CN116907214B - Preparation process and system of environment-friendly domestic ceramic - Google Patents

Preparation process and system of environment-friendly domestic ceramic Download PDF

Info

Publication number
CN116907214B
CN116907214B CN202310516727.4A CN202310516727A CN116907214B CN 116907214 B CN116907214 B CN 116907214B CN 202310516727 A CN202310516727 A CN 202310516727A CN 116907214 B CN116907214 B CN 116907214B
Authority
CN
China
Prior art keywords
ceramic
feature
time sequence
heat distribution
temperature
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310516727.4A
Other languages
Chinese (zh)
Other versions
CN116907214A (en
Inventor
黄培煌
张晓文
黄培元
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Xiahe Porcelain Co ltd
Original Assignee
Guangdong Xiahe Porcelain Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Xiahe Porcelain Co ltd filed Critical Guangdong Xiahe Porcelain Co ltd
Priority to CN202310516727.4A priority Critical patent/CN116907214B/en
Publication of CN116907214A publication Critical patent/CN116907214A/en
Application granted granted Critical
Publication of CN116907214B publication Critical patent/CN116907214B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F27—FURNACES; KILNS; OVENS; RETORTS
    • F27B—FURNACES, KILNS, OVENS, OR RETORTS IN GENERAL; OPEN SINTERING OR LIKE APPARATUS
    • F27B17/00—Furnaces of a kind not covered by any preceding group
    • F27B17/0016—Chamber type furnaces
    • F27B17/0041—Chamber type furnaces specially adapted for burning bricks or pottery
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F27—FURNACES; KILNS; OVENS; RETORTS
    • F27D—DETAILS OR ACCESSORIES OF FURNACES, KILNS, OVENS, OR RETORTS, IN SO FAR AS THEY ARE OF KINDS OCCURRING IN MORE THAN ONE KIND OF FURNACE
    • F27D19/00—Arrangements of controlling devices
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
    • G01J5/48—Thermography; Techniques using wholly visual means
    • G01J5/485—Temperature profile
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P40/00—Technologies relating to the processing of minerals
    • Y02P40/60—Production of ceramic materials or ceramic elements, e.g. substitution of clay or shale by alternative raw materials, e.g. ashes

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computational Linguistics (AREA)
  • Databases & Information Systems (AREA)
  • Mechanical Engineering (AREA)
  • Medical Informatics (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Radiation Pyrometers (AREA)

Abstract

The application relates to the technical field of intelligent preparation, and particularly discloses a preparation process and a system of an environment-friendly daily ceramic, which comprehensively utilize a thermal infrared monitoring video of the environment-friendly daily ceramic and firing temperature values of a plurality of preset time points, and carry out self-adaptive control on firing temperature based on real-time conditions of ceramic firing by combining a deep learning technology and an artificial intelligence technology, so as to provide an intelligent firing temperature control scheme. The firing temperature control scheme can control the firing temperature of the ceramic in a self-adaptive manner based on the real-time condition of ceramic firing, so that manual intervention is reduced as much as possible, the production efficiency is improved, and meanwhile, the stable quality of the environment-friendly daily ceramic can be effectively ensured.

Description

Preparation process and system of environment-friendly domestic ceramic
Technical Field
The application relates to the technical field of intelligent preparation, and more particularly relates to a preparation process and a system of environment-friendly household ceramic.
Background
In the firing process of the ceramic, the whole kiln is subjected to the processes of temperature rise, heat preservation and temperature reduction once, the ceramic changes in a complex way along with the temperature in the kiln, and the temperature in the kiln plays a key role in the firing quality of the ceramic.
The firing technology of the traditional ceramic kiln is still relatively backward, the temperature control of most small and medium-sized ceramic kilns basically stays at the level of manual operation and simple instrument operation, and the kiln temperature is manually controlled according to firing experience of firing workers, so that the efficiency is low, and the firing quality is unstable.
When the environment-friendly daily ceramic is prepared, the ceramic is required to be placed into a kiln for firing, and when the ceramic is fired, the temperature control in the kiln is very important, so that the quality of the finally formed ceramic is directly influenced, and an optimized preparation process of the environment-friendly daily ceramic is expected aiming at the existing technical problems.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a preparation process and a system of environment-friendly daily ceramic, which comprehensively utilize thermal infrared monitoring videos of the environment-friendly daily ceramic and firing temperature values of a plurality of preset time points, and combine a deep learning technology and an artificial intelligence technology to carry out self-adaptive control on the firing temperature based on the real-time condition of ceramic firing, so as to further provide an intelligent firing temperature control scheme. The firing temperature control scheme can control the firing temperature of the ceramic in a self-adaptive manner based on the real-time condition of ceramic firing, so that manual intervention is reduced as much as possible, the production efficiency is improved, and meanwhile, the stable quality of the environment-friendly daily ceramic can be effectively ensured.
Correspondingly, according to one aspect of the application, a preparation process of the environment-friendly domestic ceramic is provided, which comprises the following steps: acquiring thermal infrared monitoring videos of the environment-friendly domestic ceramic in a preset time period acquired by a thermal infrared camera, and firing temperature values of a plurality of preset time points in the preset time period; extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic; the thermal infrared monitoring keys are respectively processed through a first convolution neural network model using spatial attention so as to obtain a plurality of ceramic firing heat distribution characteristic matrixes; arranging the ceramic firing heat distribution feature matrixes into three-dimensional input tensors according to time dimension, and then obtaining a ceramic firing heat distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel; reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram to be a ceramic firing heat distribution time sequence characteristic vector; the firing temperature values of the plurality of preset time points are arranged into input vectors according to time dimensions, and then a one-dimensional convolutional neural network model is used for obtaining temperature time sequence feature vectors; calculating the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector to obtain a classification feature matrix; and passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point is increased or decreased.
In the preparation process of the environment-friendly daily ceramic, the plurality of thermal infrared monitoring keys are respectively processed by a first convolution neural network model using spatial attention to obtain a plurality of ceramic firing heat distribution characteristic matrixes, and the preparation process comprises the steps of respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on the plurality of thermal infrared monitoring keys in forward transmission of layers by using each layer of the first convolution neural network model to output a plurality of initial ceramic firing heat distribution characteristic matrixes by the last layer of the first convolution neural network model; and inputting the plurality of initial ceramic firing heat distribution feature matrices into a spatial attention layer of the first convolutional neural network model to obtain the plurality of ceramic firing heat distribution feature matrices.
In the above preparation process of the environmental protection domestic ceramic, the step of arranging the ceramic firing heat distribution feature matrices into three-dimensional input tensors according to a time dimension and then obtaining a ceramic firing heat distribution time sequence feature diagram by using a second convolution neural network model of a three-dimensional convolution kernel includes: input data are respectively carried out in forward transfer by using each layer of the second convolution neural network model using the three-dimensional convolution kernel: carrying out convolution processing on input data to obtain a convolution characteristic diagram; pooling the convolution feature images based on the local feature matrix to obtain pooled feature images; performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the second convolutional neural network using the three-dimensional convolutional kernel is the ceramic firing thermal distribution time sequence characteristic diagram, and the input of the first layer of the second convolutional neural network using the three-dimensional convolutional kernel is the three-dimensional characteristic tensor.
In the above process for preparing the environmentally-friendly domestic ceramic, the step of arranging firing temperature values at the plurality of predetermined time points into input vectors according to a time dimension and then obtaining temperature time sequence feature vectors through a one-dimensional convolutional neural network model includes: and respectively carrying out convolution processing, mean pooling processing and nonlinear activation processing based on a one-dimensional convolution kernel on an input vector in forward transfer of layers by using each layer of the one-dimensional convolution neural network model so as to take the output of the last layer of the one-dimensional convolution neural network model as the temperature time sequence feature vector.
In the above preparation process of the environmental-friendly domestic ceramic, the calculating the responsiveness estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector to obtain the classification feature matrix includes: calculating Helmholtz free energy factors of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector to obtain a first Helmholtz free energy factor and a second Helmholtz free energy factor; weighting the ceramic firing heat distribution timing feature vector and the temperature timing feature vector based on the first helmholtz class free energy factor and the second helmholtz class free energy factor to obtain a weighted ceramic firing heat distribution timing feature vector and a weighted temperature timing feature vector; and calculating a responsiveness estimate of the weighted ceramic firing profile timing feature vector relative to the weighted temperature timing feature vector to obtain the classification feature matrix.
In the above preparation process of the environmental protection domestic ceramic, the calculating the helmholtz free energy factors of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector to obtain a first helmholtz free energy factor and a second helmholtz free energy factor includes: calculating the time sequence characteristic vector of the ceramic firing heat distribution according to the following formulaAnd the temperature timing feature vector +.>Is of the Helmholtz class free energyA measurement factor; wherein, the formula is:wherein (1)>Classification probability value representing the ceramic firing heat distribution time sequence feature vector, < >>Classification probability value representing the temperature timing feature vector,/->Representing the ceramic firing heat distribution time sequence characteristic vector +.>Characteristic value of the location->Representing the temperature time sequence characteristic vector +.>Characteristic value of the location->A Helmholtz-like free energy factor representing a timing characteristic vector of the firing profile of the ceramic, +.>Helmholtz class free energy factor representing the temperature timing feature vector, ++>Represents a logarithmic function with base 2, +.>Represents an exponential operation based on a natural constant e, and +.>Is the length of the feature vector.
In the above process for preparing an environmental friendly domestic ceramic, the calculating the responsiveness estimation of the weighted ceramic firing heat distribution time sequence feature vector relative to the weighted temperature time sequence feature vector to obtain the classification feature matrix includes: calculating a responsiveness estimate of the weighted ceramic firing profile timing feature vector relative to the weighted temperature timing feature vector to obtain a classification feature matrix; wherein, the formula is:wherein->Representing the time sequence characteristic vector of the heat distribution of the ceramic firing after weighting>Representing the weighted temperature timing feature vector, < >>Representing the classification feature matrix,/->Representing matrix multiplication.
In the above process for preparing the environmentally friendly domestic ceramic, the step of passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the firing temperature value at the current time point should be increased or decreased, and comprises the following steps: expanding the classification feature matrix into classification feature vectors according to row vectors or column vectors; performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.
According to another aspect of the present application, there is provided a preparation system of an environmental-friendly domestic ceramic, comprising: the monitoring module is used for acquiring thermal infrared monitoring videos of the environment-friendly daily ceramic in a preset time period and firing temperature values of a plurality of preset time points in the preset time period, wherein the thermal infrared monitoring videos are acquired by the thermal infrared camera; the sampling module is used for extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic; the spatial feature extraction module is used for obtaining a plurality of ceramic firing heat distribution feature matrixes by using a first convolution neural network model of spatial attention; the thermal distribution time sequence feature extraction module is used for arranging the ceramic firing thermal distribution feature matrixes into three-dimensional input tensors according to a time dimension and then obtaining a ceramic firing thermal distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel; the dimension change module is used for reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram into a ceramic firing heat distribution time sequence characteristic vector; the temperature time sequence feature extraction module is used for arranging firing temperature values of the plurality of preset time points into input vectors according to time dimensions and obtaining temperature time sequence feature vectors through a one-dimensional convolutional neural network model; the responsiveness estimation module is used for calculating responsiveness estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector so as to obtain a classification feature matrix; and the temperature control result generation module is used for passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point should be increased or decreased.
In the preparation system of the environment-friendly daily ceramic, the spatial feature extraction module comprises a depth convolution coding unit, a first convolution neural network module and a second convolution neural network module, wherein the depth convolution coding unit is used for respectively carrying out convolution processing, pooling processing and nonlinear activation processing on the thermal infrared monitoring keys in forward transmission of layers by using each layer of the first convolution neural network module so as to output a plurality of initial ceramic firing thermal distribution feature matrixes from the last layer of the first convolution neural network module; and a spatial attention unit for inputting the plurality of initial ceramic firing heat distribution feature matrices into a spatial attention layer of the first convolutional neural network model to obtain the plurality of ceramic firing heat distribution feature matrices.
Compared with the prior art, the preparation process and the system of the environment-friendly daily ceramic provided by the application comprehensively utilize the thermal infrared monitoring video of the environment-friendly daily ceramic and firing temperature values of a plurality of preset time points, and carry out self-adaptive control on the firing temperature based on the real-time condition of ceramic firing by combining a deep learning technology and an artificial intelligence technology, so that an intelligent firing temperature control scheme is provided. The firing temperature control scheme can control the firing temperature of the ceramic in a self-adaptive manner based on the real-time condition of ceramic firing, so that manual intervention is reduced as much as possible, the production efficiency is improved, and meanwhile, the stable quality of the environment-friendly daily ceramic can be effectively ensured.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
Fig. 1 is an application scenario diagram of a preparation process of an environment-friendly domestic ceramic according to an embodiment of the application.
Fig. 2 is a flowchart of a preparation process of the environment-friendly domestic ceramic according to the embodiment of the application.
Fig. 3 is a schematic diagram of the construction of a process for preparing an environment-friendly domestic ceramic according to an embodiment of the application.
Fig. 4 is a flowchart of a process for preparing an environment-friendly domestic ceramic according to an embodiment of the application, wherein the plurality of thermal infrared monitoring keys are used for obtaining a plurality of ceramic firing heat distribution feature matrices by using a first convolution neural network model of spatial attention.
Fig. 5 is a flowchart of a ceramic firing heat distribution time sequence feature diagram obtained by using a second convolution neural network model of a three-dimensional convolution kernel after arranging the ceramic firing heat distribution feature matrices into a three-dimensional input tensor according to a time dimension in the preparation process of the environment-friendly domestic ceramic according to the embodiment of the application.
Fig. 6 is a flowchart of calculating a responsiveness estimation of the ceramic firing heat distribution time series eigenvector relative to the temperature time series eigenvector to obtain a classification eigenvector in the preparation process of the environmental friendly domestic ceramic according to the embodiment of the application.
Fig. 7 is a flowchart showing that the firing temperature value at the current time point should be increased or decreased by passing the classification feature matrix through a classifier in the preparation process of the environmentally-friendly domestic ceramic according to the embodiment of the present application to obtain a classification result.
Fig. 8 is a block diagram of a preparation system of an environmentally friendly domestic ceramic according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Summary of the application: aiming at the technical problems, the technical conception of the application is as follows: the method is characterized by comprehensively utilizing a thermal infrared monitoring video of the environment-friendly daily ceramic and firing temperature values of a plurality of preset time points, and carrying out self-adaptive control on the firing temperature based on the real-time condition of ceramic firing by combining a deep learning technology and an artificial intelligence technology.
Specifically, in the technical scheme of the application, firstly, a thermal infrared monitoring video of the environment-friendly daily ceramic in a preset time period acquired by a thermal infrared camera and firing temperature values of a plurality of preset time points in the preset time period are acquired. In the firing process of the ceramic, the temperature is one of the key factors, and the acquisition of the thermal infrared monitoring video can quickly acquire the temperature distribution condition of the ceramic surface in real time.
Since the thermal infrared monitoring video contains a large number of redundant frames, such as continuous similar frames, the direct use of the video for feature extraction causes huge calculation overhead and is easy to cause over-fitting problem. In the technical scheme, a plurality of thermal infrared monitoring keys are extracted from the thermal infrared monitoring video of the environment-friendly daily ceramic so as to reduce redundant information while keeping important temperature distribution information, thereby accelerating the calculation speed of the model and improving the generalization capability of the model to the ceramic firing process.
Considering that in the thermal infrared monitoring video, the degree of influence of temperature information at different positions on the firing quality of ceramics may be different, and the temperature at some positions may have a greater influence on the quality of the finally formed ceramics. In the technical scheme of the application, the plurality of thermal infrared monitoring keys are respectively used for obtaining a plurality of ceramic firing heat distribution characteristic matrixes by using a first convolution neural network model of spatial attention. Here, the spatial attention mechanism is a technology commonly used in deep learning, and different weights can be given to different areas of input data, so as to emphasize or suppress features of different positions. That is, the attention mechanism can be introduced into the model by using the first convolution neural network model of the spatial attention by the plurality of thermal infrared monitoring keys, so that the model can automatically learn and utilize the characteristic information of different positions to model and predict, and the accuracy and the robustness of the model are improved.
During firing of the ceramic, the temperature profile is time-varying. In order to capture implicit associated distribution information of the temperature of the ceramic surface in the time dimension, in the technical scheme of the application, the ceramic firing heat distribution characteristic matrixes are arranged into three-dimensional input tensors according to the time dimension, and then a second convolution neural network model of a three-dimensional convolution kernel is used for obtaining a ceramic firing heat distribution time sequence characteristic diagram. Here, the second convolutional neural network model utilizes its three-dimensional convolutional kernel to effectively capture the time sequence features in the ceramic firing process, improving the understanding and predicting capabilities of the model with respect to time sequence data.
And then, arranging firing temperature values of the plurality of preset time points into input vectors according to a time dimension, and obtaining temperature time sequence feature vectors through a one-dimensional convolutional neural network model. Here, the one-dimensional convolutional neural network is a model for processing data having time-series characteristics, and can automatically learn the time-series characteristics of the data, thereby realizing modeling of complex relationships. In other words, the firing temperature also shows dynamic change in the time dimension, and in the technical scheme of the application, the firing temperature value is converted into the time sequence feature vector through the one-dimensional convolutional neural network model, so that the model considers the dynamic change and trend of the firing temperature more comprehensively.
The change of the firing temperature can cause the change of the surface temperature of the environment-friendly daily ceramic, and if the response association relationship between the firing temperature and the surface temperature can be utilized, the accuracy of the subsequent classification can be obviously improved. Thus, in the technical solution of the present application, first, the ceramic firing heat distribution time series feature map is reduced in dimension to a ceramic firing heat distribution time series feature vector to convert a three-dimensional feature map to a one-dimensional feature vector. Then, a responsiveness estimate of the ceramic firing heat distribution timing feature vector relative to the temperature timing feature vector is calculated to obtain a classification feature matrix. That is, the influence of firing temperature variation on the thermal distribution in the ceramic firing process is characterized by the classification characteristic matrix.
After the classification feature matrix is obtained, it is passed through a classifier to obtain a classification result indicating whether the firing temperature value at the current time point should be increased or decreased. Specifically, after the classifier is trained, the classification feature matrix may be matched with preset label information (that is, the firing temperature value at the current time point should be increased and the firing temperature value at the current time point should be decreased), so as to determine the adjustment direction of the firing temperature at the current time point. In this way, the current firing temperature is adapted to the real-time temperature variation during the firing of the ceramic to improve the stability of the ceramic formation.
In the technical scheme of the application, when the ceramic firing heat distribution time sequence feature vector is calculated to obtain the classification feature matrix relative to the temperature time sequence feature vector, as the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector respectively express the channel dimension distribution of the image semantic features of the thermal infrared monitoring key frame and the time sequence distribution of firing temperature values, the significant difference of the source data and the feature distribution can cause the existence of weak class correlation distribution examples relative to class labels of the classifier in the integral feature distribution of each of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector, that is, the compatibility of the ceramic firing heat distribution time sequence feature vector and the integral feature distribution of the temperature time sequence feature vector under the class labels of the classifier is lower, which can influence the calculation accuracy of the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector, thereby influencing the accuracy of classification results obtained by the classifier of the classification feature matrix.
Based on this, it is preferable to calculate the ceramic firing heat distribution timing characteristic vector separately And the temperature timing feature vector +.>Is a helmholtz-like free energy factor; wherein, the formula is:wherein (1)>Classification probability value representing the ceramic firing heat distribution time sequence feature vector, < >>Classification probability value representing the temperature timing feature vector,/->Representing the ceramic firing heat distribution time sequence characteristic vector +.>Characteristic value of the location->Representing the temperature time sequence characteristic vector +.>Characteristic value of the location->A Helmholtz-like free energy factor representing a timing characteristic vector of the firing profile of the ceramic, +.>Helmholtz class free energy factor representing the temperature timing feature vector, ++>Represents a logarithmic function with base 2, +.>Represents an exponential operation based on a natural constant e, and +.>Is the length of the feature vector.
Here, the ceramic firing heat distribution timing characteristic vector can be calculated based on the helmholtz free energy formulaAnd the temperature timing feature vector +.>The respective feature value sets describe the energy values of the predetermined class labels as class free energies of the feature vector as a whole by using the feature values as the feature vectors for the ceramic firing heat distribution timing sequence feature vector ∈ ->And the temperature timing feature vector +.>Weighting is performed to obtain the ceramic firing heat distribution timing characteristic vector +. >And the temperature timing feature vector +.>Focusing on class-related prototype example distributions of features overlapping with true example distributions in class target domain so as to be in the ceramic firing heat distribution timing feature vector +.>And the temperature timing feature vector +.>And under the condition that a similar weak correlation distribution example exists in the integral feature distribution, incremental learning is realized by carrying out fuzzy labeling on the integral feature distribution, so that the compatibility of the integral feature distribution under a similar label is improved, the calculation accuracy of the response estimation of the time sequence feature vector of the ceramic firing thermal distribution relative to the temperature time sequence feature vector is improved, and the accuracy of a classification result obtained by a classifier through the classification feature matrix is improved.
Based on this, this application provides a preparation technology of environmental protection domestic ceramics, it includes: acquiring thermal infrared monitoring videos of the environment-friendly domestic ceramic in a preset time period acquired by a thermal infrared camera, and firing temperature values of a plurality of preset time points in the preset time period; extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic; the thermal infrared monitoring keys are respectively processed through a first convolution neural network model using spatial attention so as to obtain a plurality of ceramic firing heat distribution characteristic matrixes; arranging the ceramic firing heat distribution feature matrixes into three-dimensional input tensors according to time dimension, and then obtaining a ceramic firing heat distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel; reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram to be a ceramic firing heat distribution time sequence characteristic vector; the firing temperature values of the plurality of preset time points are arranged into input vectors according to time dimensions, and then a one-dimensional convolutional neural network model is used for obtaining temperature time sequence feature vectors; calculating the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector to obtain a classification feature matrix; and passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point is increased or decreased.
Fig. 1 is an application scenario diagram of a preparation process of an environment-friendly domestic ceramic according to an embodiment of the application. As shown in fig. 1, in this application scenario, first, a thermal infrared monitoring video of an environmentally-friendly daily ceramic (e.g., L as illustrated in fig. 1) for a predetermined period of time is acquired by a thermal infrared camera (e.g., P as illustrated in fig. 1), and firing temperature values at a plurality of predetermined time points within the predetermined period of time are acquired by a thermometer (e.g., M as illustrated in fig. 1). Next, the above information is input into a server (e.g., S as illustrated in fig. 1) deployed with an algorithm for environmentally friendly domestic ceramic preparation, wherein the server is capable of processing the above input information with the environmentally friendly domestic ceramic preparation algorithm to generate a classification result indicating that the firing temperature value at the current time point should be increased or decreased.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
An exemplary method is: fig. 2 is a flowchart of a preparation process of the environment-friendly domestic ceramic according to the embodiment of the application. As shown in fig. 2, the preparation process of the environment-friendly domestic ceramic according to the embodiment of the application comprises the following steps: s110, acquiring thermal infrared monitoring videos of the environment-friendly daily ceramic in a preset time period and firing temperature values of a plurality of preset time points in the preset time period, wherein the thermal infrared monitoring videos are acquired by a thermal infrared camera; s120, extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic; s130, the thermal infrared monitoring keys are respectively processed through a first convolution neural network model using spatial attention to obtain a plurality of ceramic firing heat distribution characteristic matrixes; s140, arranging the ceramic firing heat distribution feature matrixes into three-dimensional input tensors according to a time dimension, and then obtaining a ceramic firing heat distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel; s150, reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram to be a ceramic firing heat distribution time sequence characteristic vector; s160, arranging firing temperature values of the plurality of preset time points into input vectors according to a time dimension, and then obtaining temperature time sequence feature vectors through a one-dimensional convolutional neural network model; s170, calculating the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector to obtain a classification feature matrix; and S180, passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point is increased or decreased.
Fig. 3 is a schematic diagram of the construction of a process for preparing an environment-friendly domestic ceramic according to an embodiment of the application. As shown in fig. 3, first, a thermal infrared monitoring video of an environmental-friendly domestic ceramic for a predetermined period of time acquired by a thermal infrared camera and firing temperature values at a plurality of predetermined time points within the predetermined period of time are acquired; then, extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic; then, the thermal infrared monitoring keys are respectively processed through a first convolution neural network model using spatial attention so as to obtain a plurality of ceramic firing heat distribution characteristic matrixes; arranging the ceramic firing heat distribution feature matrixes into three-dimensional input tensors according to time dimension, and obtaining a ceramic firing heat distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel; then, reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram to be a ceramic firing heat distribution time sequence characteristic vector; meanwhile, the firing temperature values of the plurality of preset time points are arranged into input vectors according to time dimensions, and then a one-dimensional convolutional neural network model is used for obtaining temperature time sequence feature vectors; then, calculating the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector to obtain a classification feature matrix; finally, the classification feature matrix is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point should be increased or decreased.
In step S110, a thermal infrared monitoring video of the environmental-friendly domestic ceramic for a predetermined period of time acquired by the thermal infrared camera and firing temperature values at a plurality of predetermined time points within the predetermined period of time are acquired. In the firing process of the ceramic, the temperature is one of the key factors, and the acquisition of the thermal infrared monitoring video can quickly acquire the temperature distribution condition of the ceramic surface in real time. Therefore, in the technical scheme of the application, the thermal infrared monitoring video of the environment-friendly daily ceramic and firing temperature values of a plurality of preset time points are utilized, and the deep learning technology and the artificial intelligence technology are combined to carry out self-adaptive control on the firing temperature based on the real-time condition of ceramic firing. Based on the above, firstly, a thermal infrared monitoring video of the environment-friendly domestic ceramic in a preset time period is acquired through a thermal infrared camera, and firing temperature values of a plurality of preset time points in the preset time period are acquired through a thermometer.
In step S120, a plurality of thermal infrared monitoring keys are extracted from the thermal infrared monitoring video of the environmental-friendly domestic ceramic. Since the thermal infrared monitoring video contains a large number of redundant frames, such as continuous similar frames, the direct use of the video for feature extraction causes huge calculation overhead and is easy to cause over-fitting problem. In the technical scheme, a plurality of thermal infrared monitoring keys are extracted from the thermal infrared monitoring video of the environment-friendly daily ceramic so as to reduce redundant information while keeping important temperature distribution information, thereby accelerating the calculation speed of the model and improving the generalization capability of the model to the ceramic firing process. In specific implementation, a plurality of thermal infrared monitoring keys can be extracted from the thermal infrared monitoring video of the environment-friendly daily ceramic at a preset sampling frequency.
In step S130, the plurality of thermal infrared monitoring keys are respectively used to obtain a plurality of ceramic firing heat distribution feature matrices through a first convolution neural network model using spatial attention. Considering that in the thermal infrared monitoring video, the degree of influence of temperature information at different positions on the firing quality of ceramics may be different, and the temperature at some positions may have a greater influence on the quality of the finally formed ceramics. In the technical scheme of the application, the plurality of thermal infrared monitoring keys are respectively used for obtaining a plurality of ceramic firing heat distribution characteristic matrixes by using a first convolution neural network model of spatial attention. Here, the spatial attention mechanism is a technology commonly used in deep learning, and different weights can be given to different areas of input data, so as to emphasize or suppress features of different positions. That is, the attention mechanism can be introduced into the model by using the first convolution neural network model of the spatial attention by the plurality of thermal infrared monitoring keys, so that the model can automatically learn and utilize the characteristic information of different positions to model and predict, and the accuracy and the robustness of the model are improved.
Fig. 4 is a flowchart of a process for preparing an environment-friendly domestic ceramic according to an embodiment of the application, wherein the plurality of thermal infrared monitoring keys are used for obtaining a plurality of ceramic firing heat distribution feature matrices by using a first convolution neural network model of spatial attention. As shown in fig. 4, the step S130 includes: s131, respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on the thermal infrared monitoring keys in forward transmission of layers by using each layer of the first convolutional neural network model so as to output a plurality of initial ceramic firing thermal distribution feature matrixes by the last layer of the first convolutional neural network model; and S132, inputting the plurality of initial ceramic firing heat distribution feature matrices into a spatial attention layer of the first convolutional neural network model to obtain the plurality of ceramic firing heat distribution feature matrices.
In step S140, the plurality of ceramic firing heat distribution feature matrices are arranged into three-dimensional input tensors according to a time dimension, and then a ceramic firing heat distribution time sequence feature diagram is obtained through a second convolution neural network model using a three-dimensional convolution kernel. During firing of the ceramic, the temperature profile is time-varying. In order to capture implicit associated distribution information of the temperature of the ceramic surface in the time dimension, in the technical scheme of the application, the ceramic firing heat distribution characteristic matrixes are arranged into three-dimensional input tensors according to the time dimension, and then a second convolution neural network model of a three-dimensional convolution kernel is used for obtaining a ceramic firing heat distribution time sequence characteristic diagram. Here, the second convolutional neural network model utilizes its three-dimensional convolutional kernel to effectively capture the time sequence features in the ceramic firing process, improving the understanding and predicting capabilities of the model with respect to time sequence data.
Fig. 5 is a flowchart of a ceramic firing heat distribution time sequence feature diagram obtained by using a second convolution neural network model of a three-dimensional convolution kernel after arranging the ceramic firing heat distribution feature matrices into a three-dimensional input tensor according to a time dimension in the preparation process of the environment-friendly domestic ceramic according to the embodiment of the application. As shown in fig. 5, the step S140 includes: input data are respectively carried out in forward transfer by using each layer of the second convolution neural network model using the three-dimensional convolution kernel: s141, carrying out convolution processing on input data to obtain a convolution characteristic diagram; s142, pooling the convolution feature images based on the local feature matrix to obtain pooled feature images; s143, carrying out nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the second convolutional neural network using the three-dimensional convolutional kernel is the ceramic firing thermal distribution time sequence characteristic diagram, and the input of the first layer of the second convolutional neural network using the three-dimensional convolutional kernel is the three-dimensional characteristic tensor.
In step S150, the ceramic firing heat distribution time series characteristic map is reduced in dimension to a ceramic firing heat distribution time series characteristic vector. In order to perform subsequent data processing, a response association relation is established with firing temperature, and the ceramic firing heat distribution time sequence feature map is reduced in dimension to be a ceramic firing heat distribution time sequence feature vector so as to convert the three-dimensional feature map into a one-dimensional feature vector.
In step S160, the firing temperature values at the predetermined time points are arranged according to the time dimension as input vectors, and then the input vectors are passed through a one-dimensional convolutional neural network model to obtain temperature time sequence feature vectors. Here, the one-dimensional convolutional neural network is a model for processing data having time-series characteristics, and can automatically learn the time-series characteristics of the data, thereby realizing modeling of complex relationships. In other words, the firing temperature also shows dynamic change in the time dimension, and in the technical scheme of the application, the firing temperature value is converted into the time sequence feature vector through the one-dimensional convolutional neural network model, so that the model considers the dynamic change and trend of the firing temperature more comprehensively. Specifically, each layer of the one-dimensional convolutional neural network model is used for respectively carrying out convolution processing based on a one-dimensional convolutional kernel, mean pooling processing and nonlinear activation processing on an input vector in forward transfer of the layer so that the output of the last layer of the one-dimensional convolutional neural network model is used as the temperature time sequence feature vector.
In step S170, a responsiveness estimate of the ceramic firing heat distribution timing eigenvector relative to the temperature timing eigenvector is calculated to obtain a classification eigenvector. The change of the firing temperature can cause the change of the surface temperature of the environment-friendly daily ceramic, and if the response association relationship between the firing temperature and the surface temperature can be utilized, the accuracy of the subsequent classification can be obviously improved. Therefore, in the technical scheme of the application, the responsiveness estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector is calculated to obtain a classification feature matrix. That is, the influence of firing temperature variation on the thermal distribution in the ceramic firing process is characterized by the classification characteristic matrix.
Fig. 6 is a flowchart of calculating a responsiveness estimation of the ceramic firing heat distribution time series eigenvector relative to the temperature time series eigenvector to obtain a classification eigenvector in the preparation process of the environmental friendly domestic ceramic according to the embodiment of the application. As shown in fig. 6, the step S170 includes: s171, calculating Helmholtz type free energy factors of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector to obtain a first Helmholtz type free energy factor and a second Helmholtz type free energy factor; s172, weighting the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector based on the first Helmholtz class free energy factor and the second Helmholtz class free energy factor to obtain a weighted ceramic firing heat distribution time sequence feature vector and a weighted temperature time sequence feature vector; and S173, calculating the response estimation of the weighted ceramic firing heat distribution time sequence feature vector relative to the weighted temperature time sequence feature vector to obtain the classification feature matrix.
In step S171, the helmholtz-type free energy factors of the ceramic firing heat distribution timing eigenvector and the temperature timing eigenvector are calculated to obtain a first helmholtz-type free energy factor and a second helmholtz-type free energy factor. In the technical scheme of the application, when the ceramic firing heat distribution time sequence feature vector is calculated to obtain the classification feature matrix relative to the temperature time sequence feature vector, as the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector respectively express the channel dimension distribution of the image semantic features of the thermal infrared monitoring key frame and the time sequence distribution of firing temperature values, the significant difference of the source data and the feature distribution can cause the existence of weak class correlation distribution examples relative to class labels of the classifier in the integral feature distribution of each of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector, that is, the compatibility of the ceramic firing heat distribution time sequence feature vector and the integral feature distribution of the temperature time sequence feature vector under the class labels of the classifier is lower, which can influence the calculation accuracy of the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector, thereby influencing the accuracy of classification results obtained by the classifier of the classification feature matrix.
Based on this, it is preferable to calculate the ceramic firing heat distribution timing characteristic vector separatelyAnd the temperature timing feature vector +.>The helmholtz-like free energy factor of (c) is specifically: />Wherein,classification probability value representing the ceramic firing heat distribution time sequence feature vector, < >>Classification probability value representing the temperature timing feature vector,/->Representing the ceramic firing heat distribution time sequence characteristic vector +.>Characteristic value of the location->Representing the temperature time sequence characteristic vector +.>Characteristic value of the location->A Helmholtz-like free energy factor representing a timing characteristic vector of the firing profile of the ceramic, +.>A helmholtz-like free energy factor representing the temperature timing feature vector,represents a logarithmic function with base 2, +.>Represents an exponential operation based on a natural constant e, and +.>Is the length of the feature vector.
Here, the ceramic firing heat distribution timing characteristic vector can be calculated based on the helmholtz free energy formulaAnd the temperature timing feature vector +.>Each of which is a single pieceThe characteristic value set describes the energy value of the predetermined class label as the class free energy of the characteristic vector whole, and the ceramic firing heat distribution time sequence characteristic vector is +. >And the temperature timing feature vector +.>Weighting is performed to obtain the ceramic firing heat distribution timing characteristic vector +.>And the temperature timing feature vector +.>Focusing on class-related prototype example distributions of features overlapping with true example distributions in class target domain so as to be in the ceramic firing heat distribution timing feature vector +.>And the temperature timing feature vector +.>And under the condition that a similar weak correlation distribution example exists in the integral feature distribution, incremental learning is realized by carrying out fuzzy labeling on the integral feature distribution, so that the compatibility of the integral feature distribution under a similar label is improved, the calculation accuracy of the response estimation of the time sequence feature vector of the ceramic firing thermal distribution relative to the temperature time sequence feature vector is improved, and the accuracy of a classification result obtained by a classifier through the classification feature matrix is improved.
In step S173, calculating a responsiveness estimate of the weighted ceramic firing profile timing feature vector relative to the weighted temperature timing feature vector to obtain the classification feature matrix, comprising: calculating a responsiveness estimate of the weighted ceramic firing profile timing feature vector relative to the weighted temperature timing feature vector to obtain a classification feature matrix; wherein the male part The formula is:wherein->Representing the time sequence characteristic vector of the heat distribution of the ceramic firing after weighting>Representing the weighted temperature timing feature vector, < >>Representing the matrix of the classification characteristic,representing matrix multiplication.
In step S180, the classification feature matrix is passed through a classifier to obtain a classification result indicating whether the firing temperature value at the current time point should be increased or decreased. After the classification feature matrix is obtained, it is passed through a classifier to obtain a classification result indicating whether the firing temperature value at the current time point should be increased or decreased. Specifically, after the classifier is trained, the classification feature matrix may be matched with preset label information (that is, the firing temperature value at the current time point should be increased and the firing temperature value at the current time point should be decreased), so as to determine the adjustment direction of the firing temperature at the current time point. In this way, the current firing temperature is adapted to the real-time temperature variation during the firing of the ceramic to improve the stability of the ceramic formation.
Fig. 7 is a flowchart showing that the firing temperature value at the current time point should be increased or decreased by passing the classification feature matrix through a classifier in the preparation process of the environmentally-friendly domestic ceramic according to the embodiment of the present application to obtain a classification result. As shown in fig. 7, the step S180 includes: s181, expanding the classification feature matrix into classification feature vectors according to row vectors or column vectors; s182, performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and S183, inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.
In summary, the preparation process of the environmental-friendly daily ceramic according to the embodiment of the application is clarified, which comprehensively utilizes the thermal infrared monitoring video of the environmental-friendly daily ceramic and firing temperature values of a plurality of preset time points, and combines a deep learning technology and an artificial intelligence technology to carry out self-adaptive control on the firing temperature based on the real-time condition of ceramic firing, so as to further provide an intelligent firing temperature control scheme. The firing temperature control scheme can control the firing temperature of the ceramic in a self-adaptive manner based on the real-time condition of ceramic firing, so that manual intervention is reduced as much as possible, the production efficiency is improved, and meanwhile, the stable quality of the environment-friendly daily ceramic can be effectively ensured.
Exemplary System: fig. 8 is a block diagram of a preparation system of an environmentally friendly domestic ceramic according to an embodiment of the application. As shown in fig. 8, a preparation system 100 of an environment-friendly domestic ceramic according to an embodiment of the present application includes: the monitoring module 110 is used for acquiring the thermal infrared monitoring video of the environmental-friendly daily ceramic in a preset time period acquired by the thermal infrared camera and firing temperature values of a plurality of preset time points in the preset time period; the sampling module 120 is used for extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic; the spatial feature extraction module 130 is configured to obtain a plurality of ceramic firing thermal distribution feature matrices by using the first convolutional neural network model of spatial attention for the plurality of thermal infrared monitoring keys respectively; the thermal distribution time sequence feature extraction module 140 is configured to arrange the plurality of ceramic firing thermal distribution feature matrices into a three-dimensional input tensor according to a time dimension, and obtain a ceramic firing thermal distribution time sequence feature map by using a second convolution neural network model of a three-dimensional convolution kernel; the dimension change module 150 is configured to reduce the dimension of the ceramic firing heat distribution time sequence feature map to a ceramic firing heat distribution time sequence feature vector; the temperature time sequence feature extraction module 160 is configured to arrange firing temperature values of the plurality of predetermined time points into input vectors according to a time dimension, and then obtain temperature time sequence feature vectors through a one-dimensional convolutional neural network model; a responsiveness estimation module 170 for calculating a responsiveness estimate of the ceramic firing heat distribution timing feature vector relative to the temperature timing feature vector to obtain a classification feature matrix; and a temperature control result generation module 180, configured to pass the classification feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate that the firing temperature value at the current time point should be increased or decreased.
In one example, in the preparation system 100 of the environmental-friendly domestic ceramic, the spatial feature extraction module 130 includes: the depth convolution coding unit is used for respectively carrying out convolution processing, pooling processing and nonlinear activation processing on the thermal infrared monitoring keys in forward transmission of layers by using each layer of the first convolution neural network model so as to output a plurality of initial ceramic firing thermal distribution feature matrixes by the last layer of the first convolution neural network model; and a spatial attention unit for inputting the plurality of initial ceramic firing heat distribution feature matrices into a spatial attention layer of the first convolutional neural network model to obtain the plurality of ceramic firing heat distribution feature matrices.
In one example, in the preparation system 100 of the environmental-friendly domestic ceramic, the thermal distribution timing feature extraction module 140 is further configured to: input data are respectively carried out in forward transfer by using each layer of the second convolution neural network model using the three-dimensional convolution kernel: performing convolution processing on the input data based on the three-dimensional convolution kernel to obtain a convolution feature map; pooling the convolution feature images based on the local feature matrix to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the output of the last layer of the second convolutional neural network using the three-dimensional convolutional kernel is the ceramic firing thermal distribution time sequence characteristic diagram, and the input of the first layer of the second convolutional neural network using the three-dimensional convolutional kernel is the three-dimensional characteristic tensor.
In one example, in the preparation system 100 of the environmental-friendly domestic ceramic, the temperature time sequence feature extraction module 160 includes: and respectively carrying out convolution processing, mean pooling processing and nonlinear activation processing based on a one-dimensional convolution kernel on an input vector in forward transfer of layers by using each layer of the one-dimensional convolution neural network model so as to take the output of the last layer of the one-dimensional convolution neural network model as the temperature time sequence feature vector.
In one example, in the preparation system 100 of the environmental friendly domestic ceramic, the responsiveness estimation module 170 includes: the computing unit is used for computing the Helmholtz free energy factors of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector to obtain a first Helmholtz free energy factor and a second Helmholtz free energy factor; the weight applying unit is used for weighting the ceramic firing heat distribution time sequence characteristic vector and the temperature time sequence characteristic vector based on the first Helmholtz class free energy factor and the second Helmholtz class free energy factor to obtain a weighted ceramic firing heat distribution time sequence characteristic vector and a weighted temperature time sequence characteristic vector; and a responsiveness unit for calculating responsiveness estimation of the weighted ceramic firing heat distribution time sequence eigenvector relative to the weighted temperature time sequence eigenvector to obtain the classification eigenvalue matrix.
In one example, in the preparation system 100 of the above-described environmentally friendly domestic ceramic, the calculation unit includes: calculating the time sequence characteristic vector of the ceramic firing heat distribution according to the following formulaAnd the temperature timing feature vector +.>Is a helmholtz-like free energy factor; wherein, the formula is: />Wherein (1)>Classification probability value representing the ceramic firing heat distribution time sequence feature vector, < >>Classification probability value representing the temperature timing feature vector,/->Representing the ceramic firing heat distribution time sequence characteristic vector +.>Characteristic value of the location->Representing the temperature time sequence characteristic vector +.>Characteristic value of the location->A Helmholtz-like free energy factor representing a timing characteristic vector of the firing profile of the ceramic, +.>Helmholtz class free energy factor representing the temperature timing feature vector, ++>Represents a logarithmic function with base 2, +.>Represents an exponential operation based on a natural constant e, and +.>Is the length of the feature vector.
In one example, in the preparation system 100 of the above-described environmentally friendly domestic ceramic, the responsive unit includes: calculating a responsiveness estimate of the weighted ceramic firing heat distribution timing eigenvector relative to the weighted temperature timing eigenvector to obtain a classification characteristic by the formula A sign matrix; wherein, the formula is:wherein->Representing the time sequence characteristic vector of the heat distribution of the ceramic firing after weighting>Representing the weighted temperature timing feature vector, < >>Representing the classification feature matrix,/->Representing matrix multiplication.
In one example, in the preparation system 100 of the above-described environmentally friendly domestic ceramic, the temperature control result generation module 180 includes: expanding the classification feature matrix into classification feature vectors according to row vectors or column vectors; performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.
Here, it will be understood by those skilled in the art that the specific functions and operations of the respective units and modules in the above-described production system 100 of the eco-friendly daily ceramic have been described in detail in the above description of the production process of the eco-friendly daily ceramic with reference to fig. 2 to 7, and thus, repetitive descriptions thereof will be omitted.

Claims (3)

1. The preparation process of the environment-friendly domestic ceramic is characterized by comprising the following steps of:
acquiring thermal infrared monitoring videos of the environment-friendly domestic ceramic in a preset time period acquired by a thermal infrared camera, and firing temperature values of a plurality of preset time points in the preset time period;
Extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic;
the thermal infrared monitoring keys are respectively processed through a first convolution neural network model using spatial attention so as to obtain a plurality of ceramic firing heat distribution characteristic matrixes;
arranging the ceramic firing heat distribution feature matrixes into three-dimensional input tensors according to time dimension, and then obtaining a ceramic firing heat distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel;
reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram to be a ceramic firing heat distribution time sequence characteristic vector;
the firing temperature values of the plurality of preset time points are arranged into input vectors according to time dimensions, and then a one-dimensional convolutional neural network model is used for obtaining temperature time sequence feature vectors;
calculating the response estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector to obtain a classification feature matrix; and
the classification feature matrix is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point should be increased or decreased;
wherein, the plurality of thermal infrared monitoring keys are respectively used for obtaining a plurality of ceramic firing heat distribution characteristic matrixes through a first convolution neural network model using spatial attention, and the method comprises the following steps:
Performing convolution processing, pooling processing along a channel dimension and nonlinear activation processing on the thermal infrared monitoring keys in forward transfer of layers by using each layer of the first convolutional neural network model to output a plurality of initial ceramic firing thermal distribution feature matrices from the last layer of the first convolutional neural network model; and
inputting the plurality of initial ceramic firing heat distribution feature matrices into a spatial attention layer of the first convolutional neural network model to obtain the plurality of ceramic firing heat distribution feature matrices;
the method for obtaining the ceramic firing heat distribution time sequence feature map by using the second convolution neural network model of the three-dimensional convolution kernel after arranging the ceramic firing heat distribution feature matrixes into three-dimensional input tensors according to the time dimension comprises the following steps: input data are respectively carried out in forward transfer by using each layer of the second convolution neural network model using the three-dimensional convolution kernel:
carrying out convolution processing on input data to obtain a convolution characteristic diagram;
pooling the convolution feature images based on the local feature matrix to obtain pooled feature images; and
non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map;
The output of the last layer of the second convolutional neural network using the three-dimensional convolutional kernel is the ceramic firing thermal distribution time sequence characteristic diagram, and the input of the first layer of the second convolutional neural network using the three-dimensional convolutional kernel is the three-dimensional input tensor;
the method for obtaining the temperature time sequence feature vector by the one-dimensional convolutional neural network model after arranging firing temperature values of a plurality of preset time points into input vectors according to a time dimension comprises the following steps: using each layer of the one-dimensional convolutional neural network model to respectively perform convolutional processing, mean pooling processing and nonlinear activation processing based on a one-dimensional convolutional kernel on an input vector in forward transfer of the layer so that the output of the last layer of the one-dimensional convolutional neural network model is the temperature time sequence feature vector;
wherein calculating a responsiveness estimate of the ceramic firing heat distribution timing feature vector relative to the temperature timing feature vector to obtain a classification feature matrix comprises:
calculating Helmholtz free energy factors of the ceramic firing heat distribution time sequence feature vector and the temperature time sequence feature vector to obtain a first Helmholtz free energy factor and a second Helmholtz free energy factor;
Weighting the ceramic firing heat distribution timing feature vector and the temperature timing feature vector based on the first helmholtz class free energy factor and the second helmholtz class free energy factor to obtain a weighted ceramic firing heat distribution timing feature vector and a weighted temperature timing feature vector; and
calculating a responsiveness estimate of the weighted ceramic firing heat distribution time sequence feature vector relative to the weighted temperature time sequence feature vector to obtain the classification feature matrix;
wherein calculating the helmholtz-like free energy factors of the ceramic firing heat distribution timing feature vector and the temperature timing feature vector to obtain a first helmholtz-like free energy factor and a second helmholtz-like free energy factor comprises: calculating the time sequence characteristic vector of the ceramic firing heat distribution according to the following formulaAnd the temperature timing feature vectorIs a helmholtz-like free energy factor;
wherein, the formula is:
，
wherein,classification probability value representing the ceramic firing heat distribution time sequence feature vector, < >>Classification probability value representing the temperature timing feature vector,/->Representing the ceramic firing heat distribution time sequence characteristic vector +. >Characteristic value of the location->Representing the temperature time sequence characteristic vector +.>Characteristic value of the location->A Helmholtz-like free energy factor representing a timing characteristic vector of the firing profile of the ceramic, +.>Helmholtz class free energy factor representing the temperature timing feature vector, ++>Represents a logarithmic function with base 2, +.>Represents an exponential operation based on a natural constant e, and +.>Is the length of the feature vector;
wherein calculating a responsiveness estimate of the weighted ceramic firing heat distribution timing feature vector relative to the weighted temperature timing feature vector to obtain the classification feature matrix comprises:
calculating a responsiveness estimate of the weighted ceramic firing profile timing feature vector relative to the weighted temperature timing feature vector to obtain a classification feature matrix;
wherein, the formula is:
，
wherein the method comprises the steps ofRepresenting the weighted ceramic firing heat distributionTiming feature vector,/->Representing the weighted temperature timing feature vector, < >>Representing the classification feature matrix,/->Representing matrix multiplication;
wherein, the classification feature matrix is passed through a classifier to obtain a classification result, the classification result is used for indicating that the firing temperature value of the current time point should be increased or decreased, and the method comprises the following steps:
Expanding the classification feature matrix into classification feature vectors according to row vectors or column vectors;
performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and
and inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.
2. A production system of an environment-friendly domestic ceramic using the production process of an environment-friendly domestic ceramic according to claim 1, characterized by comprising:
the monitoring module is used for acquiring thermal infrared monitoring videos of the environment-friendly daily ceramic in a preset time period and firing temperature values of a plurality of preset time points in the preset time period, wherein the thermal infrared monitoring videos are acquired by the thermal infrared camera;
the sampling module is used for extracting a plurality of thermal infrared monitoring keys from the thermal infrared monitoring video of the environment-friendly daily ceramic;
the spatial feature extraction module is used for obtaining a plurality of ceramic firing heat distribution feature matrixes by using a first convolution neural network model of spatial attention;
the thermal distribution time sequence feature extraction module is used for arranging the ceramic firing thermal distribution feature matrixes into three-dimensional input tensors according to a time dimension and then obtaining a ceramic firing thermal distribution time sequence feature diagram through a second convolution neural network model using a three-dimensional convolution kernel;
The dimension change module is used for reducing the dimension of the ceramic firing heat distribution time sequence characteristic diagram into a ceramic firing heat distribution time sequence characteristic vector;
the temperature time sequence feature extraction module is used for arranging firing temperature values of the plurality of preset time points into input vectors according to time dimensions and obtaining temperature time sequence feature vectors through a one-dimensional convolutional neural network model;
the responsiveness estimation module is used for calculating responsiveness estimation of the ceramic firing heat distribution time sequence feature vector relative to the temperature time sequence feature vector so as to obtain a classification feature matrix; and
and the temperature control result generation module is used for passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating that the firing temperature value of the current time point should be increased or decreased.
3. The system for preparing an environmentally friendly domestic ceramic as claimed in claim 2, wherein the spatial feature extraction module comprises:
the depth convolution coding unit is used for respectively carrying out convolution processing, pooling processing and nonlinear activation processing on the thermal infrared monitoring keys in forward transmission of layers by using each layer of the first convolution neural network model so as to output a plurality of initial ceramic firing thermal distribution feature matrixes by the last layer of the first convolution neural network model; and
And the spatial attention unit is used for inputting the initial ceramic firing heat distribution characteristic matrixes into the spatial attention layer of the first convolutional neural network model to obtain the ceramic firing heat distribution characteristic matrixes.
CN202310516727.4A 2023-05-09 2023-05-09 Preparation process and system of environment-friendly domestic ceramic Active CN116907214B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310516727.4A CN116907214B (en) 2023-05-09 2023-05-09 Preparation process and system of environment-friendly domestic ceramic

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310516727.4A CN116907214B (en) 2023-05-09 2023-05-09 Preparation process and system of environment-friendly domestic ceramic

Publications (2)

Publication Number Publication Date
CN116907214A CN116907214A (en) 2023-10-20
CN116907214B true CN116907214B (en) 2024-03-08

Family

ID=88365585

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310516727.4A Active CN116907214B (en) 2023-05-09 2023-05-09 Preparation process and system of environment-friendly domestic ceramic

Country Status (1)

Country Link
CN (1) CN116907214B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114815930A (en) * 2022-06-30 2022-07-29 烟台黄金职业学院 Temperature control system of calcinator and temperature control method thereof
CN115079572A (en) * 2022-06-30 2022-09-20 福建省龙德新能源有限公司 Energy management control system for preparing lithium hexafluorophosphate and control method thereof
CN115238566A (en) * 2022-06-09 2022-10-25 吉林大学 Artificial hole identification method based on frequency domain electromagnetic response
CN115291646A (en) * 2022-07-08 2022-11-04 福建龙氟化工有限公司 Energy management control system for lithium fluoride preparation and control method thereof
CN115936682A (en) * 2022-12-21 2023-04-07 江西有源工业废物回收处理有限公司 Waste recovery system and method for printed circuit board

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109815339B (en) * 2019-01-02 2022-02-08 平安科技（深圳）有限公司 Knowledge extraction method and device based on TextCNN, computer equipment and storage medium

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115238566A (en) * 2022-06-09 2022-10-25 吉林大学 Artificial hole identification method based on frequency domain electromagnetic response
CN114815930A (en) * 2022-06-30 2022-07-29 烟台黄金职业学院 Temperature control system of calcinator and temperature control method thereof
CN115079572A (en) * 2022-06-30 2022-09-20 福建省龙德新能源有限公司 Energy management control system for preparing lithium hexafluorophosphate and control method thereof
CN115291646A (en) * 2022-07-08 2022-11-04 福建龙氟化工有限公司 Energy management control system for lithium fluoride preparation and control method thereof
CN115936682A (en) * 2022-12-21 2023-04-07 江西有源工业废物回收处理有限公司 Waste recovery system and method for printed circuit board

Also Published As

Publication number Publication date
CN116907214A (en) 2023-10-20

Similar Documents

Publication Publication Date Title
CN108805083B (en) Single-stage video behavior detection method
CN110619355B (en) Automatic steel material microstructure identification method based on deep learning
CN111814875B (en) Ship sample expansion method in infrared image based on pattern generation countermeasure network
CN110728698B (en) Multi-target tracking system based on composite cyclic neural network system
CN109766813B (en) Dictionary learning face recognition method based on symmetric face expansion samples
CN109743642B (en) Video abstract generation method based on hierarchical recurrent neural network
Chen et al. Learning linear regression via single-convolutional layer for visual object tracking
CN115378344B (en) Electromagnetic motor and moisture-proof control method thereof
CN110472417B (en) Convolutional neural network-based malicious software operation code analysis method
CN103985120A (en) Remote sensing image multi-objective association method
CN110705591A (en) Heterogeneous transfer learning method based on optimal subspace learning
CN110298397A (en) The multi-tag classification method of heating metal image based on compression convolutional neural networks
CN114387473A (en) Small sample image classification method based on base class sample characteristic synthesis
Chen et al. Military image scene recognition based on CNN and semantic information
CN111340011A (en) Self-adaptive time sequence shift neural network time sequence behavior identification method and system
CN116907214B (en) Preparation process and system of environment-friendly domestic ceramic
CN116955675B (en) Hash image retrieval method and network based on fine-grained similarity relation contrast learning
CN117190657A (en) Tunnel type drying production line and method thereof
CN107184224B (en) Lung nodule diagnosis method based on bimodal extreme learning machine
CN111062406B (en) Heterogeneous domain adaptation-oriented semi-supervised optimal transmission method
CN112085096A (en) Method for detecting local abnormal heating of object based on transfer learning
CN113205484B (en) Mammary tissue classification and identification method based on transfer learning
CN112465836B (en) Thermal infrared semantic segmentation unsupervised field self-adaption method based on contour information
Thangaraj et al. Deep Learning based Real-Time Face Detection and Gender Classification using OpenCV and Inception v3
CN114067155A (en) Image classification method, device, product and storage medium based on meta learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant