the present application provides a method for producing lightweight porous daily-use ceramics, which collects the stirring state video of the mud through a camera, and introduces a video processing and analysis algorithm at the back end to analyze the stirring state video, so as to carry out automatic, precise and intelligent monitoring and control of the stirring state of the mud, thereby improving the accuracy and efficiency of stirring state monitoring. In this way, the degree of automation and production efficiency of the ceramic preparation process can be improved, the subjectivity and uncertainty of human judgment can be reduced, and the quality and consistency of lightweight porous daily-use ceramic products can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0015] Figure 1 It is a flow chart of a method for producing lightweight porous daily-use ceramics according to an embodiment of the present application;

[0016] Figure 2 A system architecture diagram of a method for producing lightweight porous daily-use ceramics according to an embodiment of the present application;

[0017] Figure 3 It is a flowchart of sub-step S1 of the method for producing lightweight porous daily-use ceramics according to an embodiment of the present application;

[0018] Figure 4 This is a flow chart of sub-step S15 of the method for producing lightweight porous daily-use ceramics according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0020] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0021] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0022] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0023] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0024] Considering that in the production process of the above-mentioned lightweight porous daily-use ceramics, the step of mixing clay, quartz, feldspar and calcium carbonate to form a slurry is particularly critical, the real-time and accurate control of the mixing affects the physical properties of the slurry such as uniformity, fluidity and viscosity, and further affects the quality and performance of the ceramic body. Therefore, it is necessary to monitor and control the mixing state in real time to ensure the quality and efficiency of the mixing.

[0025] In the technical solution of the present application, a method for producing lightweight porous daily-use ceramics is proposed. Figure 1 The present invention is a flow chart of a method for producing lightweight porous daily-use ceramics according to an embodiment of the present application. Figure 2 1 is a system architecture diagram of a method for producing lightweight porous daily-use ceramics according to an embodiment of the present application. Figure 1 and Figure 2As shown, according to the production method of the embodiment of the present application, the lightweight porous daily-use ceramics include the following steps: S1, mixing clay, quartz, feldspar and calcium carbonate, adding water and stirring to form a slurry; S2, injecting the slurry into a mold for vibration, pressure and vacuum treatment, so that the slurry fully fills the mold, and discharges excess water and bubbles to obtain a pre-treated mold; S3, placing the pre-treated mold in a drying chamber for drying treatment, so that the slurry in the mold is dried and formed to obtain a dried ceramic body; S4, after demolding the dried ceramic body, grinding, trimming and cleaning are performed to remove surface impurities of the dried ceramic body to obtain a pre-treated ceramic body; S5, placing the pre-treated ceramic body into a kiln for firing to obtain fired ceramics; S6, taking out the fired ceramics, cooling, inspecting and packaging them to obtain lightweight porous daily-use ceramics.

[0026] In particular, in S1, clay, quartz, feldspar and calcium carbonate are mixed and then water is added to form a slurry. In particular, in a specific example of the present application, Figure 3 As shown, the S1 includes: S11, acquiring a stirring state monitoring video collected by a camera; S12, performing video segmentation on the stirring state monitoring video to obtain a sequence of stirring state video clips; S13, performing feature extraction on the sequence of stirring state video clips through a stirring state temporal feature extractor based on a deep neural network model to obtain a sequence of stirring state local temporal feature graphs; S14, performing spatial feature explicitation processing on the sequence of stirring state local temporal feature graphs to obtain a sequence of spatially explicit stirring state local temporal feature graphs; S15, performing stirring state semantic change measurement on each two adjacent spatially explicit stirring state local temporal feature graphs in the sequence of spatially explicit stirring state local temporal feature graphs to obtain a stirring state semantic change measurement feature; S16, determining whether to stop stirring based on the stirring state semantic change measurement feature.

[0027] Specifically, the S11 obtains the stirring state monitoring video collected by the camera. It should be understood that the stirring state video can reflect the mixing uniformity of various components in the mud, the fluidity of the mud, the bubbles and suspended matter in the mud, the change trend of the mud, etc. In the technical solution of the present application, the stirring state monitoring video is analyzed to complete the automatic, precise and intelligent monitoring and control of the mud stirring state, thereby improving the accuracy and efficiency of the stirring state monitoring.

[0028] Specifically, in S12, the stirring state monitoring video is segmented to obtain a sequence of stirring state video clips. Considering that during the stirring process, the mud stirring state will change over time, this change will gradually occur in the time dimension, and the changes at each adjacent time point may be relatively weak. If the timing feature analysis is performed on the entire stirring state monitoring video, it is difficult to capture the timing detail features of the changes in the mud stirring state and the changes in specific nodes in the video. Therefore, in order to better extract the local detail timing features of the stirring state during the stirring process, so as to better understand and judge the changes in the stirring state, in the technical solution of the present application, the stirring state monitoring video is segmented to obtain a sequence of stirring state video clips. By dividing the entire stirring state monitoring video into multiple video clips, the stirring process can be decomposed into several small time periods, so as to better capture the details and changes of the stirring state.

[0029] Specifically, the S13 extracts features from the sequence of the stirring state video clips through a stirring state temporal feature extractor based on a deep neural network model to obtain a sequence of stirring state local temporal feature maps. That is, in the technical solution of the present application, the sequence of the stirring state video clips is subjected to feature mining in a stirring state temporal feature extractor based on a three-dimensional convolutional neural network model to extract the local temporal semantic feature information about the stirring state of each stirring state video clip in the sequence of the stirring state video clips, thereby obtaining a sequence of stirring state local temporal feature maps. More specifically, the sequence of the stirring state video clips is subjected to feature mining in a stirring state temporal feature extractor based on a three-dimensional convolutional neural network model to obtain a sequence of local temporal feature maps of the stirring state, including: using each layer of the stirring state temporal feature extractor based on the three-dimensional convolutional neural network model to perform the following on the input data in the forward pass of the layer: convolution processing on the input data to obtain a convolution feature map; pooling the convolution feature map based on a local feature matrix to obtain a pooling feature map; and nonlinear activation on the pooling feature map to obtain an activation feature map; wherein the output of the last layer of the stirring state temporal feature extractor based on the three-dimensional convolutional neural network model is the sequence of local temporal feature maps of the stirring state, and the input of the first layer of the stirring state temporal feature extractor based on the three-dimensional convolutional neural network model is the sequence of the stirring state video clips.

[0030] It is worth noting that the 3D Convolutional Neural Network (3D CNN) is a neural network model for processing three-dimensional data. Compared with the traditional two-dimensional convolutional neural network (2D CNN), 3DCNN is more effective in processing data with time and space dimensions such as videos, medical images, and dynamic voxel data. The basic structure and operation of 3DCNN are similar to those of 2D CNN, but the time dimension is introduced in the convolution operation. The following are the general steps of 3D CNN: Input data: The input of 3D CNN is three-dimensional data, usually expressed as [Batch Size, Channels, Depth, Height, Width]. Among them, Batch Size represents the batch size, Channels represents the number of channels, Depth represents the time dimension, and Height and Width represent the spatial dimension; Convolution layer: Similar to 2D CNN, 3D CNN contains multiple convolution layers. Each convolution layer uses a set of learnable convolution kernels (or filters) to extract features by sliding over all time and space positions of the input data. The size of the convolution kernel is usually [Depth, Height, Width]. The convolution operation generates a set of feature maps, each of which corresponds to a convolution kernel; Activation function: After the convolution operation of each convolution layer, a nonlinear activation function, such as ReLU, is usually applied to introduce nonlinear properties; Pooling layer: In order to reduce the size of the feature map and extract more significant features, 3D CNN usually uses pooling layers. The pooling layer downsamples the feature map in the time and space dimensions. Common pooling operations include maximum pooling and average pooling; Fully connected layer: After multiple convolutional layers and pooling layers, a fully connected layer can be added to perform classification or regression tasks. The fully connected layer flattens the feature map into a one-dimensional vector and generates the final output through a series of fully connected operations; Output layer: Depending on the specific task, the output layer can use different activation functions and loss functions. For example, for classification tasks, the commonly used output layer activation function is softmax, and the cross entropy loss function is used. The training process of 3D CNN is similar to other neural networks, and the back propagation algorithm is usually used for parameter update. During the training process, labeled data can be used for supervised learning, and the network parameters are optimized by minimizing the loss function.

[0031] Specifically, the S14 performs spatial feature explicit processing on the sequence of the local time series feature graphs of the stirring state to obtain a sequence of spatially explicit local time series feature graphs of the stirring state. It should be understood that during the stirring process, the state changes in different regions have different importance. Some regions may contain more key information about the stirring state, while other regions may be relatively less important. Therefore, in order to highlight the key semantic feature information of the stirring state to improve the sensitivity to changes in the stirring state, in the technical solution of the present application, the sequence of the local time series feature graphs of the stirring state is further subjected to a spatial attention layer for spatial feature enhancement, thereby obtaining a sequence of spatially explicit local time series feature graphs of the stirring state. By applying the spatial attention layer, the sequence of the local time series feature graphs of the stirring state can be weighted, so that the semantic feature information in the key spatial regions of the mud stirring state is highlighted, while the features of the non-key regions are suppressed, so as to further highlight the important regions and change trends of the stirring state and enhance the expressive power of the stirring state features. Specifically, the sequence of the stirring state local temporal feature maps is passed through the spatial attention layer to obtain the sequence of the spatially explicit stirring state local temporal feature maps; including: passing the sequence of the stirring state local temporal feature maps through the spatial attention module to obtain a weight score map; inputting the weight score map into the Softmax activation function to obtain a spatial attention weight feature map; and calculating the position point multiplication between the spatial attention weight feature map and the sequence of the stirring state local temporal feature maps to obtain the sequence of the spatially explicit stirring state local temporal feature maps.

[0032] It is worth noting that the spatial attention layer is a neural network layer used to enhance the model's attention to different spatial locations in the input data. It can help the model automatically learn and focus on important spatial regions when processing images, videos or other data with spatial structures.

[0033] Specifically, in S15, the semantic change measurement of the stirring state is performed on each adjacent two spatially explicit stirring state local time series feature graphs in the sequence of the spatially explicit stirring state local time series feature graphs to obtain the stirring state semantic change measurement feature. In particular, in a specific example of the present application, Figure 4 As shown, the S15 includes: S151, optimizing the sequence of the spatially explicit stirring state local temporal feature graphs to obtain a sequence of optimized spatially explicit stirring state local temporal feature graphs; S152, calculating the stirring state semantic change measurement coefficient between each adjacent two optimized spatially explicit stirring state local temporal feature graphs in the sequence of optimized spatially explicit stirring state local temporal feature graphs to obtain a stirring state semantic change measurement feature vector composed of multiple stirring state semantic change measurement coefficients as the stirring state semantic change measurement feature.

[0034] More specifically, in S151, the sequence of the spatially explicit stirring state local time-series characteristic graphs is optimized to obtain a sequence of optimized spatially explicit stirring state local time-series characteristic graphs. In particular, in the above technical scheme, each stirring state local temporal feature map in the sequence of the stirring state local temporal feature maps expresses the image semantic features of the local time domain temporal association of the video clip of the stirring state monitoring video, and after the sequence of the stirring state local temporal feature maps passes through the spatial attention layer, the semantic spatial distribution of the feature matrix dimension of the image semantic features can be strengthened, so that each spatially explicit stirring state local temporal feature map in the sequence of the spatially explicit stirring state local temporal feature maps expresses the spatial association representation of the image semantic features in the spatial distribution dimension of the feature matrix, and expresses the temporal association representation of the image semantic features in the channel distribution dimension of the feature matrix, that is, the spatially explicit stirring state local temporal feature map is based on the image semantic feature distribution representation in the spatial distribution dimension within the feature matrix and the channel distribution dimension between the feature matrices in the spatial distribution dimension, that is, it also has spatial information attributes corresponding to the spatial association distribution and temporal association distribution of the image semantic features in the overall distribution dimension. Therefore, if the spatial information expression effect of the spatially explicit stirring state local time series feature map as a high-dimensional feature can be improved, the expression effect of the spatially explicit stirring state local time series feature map can be improved, thereby improving the calculation accuracy of the stirring state semantic change measurement coefficient between each two adjacent spatially explicit stirring state local time series feature maps in the sequence of the spatially explicit stirring state local time series feature maps, and improving the accuracy of the classification result obtained by the classifier of the stirring state semantic change measurement feature vector. Based on this, the applicant of the present application for the spatially explicit stirring state local time series feature map, for example, is recorded as Optimize, expressed as: in and ,Right now and All of them are local time series characteristic diagrams of the spatially manifested stirring state The eigenvalues ​​of is the local space division coefficient. Specifically, the local time series characteristic diagram of the stirring state is visualized in the space The local segmentation space in the expanded Hilbert space is used as a reference to visualize the local time series characteristic graph of the stirring state in the space. The local integration of the surface is performed on the characteristic manifold in the high-dimensional characteristic space, so as to correct the local time series characteristic map of the spatially manifested stirring state based on the local integration processing of the integral function. The phase change discontinuity of the characteristic manifold expressed by the non-stationary data sequence after the local space expansion is obtained, so as to obtain a finer structure and geometric features of the characteristic manifold, and improve the local time series feature map of the spatial explicit stirring state. The spatial information expression effect in the high-dimensional feature space can improve the expression effect of the local time series feature map of the spatially explicit stirring state, and finally improve the accuracy of the classification result obtained by the classifier of the semantic change metric feature vector of the stirring state. In this way, the automatic, precise and intelligent monitoring and control of the mud stirring state can be carried out to improve the automation degree and production efficiency of the ceramic preparation process, reduce the subjectivity and uncertainty of human judgment, and thus improve the quality and consistency of lightweight porous daily-use ceramic products.

[0035] More specifically, the S152 is to calculate the mixing state semantic change measurement coefficient between each adjacent two optimized spatially explicit mixing state local temporal feature graphs in the sequence of the optimized spatially explicit mixing state local temporal feature graphs to obtain a mixing state semantic change measurement feature vector composed of a plurality of mixing state semantic change measurement coefficients as the mixing state semantic change measurement feature. It should be understood that during the mixing process, the change of the mixing state not only exists in each local time period, but also has a temporal change pattern in the entire time domain during the entire mixing process. In particular, it is also considered that there is volatility between the mixing state temporal features after spatial feature manifestation in each local time period, which may be the node of the mixing state change. Therefore, in order to analyze and measure the mixing state semantic change more accurately, in the technical solution of the present application, the mixing state semantic change measurement coefficient between each adjacent two spatially explicit mixing state local temporal feature graphs in the sequence of the spatially explicit mixing state local temporal feature graphs is further calculated to obtain a mixing state semantic change measurement feature vector composed of a plurality of mixing state semantic change measurement coefficients. By calculating the semantic change of the stirring state between two adjacent spatially explicit stirring state local temporal feature graphs, the overall changes in the stirring process can be comprehensively considered, and the changing trend and mutation information of the stirring state can be captured, so that the changing situation of the stirring state can be described more comprehensively, and richer information can be provided for subsequent stirring state judgment and control. More specifically, the stirring state semantic change measurement coefficient between each two adjacent optimized spatially explicit stirring state local temporal feature graphs in the sequence of the optimized spatially explicit stirring state local temporal feature graphs is calculated to obtain a stirring state semantic change measurement feature vector composed of multiple stirring state semantic change measurement coefficients as the stirring state semantic change measurement feature, including: calculating the stirring state semantic change measurement coefficient between each two adjacent optimized spatially explicit stirring state local temporal feature graphs in the sequence of the optimized spatially explicit stirring state local temporal feature graphs by the following formula to obtain the multiple stirring state semantic change measurement coefficients; wherein, the formula is: in, is the first global cross entropy weighting coefficient, is the value of each position in the first initial feature map, is the value of each position in the second initial feature map, is the width of the first initial feature map and the second initial feature map, is the height of the first initial feature map and the second initial feature map, and is the number of channels of the convolutional neural network; and, the multiple stirring state semantic change measurement coefficients are arranged in one dimension to obtain the stirring state semantic change measurement feature vector as the stirring state semantic change measurement feature.

[0036] It is worth mentioning that in other specific examples of the present application, the semantic change measurement of the stirring state can be performed on each adjacent two spatially explicit stirring state local temporal feature maps in the sequence of the spatially explicit stirring state local temporal feature maps to obtain the stirring state semantic change measurement feature, for example: input the sequence of the spatially explicit stirring state local temporal feature maps; for each feature map, apply the spatial explicit operation to highlight the local features of the stirring state. This can be achieved by introducing a spatial attention layer or other explicit operations; for two adjacent spatially explicit stirring state local temporal feature maps, we need to calculate the stirring state semantic change measurement between them. This can be done by the following steps: for each feature map, a convolutional neural network (CNN) or other feature extraction method can be used to extract the semantic features of the stirring state. This can be achieved by applying a convolution operation on the feature map; using an appropriate measurement method (such as Euclidean distance, cosine similarity, etc.), the stirring state semantic change measurement between adjacent feature maps is calculated. This will measure the difference or similarity between them; the calculated stirring state semantic change measurement is used as a feature to obtain a stirring state semantic change measurement feature sequence.

[0037] Specifically, the S16 determines whether to stop stirring based on the semantic change measurement feature of the stirring state. That is, in the technical solution of the present application, the stirring state semantic change measurement feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether to stop stirring. In other words, the classification process is performed using the stirring state semantic change measurement temporal feature information to perform automated, precise and intelligent monitoring and control of the mud stirring state, thereby improving the accuracy and efficiency of stirring state monitoring. More specifically, the stirring state semantic change measurement feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether to stop stirring, including: using multiple fully connected layers of the classifier to fully connect encode the stirring state semantic change measurement feature vector to obtain an encoded classification feature vector; and, passing the encoded classification feature vector through the Softmax classification function of the classifier to obtain the classification result.

[0038] A classifier is a machine learning model or algorithm that is used to classify input data into different categories or labels. Classifiers are part of supervised learning and perform classification tasks by learning the mapping relationship from input data to output categories.

[0039] A fully connected layer is a common type of layer in neural networks. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and each connection has a weight. This means that each neuron in the fully connected layer receives inputs from all neurons in the previous layer, and performs a weighted sum of these inputs through the weights, and then passes the result to the next layer.

[0040] The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, and the sum of these probability values ​​is equal to 1. The Softmax function is often used in the output layer of a neural network, especially for multi-classification problems, because it can map the network output to the probability distribution of each category. During the training process, the output of the Softmax function can be used to calculate the loss function and update the network parameters through the back propagation algorithm. It is worth noting that the output of the Softmax function does not change the relative size relationship between the elements, but only normalizes them. Therefore, the Softmax function does not change the characteristics of the input vector, but only converts it into a probability distribution form.

[0041] In particular, in S2, the slurry is injected into the mold for vibration, pressure and vacuum treatment, so that the slurry fully fills the mold, and the excess water and bubbles are discharged to obtain the pre-treated mold. It should be understood that vibration and pressure treatment can help the slurry flow better and fill every detail and gap in the mold. Through vibration and pressure, it can be ensured that the slurry completely fills the mold to avoid voids or uneven filling. In addition, vibration and vacuum treatment help to remove bubbles in the slurry. Vibration can make bubbles float out of the mud and rise to the surface, while vacuum treatment can extract bubbles from the mud by creating a vacuum environment. Excluding bubbles can avoid the formation of gaps or uneven structures in the mold and improve the density and quality of the product.

[0042] In particular, in S3, the pre-treated mold is placed in a drying chamber for drying, so that the mud in the mold is dried and formed to obtain a dried ceramic body. It should be understood that the drying process in the drying chamber can quickly remove the residual moisture in the mold. The presence of moisture may cause cracking, deformation or unstable quality of the molded product during the curing process. Through the drying process, the moisture can be evaporated and discharged, so that the mud in the mold is gradually dried and formed. Through the drying process, the mud in the mold can be gradually dried and formed, thereby improving the quality of the product. Drying can make the surface of the product more uniform and smooth, and avoid shrinkage, deformation or surface defects of the product during the curing process. The drying process helps to ensure that the geometric shape and details of the product are accurately retained, and improves the strength and durability of the product.

[0043] In particular, in S4, after the dried ceramic body is demolded, grinding, trimming and cleaning are performed to remove surface impurities of the dried ceramic body to obtain a pretreated ceramic body. It should be understood that grinding, trimming and cleaning can effectively remove impurities on the surface of the dried ceramic body, such as dust, residual mold material, particles, etc. These impurities may affect the appearance and quality of the ceramic body, and a cleaner and smoother surface can be obtained by removing them.

[0044] In particular, the S5, the pre-treated ceramic body is placed in a kiln for firing to obtain fired ceramics. It should be understood that during the firing process, the ceramic body undergoes physical and chemical changes to achieve the formation of the final ceramic product. Under high temperature conditions, the original materials in the ceramic body will undergo processes such as thermal decomposition, sintering and crystallization, so that the ceramic body is gradually transformed into a hard and dense ceramic product. And, during the firing process, the particles of the ceramic body will fuse and solidify at high temperature to form a dense crystal structure. This makes the ceramic product have higher strength, hardness and heat resistance, and can withstand various mechanical and thermal stresses.

[0045] In particular, the fired ceramic is taken out in S6, and cooled, inspected and packaged to obtain lightweight porous daily-use ceramics. It should be understood that fired ceramic products usually need to go through a cooling process to reduce their temperature to a safe level. The cooling process helps the structure of the ceramic product to further solidify and stabilize, ensuring that it has the required physical and chemical properties. After cooling, the ceramic product needs to be inspected and quality controlled.

[0046] In summary, the production method of lightweight porous daily-use ceramics according to the embodiment of the present application is explained, which collects the stirring state video of the mud through a camera, and introduces a video processing and analysis algorithm at the back end to analyze the stirring state video, so as to carry out automatic, precise and intelligent monitoring and control of the stirring state of the mud, thereby improving the accuracy and efficiency of stirring state monitoring. In this way, the degree of automation and production efficiency of the ceramic preparation process can be improved, the subjectivity and uncertainty of human judgment can be reduced, and the quality and consistency of lightweight porous daily-use ceramic products can be improved.

[0047] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for producing lightweight porous daily-use ceramics, characterized in that: include: Clay, quartz, feldspar and calcium carbonate are mixed and stirred with water to form a slurry; Injecting the slurry into a mold for vibration, pressure and vacuum treatment, so that the slurry fully fills the mold, and exhausts excess water and bubbles to obtain a pre-treated mold; placing the pretreated mold in a drying chamber for drying, so that the slurry in the mold is dried and formed to obtain a dried ceramic body; After the dried ceramic body is demolded, grinding, trimming and cleaning are performed to remove surface impurities of the dried ceramic body to obtain a pretreated ceramic body; placing the pretreated ceramic body into a kiln for firing to obtain fired ceramic; The fired ceramic is taken out, cooled, inspected and packaged to obtain a lightweight porous daily-use ceramic; The clay, quartz, feldspar and calcium carbonate are mixed and then stirred with water to form a slurry, including: Obtain stirring status monitoring video collected by the camera; Performing video segmentation on the stirring state monitoring video to obtain a sequence of stirring state video segments; Extracting features from the sequence of the stirring state video clips by a stirring state temporal feature extractor based on a deep neural network model to obtain a sequence of stirring state local temporal feature graphs; Performing spatial feature explicitation processing on the sequence of the local time-series characteristic graphs of the stirring state to obtain a sequence of spatially explicit local time-series characteristic graphs of the stirring state; Performing stirring state semantic change measurement on every two adjacent spatially explicit stirring state local time series feature graphs in the sequence of the spatially explicit stirring state local time series feature graphs to obtain stirring state semantic change measurement features; Determining whether to stop stirring based on the stirring state semantic change measurement feature; The sequence of local temporal feature maps of the stirring state is subjected to spatial feature explicitation processing to obtain a sequence of spatially explicit local temporal feature maps of the stirring state, comprising: passing the sequence of local temporal feature maps of the stirring state through a spatial attention layer to obtain a sequence of spatially explicit local temporal feature maps of the stirring state; The stirring state semantic change measurement is performed on each two adjacent spatially explicit stirring state local time series feature graphs in the sequence of the spatially explicit stirring state local time series feature graphs to obtain the stirring state semantic change measurement feature, including: Optimizing the sequence of the spatially explicit stirring state local time-series characteristic graphs to obtain an optimized sequence of the spatially explicit stirring state local time-series characteristic graphs; Calculating the stirring state semantic change measurement coefficient between each two adjacent optimized spatial explicit stirring state local time series feature graphs in the sequence of the optimized spatial explicit stirring state local time series feature graphs to obtain a stirring state semantic change measurement feature vector composed of a plurality of stirring state semantic change measurement coefficients as the stirring state semantic change measurement feature; Wherein, calculating the stirring state semantic change measurement coefficient between each two adjacent optimized spatial explicit stirring state local temporal feature graphs in the sequence of the optimized spatial explicit stirring state local temporal feature graphs to obtain a stirring state semantic change measurement feature vector composed of a plurality of stirring state semantic change measurement coefficients as the stirring state semantic change measurement feature, including: calculating the stirring state semantic change measurement coefficient between each two adjacent optimized spatial explicit stirring state local temporal feature graphs in the sequence of the optimized spatial explicit stirring state local temporal feature graphs by the following formula to obtain the plurality of stirring state semantic change measurement coefficients; Wherein, the formula is: ;in, is the first global cross entropy weighting coefficient, is the value of each position in the first initial feature map, is the value of each position in the second initial feature map, is the width of the first initial feature map and the second initial feature map, is the height of the first initial feature map and the second initial feature map, and is the number of channels of the convolutional neural network; Arrange the plurality of stirring state semantic change measurement coefficients in one dimension to obtain the stirring state semantic change measurement feature vector as the stirring state semantic change measurement feature; The sequence of the spatially explicit stirring state local time series characteristic diagrams is optimized to obtain a sequence of optimized spatially explicit stirring state local time series characteristic diagrams, which is expressed as: ;in, and ,Right now and All of them are local time series characteristic diagrams of the spatially manifested stirring state The eigenvalues ​​of is the local space partition coefficient.

2. The method for producing lightweight porous daily-use ceramics according to claim 1, characterized in that: The deep neural network model is a three-dimensional convolutional neural network model.

3. The method for producing lightweight porous daily-use ceramics according to claim 2, characterized in that: Determining whether to stop stirring based on the stirring state semantic change measurement feature includes: passing the stirring state semantic change measurement feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether to stop stirring.

4. The method for producing lightweight porous daily-use ceramics according to claim 3, characterized in that: The stirring state semantic change measurement feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether to stop stirring, including: Using multiple fully connected layers of the classifier to perform fully connected encoding on the stirring state semantic change metric feature vector 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 classification result.

Images