CN115963229A - Gas monitoring system and monitoring method thereof - Google Patents
Gas monitoring system and monitoring method thereof Download PDFInfo
- Publication number
- CN115963229A CN115963229A CN202310028948.7A CN202310028948A CN115963229A CN 115963229 A CN115963229 A CN 115963229A CN 202310028948 A CN202310028948 A CN 202310028948A CN 115963229 A CN115963229 A CN 115963229A
- Authority
- CN
- China
- Prior art keywords
- response
- sensor
- topological
- matrix
- feature
- 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.)
- Withdrawn
Links
Images
Classifications
-
- 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
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A50/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE in human health protection, e.g. against extreme weather
- Y02A50/20—Air quality improvement or preservation, e.g. vehicle emission control or emission reduction by using catalytic converters
Landscapes
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
The application relates to the field of gas monitoring, and particularly discloses a gas monitoring system and a monitoring method thereof, wherein high-dimensional associated feature extraction is carried out on response values of each sensor at each time point through a context-based encoder model, implicit feature extraction is carried out on topological features of the sensors by adopting a convolutional neural network, the response feature matrix and the topological feature matrix are further input into a graph neural network together to extract associated information of a data sample due to feature information and irregular topological structure information, and the response topological feature vector is corrected based on temperature drift of the sensors so as to ensure the accuracy of decoding regression. By the method, concentration values of three gas concentrations in the closed laying hen house environment can be accurately measured, and further the health and production performance of laying hens are ensured.
Description
Technical Field
The present invention relates to the field of gas monitoring, and more particularly, to a gas monitoring system and a monitoring method thereof.
Background
A plurality of harmful gases exist in the closed environment of the laying hen house, and mainly comprise ammonia gas, hydrogen sulfide, carbon dioxide and the like. The existence of harmful gases can cause harm to the health and the production performance of the laying hens in different degrees. Hydrogen sulfide in harmful gas easily causes pulmonary edema of the chickens, ammonia gas easily causes conjunctivitis and conjunctival inflammation, and overhigh carbon dioxide concentration easily causes dyspnea of the chickens.
A part of barns are provided with breeding environment monitoring equipment, but harmful gas concentration data measured by sensors are often directly used for system regulation and control, and the data are not analyzed and processed. The cross-sensitivity problem of the gas sensor often causes the measurement result to have larger error. Therefore, a gas monitoring system is desired for accurately measuring concentration values of gas concentrations in an enclosed hen house environment.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides a gas monitoring system and a monitoring method thereof, wherein high-dimensional correlation characteristic extraction is carried out on the response value of each sensor at each time point through a context-based encoder model, a convolution neural network is adopted to carry out implicit characteristic extraction on the topological characteristic of the sensor, the response characteristic matrix and the topological characteristic matrix are further input into the graph neural network together to extract correlation information of a data sample due to characteristic information and irregular topological structure information, and the response topological characteristic vector is corrected based on the temperature drift of the sensor so as to ensure the accuracy of decoding regression. By the method, the concentration values of the three gas concentrations in the closed laying hen house environment can be accurately measured, and the health and the production performance of laying hens are further ensured.
According to one aspect of the present application, there is provided a gas monitoring system comprising:
the sensor data acquisition unit is used for acquiring the response value of each sensor in the array sensor at each time point;
a response value encoding unit, configured to pass a response value of each of the sensors at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and concatenate the plurality of feature vectors to obtain a response feature vector of each of the sensors;
the matrix splicing unit is used for carrying out two-dimensional splicing on the response characteristic vector of each sensor to obtain a response characteristic matrix;
a topology data unit, configured to obtain a topology matrix of the array sensors, where a value of each position at a non-diagonal position in the topology matrix is a distance between two corresponding sensors, and a characteristic value of each position at a diagonal position in the topology matrix is 0;
the topological coding unit is used for enabling the topological matrix to pass through a convolutional neural network so as to obtain a topological characteristic matrix;
a fusion unit, configured to pass the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, where the neural network generates the topological feature matrix including response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor;
the temperature drift value acquisition unit is used for acquiring the temperature drift value of each sensor in the array sensor at each time point;
the temperature drift coding unit is used for enabling the temperature drift value of each sensor at each time point to pass through the context-based coder model containing the embedded layer so as to obtain a temperature drift characteristic vector of each time point, and cascading the temperature drift characteristic vectors of each time point so as to obtain a global temperature drift characteristic vector of each sensor;
a correction unit, configured to correct the response topology feature vector of each sensor based on the global temperature drift feature vector of each sensor and using a periodic chaotic map to obtain a corrected response topology feature vector of each sensor, where the periodic chaotic map represents that the global temperature drift feature vector is incorporated into a periodic function;
the splicing unit is used for splicing the corrected response topological characteristic vector of each sensor into a regression matrix in a two-dimensional mode; and
and the decoding unit is used for decoding and regressing the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment.
In the above gas monitoring system, the response value encoding unit is further configured to: passing the response value of each of the sensors at various points in time through an embedding layer of the encoder model to convert the response value into a response embedding vector to obtain a sequence of response embedding vectors; inputting the sequence of response embedding vectors into a converter of the encoder model to obtain the plurality of feature vectors; and cascading the plurality of feature vectors to obtain a response feature vector for each of the sensors.
In the above gas monitoring system, each layer of the convolutional neural network performs convolution processing, pooling processing along channel dimension, and activation processing on input data during forward transfer of the layer to output the topological feature matrix from the last layer of the convolutional neural network, wherein the input of the first layer of the convolutional neural network is the topological matrix.
In the above gas monitoring system, the correction unit is further configured to: modifying the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using a periodic chaotic mapping according to the following formula to obtain a modified response topological characteristic vector of each sensor; wherein the formula is:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all locations of the global temperature drift eigenvector, and [ ·]For the rounding function, a is the hyperparameter.
In the above gas monitoring system, the decoding unit is further configured to: decoding and regressing the regression matrix by the following formula to generate concentration values of three gas concentrations in the closed hen house environment; wherein the formula is:
where X is the input matrix, Y is the output vector, W is the weight vector, and B is the offset vector.
In the above gas monitoring system, the three gases are ammonia, hydrogen sulfide and carbon dioxide.
According to another aspect of the present application, a monitoring method of a gas monitoring system includes:
acquiring a response value of each sensor in the array sensor at each time point;
passing the response value of each sensor at each time point through a context-based encoder model comprising an embedded layer to obtain a plurality of feature vectors, and concatenating the plurality of feature vectors to obtain a response feature vector for each sensor;
performing two-dimensional splicing on the response characteristic vector of each sensor to obtain a response characteristic matrix;
obtaining a topological matrix of the array sensors, wherein the value of each position at a non-diagonal position in the topological matrix is the distance between two corresponding sensors, and the characteristic value of each position at a diagonal position in the topological matrix is 0;
passing the topological matrix through a convolutional neural network to obtain a topological feature matrix;
passing the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, wherein the neural network generates the topological feature matrix comprising response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor;
acquiring the temperature drift value of each sensor in the array sensor at each time point;
passing the temperature drift value of each sensor at each time point through the context-based encoder model comprising the embedded layer to obtain a temperature drift feature vector of each time point, and cascading the temperature drift feature vectors of each time point to obtain a global temperature drift feature vector of each sensor;
modifying the response topological eigenvector of each sensor based on the global temperature drift eigenvector of each sensor and using a periodic chaotic map to obtain a modified response topological eigenvector of each sensor, wherein the periodic chaotic map represents the incorporation of the global temperature drift eigenvector into a periodic function;
two-dimensionally splicing the corrected response topological characteristic vectors of each sensor into a regression matrix; and
and decoding and regressing the regression matrix to generate concentration values of the three gas concentrations in the closed hen house environment.
In the monitoring method of the gas monitoring system, the step of passing the response value of each sensor at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and the step of concatenating the plurality of feature vectors to obtain the response feature vector of each sensor comprises: passing the response value of each of the sensors at various points in time through an embedding layer of the encoder model to convert the response value into a response embedding vector to obtain a sequence of response embedding vectors; inputting the sequence of response embedding vectors into a converter of the encoder model to obtain the plurality of feature vectors; and cascading the plurality of feature vectors to obtain a response feature vector for each of the sensors.
In the monitoring method of the gas monitoring system, each layer of the convolutional neural network performs convolution processing, pooling processing along channel dimensions, and activation processing on input data in the forward transfer process of the layer to output the topological characteristic matrix from the last layer of the convolutional neural network, wherein the input of the first layer of the convolutional neural network is the topological matrix.
In the monitoring method of the gas monitoring system, the modifying the response topology feature vector of each sensor based on the global temperature drift feature vector of each sensor and using periodic chaotic mapping to obtain a modified response topology feature vector of each sensor includes: modifying the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using a periodic chaotic mapping according to the following formula to obtain a modified response topological characteristic vector of each sensor; wherein the formula is:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all positions of the global temperature drift eigenvector, and [ ·]For the rounding function, a is the hyperparameter. />
In the monitoring method of the gas monitoring system, performing decoding regression on the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment includes: decoding and regressing the regression matrix by the following formula to generate concentration values of three gas concentrations in the closed hen house environment; wherein the formula is:
where X is the input matrix, Y is the output vector, W is the weight vector, and B is the offset vector.
In the monitoring method of the gas monitoring system, the three gases are ammonia gas, hydrogen sulfide and carbon dioxide.
Compared with the prior art, the gas monitoring system and the monitoring method thereof have the advantages that the context-based encoder model is used for extracting the high-dimensional correlation characteristics of the response value of each sensor at each time point, the convolutional neural network is adopted for extracting the implicit characteristics of the topological characteristics of the sensors, the response characteristic matrix and the topological characteristic matrix are further input into the graph neural network together to extract the correlation information of a data sample due to characteristic information and irregular topological structure information, and the response topological characteristic vector is corrected based on the temperature drift of the sensors so as to ensure the accuracy of decoding regression. By the method, concentration values of three gas concentrations in the closed laying hen house environment can be accurately measured, and further the health and production performance of laying hens are ensured.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally represent like parts or steps.
Fig. 1 is a diagram of an application scenario of a gas monitoring system according to an embodiment of the present application.
FIG. 2 is a block diagram of a gas monitoring system according to an embodiment of the present application.
Fig. 3 is a flow chart of a monitoring method of a gas monitoring system according to an embodiment of the present application.
Fig. 4 is a schematic diagram of a monitoring method of a gas monitoring system according to an embodiment of the present 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 a few embodiments of the present application, and not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.
Overview of a scene
As mentioned above, there are many harmful gases in the closed hen house environment, mainly including ammonia, hydrogen sulfide, carbon dioxide, etc. The existence of harmful gases can cause harm to the health and the production performance of the laying hens in different degrees. Hydrogen sulfide in harmful gas easily causes pulmonary edema of the chickens, ammonia gas easily causes conjunctivitis and conjunctival inflammation, and overhigh carbon dioxide concentration easily causes dyspnea of the chickens.
A part of barns are provided with breeding environment monitoring equipment, but harmful gas concentration data measured by sensors are often directly used for system regulation and control, and the data are not analyzed and processed. The cross-sensitivity problem of the gas sensor often causes the measurement result to have larger error. Therefore, a gas monitoring system is desired for accurately measuring concentration values of gas concentrations in an enclosed hen house environment.
Correspondingly, the hen house harmful gas monitoring system adopts an array formed by 5 sensors to detect 3 kinds of harmful gas in the hen house. The key in the solution of the present application is how to characterize the cross-sensitivity between the various other sensors.
Specifically, in the technical scheme of the application, response values of each sensor at each time point are obtained, a plurality of feature vectors are obtained through a context-based encoder model including an embedded layer, and the feature vectors are obtained through cascade connection. And performing two-dimensional splicing on the response characteristic vectors of the 5 sensors to obtain a response characteristic matrix.
And obtaining a topological matrix of 5 sensors, processing the topological matrix by using a convolutional neural network to generate a topological characteristic matrix, and inputting the topological characteristic matrix and the response characteristic matrix into the graph neural network to obtain a response topological characteristic matrix, wherein each row of the matrix is a response topological characteristic vector of one sensor.
And correcting the response topology characteristic vector based on the temperature drift of the sensor, specifically, obtaining the temperature drift value of the sensor at each time point, obtaining the temperature drift characteristic vector of each time point through a context-based encoder model containing an embedded layer, and then cascading to obtain the global temperature drift characteristic vector. Because the temperature drift characteristics have periodic chaos, the characteristic distribution of the response topological characteristics is improved by using periodic chaos mapping, so that the diversity of the characteristic value distribution is enhanced, and the global optimization capability during subsequent decoding is improved:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all positions of the global temperature drift eigenvector, and [ ·]For the rounding function, a is a hyperparameter, for example, with an initial value of 0.5.
Then, the corrected response topological characteristic vectors of the 5 sensors are two-dimensionally spliced into a regression matrix, and the regression is decoded by a decoder.
Based on this, the present application proposes a gas monitoring system comprising: the sensor data acquisition unit is used for acquiring the response value of each sensor in the array sensor at each time point; a response value encoding unit, configured to pass a response value of each of the sensors at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and concatenate the plurality of feature vectors to obtain a response feature vector of each of the sensors; the matrix splicing unit is used for carrying out two-dimensional splicing on the response characteristic vector of each sensor to obtain a response characteristic matrix; a topology data unit, configured to obtain a topology matrix of the array sensors, where a value of each position at a non-diagonal position in the topology matrix is a distance between two corresponding sensors, and a characteristic value of each position at a diagonal position in the topology matrix is 0; the topology coding unit is used for enabling the topology matrix to pass through a convolutional neural network so as to obtain a topology characteristic matrix; a fusion unit, configured to pass the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, where the neural network generates the topological feature matrix including response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor; the temperature drift value acquisition unit is used for acquiring the temperature drift value of each sensor in the array sensor at each time point; the temperature drift coding unit is used for enabling the temperature drift value of each sensor at each time point to pass through the context-based coder model containing the embedded layer so as to obtain a temperature drift characteristic vector of each time point, and cascading the temperature drift characteristic vectors of each time point so as to obtain a global temperature drift characteristic vector of each sensor; a correction unit, configured to correct the response topology feature vector of each sensor based on the global temperature drift feature vector of each sensor and using a periodic chaotic map to obtain a corrected response topology feature vector of each sensor, where the periodic chaotic map represents that the global temperature drift feature vector is incorporated into a periodic function; the splicing unit is used for splicing the corrected response topological characteristic vector of each sensor into a regression matrix in a two-dimensional mode; and the decoding unit is used for decoding and regressing the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment.
Fig. 1 illustrates an application scenario of a gas monitoring system according to an embodiment of the present application. As shown in fig. 1, in this application scenario, first, response values of each sensor at various time points are obtained through an array sensor (e.g., R as illustrated in fig. 1) disposed in a hen house (e.g., E as illustrated in fig. 1), and a topological matrix of the array sensor is obtained based on a distance between each two sensors, and a temperature drift value of each sensor in the array sensor at various time points is obtained through a temperature sensor (e.g., T as illustrated in fig. 1). Then, the obtained response value and temperature drift value of each sensor at each time point and the obtained topological matrix of the array sensor are input into a server (for example, a cloud server S as illustrated in fig. 1) deployed with a gas monitoring algorithm, wherein the server can process the response value and temperature drift value of each sensor at each time point and the topological matrix of the array sensor with the gas monitoring algorithm to generate concentration values of three gas concentrations in the closed hen house environment. And then, regulating and controlling the system environment of the closed laying hen house based on the concentration value of the gas so as to ensure the health and the production performance of the laying hens.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary System
FIG. 2 illustrates a block diagram of a gas monitoring system according to an embodiment of the present application. As shown in fig. 2, a gas monitoring system 200 according to an embodiment of the present application includes: a sensor data acquisition unit 210 for acquiring a response value of each sensor in the array sensor at each time point; a response value encoding unit 220, configured to pass the response value of each of the sensors at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and concatenate the plurality of feature vectors to obtain a response feature vector of each of the sensors; a matrix stitching unit 230, configured to perform two-dimensional stitching on the response feature vector of each sensor to obtain a response feature matrix; a topology data unit 240, configured to obtain a topology matrix of the array sensors, where a value of each position at a non-diagonal position in the topology matrix is a distance between two corresponding sensors, and a characteristic value of each position at a diagonal position in the topology matrix is 0; a topology coding unit 250, configured to pass the topology matrix through a convolutional neural network to obtain a topology feature matrix; a fusion unit 260, configured to pass the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, where the neural network generates the topological feature matrix including response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor; a temperature drift value obtaining unit 270, configured to obtain a temperature drift value of each sensor in the array sensor at each time point; a temperature drift encoding unit 280, configured to pass the temperature drift value of each sensor at each time point through the context-based encoder model including the embedded layer to obtain a temperature drift feature vector of each time point, and concatenate the temperature drift feature vectors of each time point to obtain a global temperature drift feature vector of each sensor; a correcting unit 290, configured to correct the response topology feature vector of each sensor based on the global temperature drift feature vector of each sensor and using a periodic chaotic map to obtain a corrected response topology feature vector of each sensor, where the periodic chaotic map represents that the global temperature drift feature vector is incorporated into a periodic function; the splicing unit 300 is configured to splice the corrected response topological characteristic vectors of each sensor into a regression matrix in a two-dimensional manner; and a decoding unit 310, configured to perform decoding regression on the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment.
Specifically, in the embodiment of the present application, the sensor data obtaining unit 210 and the response value encoding unit 220 are configured to obtain the response value of each sensor in the array sensor at each time point, pass the response value of each sensor at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and concatenate the plurality of feature vectors to obtain the response feature vector of each sensor. As mentioned previously, in the solution of the present application, it is desirable to detect 3 harmful gases in the chicken house using an array of 5 sensors, here ammonia, hydrogen sulphide and carbon dioxide, and to take into account how the cross-sensitivity between the various other sensors is characterized. Therefore, in the technical solution of the present application, first, the response value of each sensor at each time point is obtained by an array sensor deployed in the hen house. And then, encoding the obtained response value of each sensor at each time point in a context-based encoder model comprising an embedded layer to extract global response value related information so as to obtain a plurality of feature vectors. Then, the plurality of feature vectors are concatenated to obtain a response feature vector of each sensor for subsequent processing.
More specifically, in an embodiment of the present application, the response value encoding unit includes: first, the response value of each sensor at each time point is passed through an embedding layer of the encoder model to convert the response value into a response embedding vector to obtain a sequence of response embedding vectors. The sequence of response-embedding vectors is then input to a converter of the encoder model to obtain the plurality of feature vectors. It should be understood that since the converter-based encoder model can encode the response-embedded vector based on the context, the obtained plurality of feature vectors have global response-value-associated features. Finally, the plurality of feature vectors are concatenated to obtain a response feature vector for each of the sensors.
Specifically, in this embodiment of the application, the matrix stitching unit 230 is configured to perform two-dimensional stitching on the response feature vector of each sensor to obtain a response feature matrix. That is, in the technical scheme of this application, after obtaining the response eigenvector of each sensor, carry out two-dimensional concatenation with 5 the eigenvector of sensor to integrate 5 the characteristic information of sensor, and then obtain the response characteristic matrix.
Specifically, in this embodiment, the topology data unit 240 and the topology encoding unit 250 are configured to obtain a topology matrix of the array sensors, where a value of each position at a non-diagonal position in the topology matrix is a distance between two corresponding sensors, a characteristic value of each position at a diagonal position in the topology matrix is 0, and the topology matrix is passed through a convolutional neural network to obtain a topology characteristic matrix. It will be appreciated that cross-sensitivity is taken into account between gas sensors and that this cross-sensitivity information relates to topological information between 5 of said sensors, i.e. if the distance between said two sensors is relatively far apart, the degree of said cross-sensitivity is relatively weak. Therefore, in the technical solution of the present application, a topological matrix of the array sensor needs to be obtained according to a distance between every two sensors. It is worth mentioning that, here, the value of each position at the non-diagonal position in the topology matrix is the distance between the corresponding two sensors, and the characteristic value of each position at the diagonal position in the topology matrix is 0. And then, processing the topological matrix in a convolutional neural network to extract the high-dimensional topological characteristic of the array sensor, thereby obtaining the topological characteristic matrix. Accordingly, in one particular example, each layer of the convolutional neural network performs convolution processing, pooling processing along channel dimensions, and activation processing on input data during forward pass of the layer to output the topological feature matrix by a last layer of the convolutional neural network, wherein an input of a first layer of the convolutional neural network is the topological matrix.
Specifically, in this embodiment of the present application, the fusion unit 260 is configured to pass the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, where the neural network generates the topological feature matrix including response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor. That is, in the technical solution of the present application, in order to more effectively fuse the topology feature information and the response feature information, the topology feature matrix and the response feature matrix are further passed through a graph neural network to obtain a response topology feature matrix. Here, the graph neural network generates the topological feature matrix including response feature information and irregular topological structure information by learnable neural network parameters. It should be understood that the graph neural network can be used for processing graph data in an irregular non-euclidean space, so that associated information of a data sample, which exists due to response feature information and irregular topological structure information, can be extracted, and therefore, compared with a feature matrix obtained by directly splicing the obtained response topological feature matrix, the accuracy of decoding regression can be improved.
Specifically, in this embodiment of the present application, the temperature drift value obtaining unit 270 and the temperature drift encoding unit 280 are configured to obtain the temperature drift value of each sensor in the array sensor at each time point, pass the temperature drift value of each sensor at each time point through the context-based encoder model including the embedded layer to obtain the temperature drift feature vector of each time point, and concatenate the temperature drift feature vectors of each time point to obtain the global temperature drift feature vector of each sensor. It will be appreciated that cross-sensitivity exists between gas sensors and is also related to the temperature of each of the sensors themselves, i.e. the effect of temperature drift. Therefore, in the technical solution of the present application, the response topology feature vector needs to be corrected based on the temperature drift of the sensor. That is, specifically, in the technical solution of the present application, the temperature drift value of each sensor in the array sensor at each time point is first obtained by the temperature sensor. And then, encoding the temperature drift value of each sensor at each time point through the context-based encoder model comprising the embedded layer to extract global temperature drift associated information, so as to obtain a temperature drift characteristic vector of each time point. And then cascading the obtained temperature drift characteristic vectors of each time point to integrate the temperature drift characteristic information so as to obtain the global temperature drift characteristic vector of each sensor.
Specifically, in this embodiment of the present application, the modifying unit 290 is configured to modify the response topological feature vector of each sensor based on the global temperature drift feature vector of each sensor and by using a periodic chaotic map to obtain a modified response topological feature vector of each sensor, where the periodic chaotic map indicates that the global temperature drift feature vector is incorporated into a periodic function. It should be understood that, since the temperature drift characteristics have periodic chaos, the periodic chaos mapping is used to improve the characteristic distribution of the response topology characteristics, thereby enhancing the diversity of the characteristic value distribution and improving the global optimization capability in the subsequent decoding. Therefore, in the technical solution of the present application, the response topology feature vector of each sensor is further modified based on the global temperature drift feature vector of each sensor and by using a periodic chaotic map, so as to obtain a modified response topology feature vector of each sensor. Here, the periodic chaotic map represents the incorporation of a global temperature drift feature vector into a periodic function.
More specifically, in an embodiment of the present application, the modifying unit is further configured to: correcting the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using periodic chaotic mapping according to the following formula so as to obtain a corrected response topological characteristic vector of each sensor; wherein the formula is:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all locations of the global temperature drift eigenvector, and [ ·]For the rounding function, a is a hyperparameter, for example, with an initial value of 0.5.
Specifically, in this embodiment of the application, the splicing unit 300 and the decoding unit 310 are configured to splice the corrected response topological characteristic vectors of each sensor into a regression matrix in a two-dimensional manner, and perform decoding regression on the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment. That is, in the technical solution of the present application, after obtaining the corrected response topological feature vector of each sensor, the corrected response topological feature vectors of 5 sensors are two-dimensionally spliced into a regression matrix to integrate the topological response features of all the sensors, so as to facilitate subsequent decoding regression. Then, the regression matrix is decoded and regressed to generate concentration values of the three gas concentrations in the closed hen house environment. Here, the three gases are ammonia gas, hydrogen sulfide, and carbon dioxide.
Accordingly, in a specific example, the decoding unit is further configured to: decoding and regressing the regression matrix by the following formula to generate concentration values of three gas concentrations in the environment of the closed laying hen house; wherein the formula is:
where X is the input matrix, Y is the output vector, W is the weight vector, and B is the offset vector.
In summary, the gas monitoring system 200 according to the embodiment of the present application is illustrated, which performs high-dimensional correlation feature extraction on the response value of each sensor at each time point through a context-based encoder model, performs implicit feature extraction on the topological feature of the sensor by using a convolutional neural network, further inputs the response feature matrix and the topological feature matrix into the graph neural network together to extract correlation information of a data sample due to feature information and irregular topological structure information, and corrects the response topological feature vector based on the temperature drift of the sensor, so as to ensure the accuracy of decoding regression. By the method, concentration values of three gas concentrations in the closed laying hen house environment can be accurately measured, and further the health and production performance of laying hens are ensured.
As described above, the gas monitoring system 200 according to the embodiment of the present application may be implemented in various terminal devices, such as a server of a gas monitoring algorithm, and the like. In one example, gas monitoring system 200 according to embodiments of the present application may be integrated into a terminal device as a software module and/or a hardware module. For example, the gas monitoring system 200 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the gas monitoring system 200 may also be one of many hardware modules of the terminal device.
Alternatively, in another example, the gas monitoring system 200 and the terminal device may also be separate devices, and the gas monitoring system 200 may be connected to the terminal device through a wired and/or wireless network and transmit the interaction information in an agreed data format.
Exemplary method
Fig. 3 illustrates a flow chart of a monitoring method of the gas monitoring system. As shown in fig. 3, a monitoring method of a gas monitoring system according to an embodiment of the present application includes the steps of: s110, acquiring a response value of each sensor in the array sensor at each time point; s120, enabling the response value of each sensor at each time point to pass through a context-based encoder model comprising an embedded layer to obtain a plurality of feature vectors, and cascading the plurality of feature vectors to obtain a response feature vector of each sensor; s130, performing two-dimensional splicing on the response characteristic vector of each sensor to obtain a response characteristic matrix; s140, obtaining a topological matrix of the array sensors, wherein the value of each position at a non-diagonal position in the topological matrix is the distance between two corresponding sensors, and the characteristic value of each position at a diagonal position in the topological matrix is 0; s150, passing the topological matrix through a convolutional neural network to obtain a topological characteristic matrix; s160, passing the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, wherein the neural network generates the topological feature matrix containing response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor; s170, acquiring temperature drift values of each sensor in the array sensor at each time point; s180, enabling the temperature drift value of each sensor at each time point to pass through the context-based encoder model comprising the embedded layer to obtain a temperature drift characteristic vector of each time point, and cascading the temperature drift characteristic vectors of each time point to obtain a global temperature drift characteristic vector of each sensor; s190, modifying the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using a periodic chaotic map to obtain a modified response topological characteristic vector of each sensor, wherein the periodic chaotic map represents that the global temperature drift characteristic vector is combined into a periodic function; s200, two-dimensionally splicing the corrected response topological characteristic vectors of each sensor into a regression matrix; and S210, performing decoding regression on the regression matrix to generate concentration values of the three gas concentrations in the closed hen house environment.
Fig. 4 illustrates an architecture diagram of a monitoring method of a gas monitoring system according to an embodiment of the present application. As shown in fig. 4, in the network architecture of the monitoring method of the gas monitoring system, first, the obtained response value (e.g., P1 as illustrated in fig. 4) of each of the sensors at each time point is passed through a context-based encoder model (e.g., E as illustrated in fig. 4) including an embedded layer to obtain a plurality of eigenvectors (e.g., VF1 as illustrated in fig. 4), and the plurality of eigenvectors are concatenated to obtain a response eigenvector (e.g., VF2 as illustrated in fig. 4) of each of the sensors; then, two-dimensionally concatenating the response feature vectors of each of the sensors to obtain a response feature matrix (e.g., MF1 as illustrated in fig. 4); then, passing the obtained topological matrix (e.g., M as illustrated in fig. 4) through a convolutional neural network (e.g., CNN as illustrated in fig. 4) to obtain a topological feature matrix (e.g., MF2 as illustrated in fig. 4); then, passing the topological feature matrix and the response feature matrix through a graph neural network (e.g., GNN as illustrated in fig. 4) to obtain a response topological feature matrix (e.g., MF as illustrated in fig. 4); then, passing the obtained temperature drift value (e.g., P2 as illustrated in fig. 4) of each of the sensors at each time point through the context-based encoder model (e.g., E as illustrated in fig. 4) including the embedded layer to obtain a temperature drift feature vector (e.g., VF3 as illustrated in fig. 4) of each time point, and cascading the temperature drift feature vectors of each time point to obtain a global temperature drift feature vector (e.g., VF4 as illustrated in fig. 4) of each of the sensors; then, modifying the response topological feature vector of each sensor based on the global temperature drift feature vector of each sensor and using a periodic chaotic map to obtain a modified response topological feature vector (e.g., VF as illustrated in fig. 4) of each sensor; then, two-dimensionally stitching the corrected response topological feature vectors of each of the sensors into a regression matrix (e.g., MR as illustrated in fig. 4); and finally, decoding and regressing the regression matrix to generate concentration values of the three gas concentrations in the closed hen house environment.
More specifically, in steps S110 and S120, a response value of each sensor in the array sensor at each time point is obtained, and the response value of each sensor at each time point is passed through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and the plurality of feature vectors are concatenated to obtain a response feature vector of each sensor. It will be appreciated that in the solution of the present application it is desirable to use an array of 5 sensors for the detection of 3 harmful gases in the chicken house, here ammonia, hydrogen sulphide and carbon dioxide, and to take into account how the cross-sensitivity between the various other sensors is characterized. Therefore, in the technical solution of the present application, first, the response value of each sensor at each time point is obtained through the array sensor deployed in the hen house. Then, the obtained response value of each sensor at each time point is subjected to encoding processing in a context-based encoder model comprising an embedded layer, so as to extract global response value related information, and thus a plurality of feature vectors are obtained. Then, the plurality of feature vectors are concatenated to obtain a response feature vector of each sensor for subsequent processing.
Specifically, in the embodiment of the present application, the process of passing the response value of each sensor at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and concatenating the plurality of feature vectors to obtain the response feature vector of each sensor includes: first, the response value of each sensor at each time point is passed through an embedding layer of the encoder model to convert the response value into a response embedding vector to obtain a sequence of response embedding vectors. The sequence of response-embedding vectors is then input to a converter of the encoder model to obtain the plurality of feature vectors. It should be understood that since the converter-based encoder model can encode the response-embedded vector based on the context, the obtained plurality of feature vectors have global response-value-associated features. Finally, the plurality of feature vectors are concatenated to obtain a response feature vector for each of the sensors.
More specifically, in step S130, the response feature vectors of each of the sensors are two-dimensionally concatenated to obtain a response feature matrix. That is, in the technical scheme of this application, after obtaining the response eigenvector of each sensor, carry out two-dimensional concatenation with 5 the eigenvector of sensor to integrate 5 the characteristic information of sensor, and then obtain the response characteristic matrix.
More specifically, in steps S140 and S150, a topological matrix of the array sensors is obtained, wherein a value of each position at an off-diagonal position in the topological matrix is a distance between the corresponding two sensors, a characteristic value of each position at a diagonal position in the topological matrix is 0, and the topological matrix is passed through a convolutional neural network to obtain a topological characteristic matrix. It will be appreciated that cross-sensitivity is taken into account between gas sensors and that this cross-sensitivity information relates to topological information between 5 of said sensors, i.e. if the distance between said two sensors is relatively far apart, the degree of said cross-sensitivity is relatively weak. Therefore, in the technical solution of the present application, a topological matrix of the array sensor needs to be obtained according to a distance between every two sensors. It is worth mentioning that, here, the value of each position at the non-diagonal position in the topology matrix is the distance between the corresponding two sensors, and the characteristic value of each position at the diagonal position in the topology matrix is 0. And then, processing the topological matrix in a convolutional neural network to extract the high-dimensional topological characteristics of the array sensor, thereby obtaining the topological characteristic matrix. Accordingly, in one specific example, each layer of the convolutional neural network performs convolution processing, pooling processing along channel dimension, and activation processing on input data during forward pass of the layer to output the topological feature matrix from the last layer of the convolutional neural network, wherein the input of the first layer of the convolutional neural network is the topological matrix.
More specifically, in step S160, the topological feature matrix and the response feature matrix are passed through a neural network to obtain a response topological feature matrix, wherein the neural network generates the topological feature matrix including response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor. That is, in the technical solution of the present application, in order to more effectively fuse the topology feature information and the response feature information, the topology feature matrix and the response feature matrix are further passed through a graph neural network to obtain a response topology feature matrix. Here, the graph neural network generates the topological feature matrix including response feature information and irregular topological structure information by using learnable neural network parameters. It should be understood that the graph neural network can be used for processing graph data in an irregular non-euclidean space, so that associated information of a data sample, which exists due to response feature information and irregular topological structure information, can be extracted, and therefore, compared with a feature matrix obtained by directly splicing the obtained response topological feature matrix, the accuracy of decoding regression can be improved.
More specifically, in steps S170 and S180, the temperature drift value of each sensor in the array sensor at each time point is obtained, the temperature drift value of each sensor at each time point is passed through the context-based encoder model including the embedded layer to obtain a temperature drift feature vector of each time point, and the temperature drift feature vectors of each time point are concatenated to obtain a global temperature drift feature vector of each sensor. It will be appreciated that cross-sensitivity exists between gas sensors and is also related to the temperature of each of the sensors themselves, i.e. the effect of temperature drift. Therefore, in the technical solution of the present application, the response topology feature vector needs to be corrected based on the temperature drift of the sensor. That is, specifically, in the technical solution of the present application, the temperature drift value of each sensor in the array sensor at each time point is first obtained through the temperature sensor. And then, encoding the temperature drift value of each sensor at each time point through the context-based encoder model comprising the embedded layer to extract global temperature drift associated information, so as to obtain a temperature drift characteristic vector of each time point. And then cascading the obtained temperature drift characteristic vectors of each time point to integrate the temperature drift characteristic information so as to obtain the global temperature drift characteristic vector of each sensor.
More specifically, in step S190, the response topological feature vector of each sensor is modified based on the global temperature drift feature vector of each sensor and by using a periodic chaotic map to obtain a modified response topological feature vector of each sensor, wherein the periodic chaotic map represents that the global temperature drift feature vector is incorporated into a periodic function. It should be understood that, since the temperature drift characteristics have periodic chaos, the periodic chaos mapping is used to improve the characteristic distribution of the response topology characteristics, thereby enhancing the diversity of the characteristic value distribution and improving the global optimization capability in the subsequent decoding. Therefore, in the technical solution of the present application, the response topology feature vector of each sensor is further modified based on the global temperature drift feature vector of each sensor and by using a periodic chaotic map, so as to obtain a modified response topology feature vector of each sensor. Here, the periodic chaotic map represents the incorporation of a global temperature drift feature vector into a periodic function.
Specifically, in this embodiment of the present application, the process of modifying the response topology feature vector of each sensor based on the global temperature drift feature vector of each sensor and using periodic chaotic mapping to obtain a modified response topology feature vector of each sensor includes: modifying the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using a periodic chaotic mapping according to the following formula to obtain a modified response topological characteristic vector of each sensor; wherein the formula is:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all positions of the global temperature drift eigenvector, and [ ·]For the rounding function, a is a hyperparameter, for example, with an initial value of 0.5.
More specifically, in step S200 and step S210, the corrected response topological feature vectors of each sensor are two-dimensionally spliced into a regression matrix, and the regression matrix is decoded and regressed to generate concentration values of three gas concentrations in the closed hen house environment. That is, in the technical solution of the present application, after obtaining the corrected response topological feature vector of each sensor, the corrected response topological feature vectors of 5 sensors are two-dimensionally spliced into a regression matrix to integrate the topological response features of all the sensors, so as to facilitate subsequent decoding regression. Then, decoding and regressing the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment. Here, the three gases are ammonia, hydrogen sulfide, and carbon dioxide.
In summary, the monitoring method of the gas monitoring system based on the embodiment of the present application is illustrated, which performs high-dimensional associated feature extraction on the response value of each sensor at each time point through a context-based encoder model, performs implicit feature extraction on the topological features of the sensors by using a convolutional neural network, further inputs the response feature matrix and the topological feature matrix into a graph neural network together to extract associated information of a data sample due to feature information and irregular topological structure information, and corrects the response topological feature vector based on the temperature drift of the sensors so as to ensure the accuracy of decoding regression. By the method, concentration values of three gas concentrations in the closed laying hen house environment can be accurately measured, and further the health and production performance of laying hens are ensured.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. As used herein, the words "or" and "refer to, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations should be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (10)
1. A gas monitoring system, comprising:
the sensor data acquisition unit is used for acquiring the response value of each sensor in the array sensor at each time point;
a response value encoding unit, configured to pass a response value of each of the sensors at each time point through a context-based encoder model including an embedded layer to obtain a plurality of feature vectors, and concatenate the plurality of feature vectors to obtain a response feature vector of each of the sensors;
the matrix splicing unit is used for carrying out two-dimensional splicing on the response characteristic vector of each sensor to obtain a response characteristic matrix;
a topology data unit, configured to obtain a topology matrix of the array sensors, where a value of each position at a non-diagonal position in the topology matrix is a distance between two corresponding sensors, and a characteristic value of each position at a diagonal position in the topology matrix is 0;
the topological coding unit is used for enabling the topological matrix to pass through a convolutional neural network so as to obtain a topological characteristic matrix;
a fusion unit, configured to pass the topological feature matrix and the response feature matrix through a neural network to obtain a response topological feature matrix, where the neural network generates the topological feature matrix including response feature information and irregular topological structure information according to learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor;
the temperature drift value acquisition unit is used for acquiring the temperature drift value of each sensor in the array sensor at each time point;
the temperature drift coding unit is used for enabling the temperature drift value of each sensor at each time point to pass through the context-based coder model containing the embedded layer so as to obtain a temperature drift characteristic vector of each time point, and cascading the temperature drift characteristic vectors of each time point so as to obtain a global temperature drift characteristic vector of each sensor;
the correction unit is used for correcting the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using periodic chaotic mapping to obtain the corrected response topological characteristic vector of each sensor, wherein the periodic chaotic mapping represents that the global temperature drift characteristic vector is combined into a periodic function;
the splicing unit is used for splicing the corrected response topological characteristic vectors of each sensor into a regression matrix in a two-dimensional mode; and
and the decoding unit is used for decoding and regressing the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment.
2. The gas monitoring system of claim 1, wherein the response value encoding unit is further configured to:
passing the response value of each of the sensors at various points in time through an embedding layer of the encoder model to convert the response value into a response embedding vector to obtain a sequence of response embedding vectors;
inputting the sequence of response embedding vectors into a converter of the encoder model to obtain the plurality of feature vectors; and
the plurality of feature vectors are concatenated to obtain a response feature vector for each of the sensors.
3. The gas monitoring system of claim 2, wherein each layer of the convolutional neural network convolves input data during forward pass of the layer, pooled along a channel dimension, and activated to output the topological feature matrix from a last layer of the convolutional neural network, wherein an input of the first layer of the convolutional neural network is the topological matrix.
4. The gas monitoring system of claim 3, wherein the correction unit is further configured to correct the response topology feature vector of each sensor based on the global temperature drift feature vector of each sensor and using a periodic chaotic map to obtain a corrected response topology feature vector of each sensor;
wherein the formula is:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all positions of the global temperature drift eigenvector, and [ ·]For the rounding function, a is the hyperparameter.
5. The gas monitoring system of claim 4, wherein the decoding unit is further configured to perform decoding regression on the regression matrix to generate concentration values of three gas concentrations in the closed hen house environment according to the following formula;
wherein the formula is:
where X is the input matrix, Y is the output vector, W is the weight vector, and B is the offset vector.
6. The gas monitoring system of claim 5, wherein the three gases are ammonia, hydrogen sulfide, and carbon dioxide.
7. A method of monitoring a gas monitoring system, comprising:
acquiring a response value of each sensor in the array sensor at each time point;
passing the response value of each sensor at each time point through a context-based encoder model comprising an embedded layer to obtain a plurality of feature vectors, and concatenating the plurality of feature vectors to obtain a response feature vector for each sensor;
performing two-dimensional splicing on the response characteristic vector of each sensor to obtain a response characteristic matrix;
obtaining a topological matrix of the array sensors, wherein the value of each position at a non-diagonal position in the topological matrix is the distance between two corresponding sensors, and the characteristic value of each position at a diagonal position in the topological matrix is 0;
passing the topological matrix through a convolutional neural network to obtain a topological feature matrix;
passing the topological feature matrix and the response feature matrix through a graph neural network to obtain a response topological feature matrix, wherein the graph neural network generates the topological feature matrix comprising response feature information and irregular topological structure information through learnable neural network parameters, and each row vector in the topological feature matrix corresponds to a response topological feature vector of one sensor;
acquiring the temperature drift value of each sensor in the array sensor at each time point;
passing the temperature drift value of each sensor at each time point through the context-based encoder model comprising the embedded layer to obtain a temperature drift feature vector of each time point, and cascading the temperature drift feature vectors of each time point to obtain a global temperature drift feature vector of each sensor;
modifying the response topological eigenvector of each sensor based on the global temperature drift eigenvector of each sensor and using a periodic chaotic map to obtain a modified response topological eigenvector of each sensor, wherein the periodic chaotic map represents incorporating the global temperature drift eigenvector into a periodic function;
two-dimensionally splicing the corrected response topological characteristic vectors of each sensor into a regression matrix; and
and decoding and regressing the regression matrix to generate concentration values of the three gas concentrations in the closed hen house environment.
8. The method of monitoring of a gas monitoring system of claim 7, wherein passing the response value of each of the sensors at the respective point in time through a context-based encoder model comprising an embedded layer to obtain a plurality of feature vectors, and concatenating the plurality of feature vectors to obtain the response feature vector for each of the sensors comprises:
passing the response value of each sensor at each time point through an embedding layer of the encoder model to convert the response value into a response embedding vector to obtain a sequence of response embedding vectors;
inputting the sequence of response embedding vectors into a converter of the encoder model to obtain the plurality of feature vectors; and
the plurality of feature vectors are concatenated to obtain a response feature vector for each of the sensors.
9. The method of monitoring a gas monitoring system of claim 7, wherein each layer of the convolutional neural network convolves input data during forward pass of the layer, pooled along a channel dimension, and activated to output the topological feature matrix by a last layer of the convolutional neural network, wherein an input of the first layer of the convolutional neural network is the topological matrix.
10. The monitoring method of a gas monitoring system according to claim 7, wherein modifying the response topology feature vector of each of the sensors based on the global temperature drift feature vector of each of the sensors and using periodic chaotic mapping to obtain a modified response topology feature vector of each of the sensors comprises:
modifying the response topological characteristic vector of each sensor based on the global temperature drift characteristic vector of each sensor and by using a periodic chaotic mapping according to the following formula to obtain a modified response topological characteristic vector of each sensor;
wherein the formula is:
wherein f is 1i And f 2i Characteristic values f of the ith position of the response topology characteristic vector before and after correction 3i Is the eigenvalue of the ith position of the global temperature drift eigenvector,
represents the mean of the eigenvalues of all positions of the global temperature drift eigenvector, and [ ·]For the rounding function, a is the hyperparameter. />
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310028948.7A CN115963229A (en) | 2023-01-09 | 2023-01-09 | Gas monitoring system and monitoring method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310028948.7A CN115963229A (en) | 2023-01-09 | 2023-01-09 | Gas monitoring system and monitoring method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115963229A true CN115963229A (en) | 2023-04-14 |
Family
ID=87357997
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310028948.7A Withdrawn CN115963229A (en) | 2023-01-09 | 2023-01-09 | Gas monitoring system and monitoring method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115963229A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117054968A (en) * | 2023-08-19 | 2023-11-14 | 杭州优航信息技术有限公司 | Sound source positioning system and method based on linear array microphone |
| CN117091799A (en) * | 2023-10-17 | 2023-11-21 | 湖南一特医疗股份有限公司 | Intelligent three-dimensional monitoring method and system for oxygen supply safety of medical center |
-
2023
- 2023-01-09 CN CN202310028948.7A patent/CN115963229A/en not_active Withdrawn
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117054968A (en) * | 2023-08-19 | 2023-11-14 | 杭州优航信息技术有限公司 | Sound source positioning system and method based on linear array microphone |
| CN117054968B (en) * | 2023-08-19 | 2024-03-12 | 杭州优航信息技术有限公司 | Sound source positioning system and method based on linear array microphone |
| CN117091799A (en) * | 2023-10-17 | 2023-11-21 | 湖南一特医疗股份有限公司 | Intelligent three-dimensional monitoring method and system for oxygen supply safety of medical center |
| CN117091799B (en) * | 2023-10-17 | 2024-01-02 | 湖南一特医疗股份有限公司 | Intelligent three-dimensional monitoring method and system for oxygen supply safety of medical center |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115963229A (en) | Gas monitoring system and monitoring method thereof | |
| Su et al. | Inner envelopes: efficient estimation in multivariate linear regression | |
| CN114647198B (en) | Intelligent home control method and system based on Internet of things and electronic equipment | |
| JP7237457B2 (en) | computing equipment | |
| US11906533B2 (en) | Chemical sensing system | |
| CN115099684B (en) | Enterprise safety production management system and management method thereof | |
| CN115147655A (en) | Oil gas gathering and transportation monitoring system and method thereof | |
| CN115291646A (en) | Energy management control system for lithium fluoride preparation and control method thereof | |
| CN110166055B (en) | Multi-channel compressive sensing optimization method and system based on compressive sensing | |
| CN115144312A (en) | Indoor air fine particle measuring method and system based on Internet of things | |
| CN114548246A (en) | Control method and system of intelligent device and electronic device | |
| CN114580520A (en) | Monitoring system and monitoring method for secondary equipment of power system | |
| CN115186774A (en) | Intelligent cable stripping equipment and method thereof | |
| CN115834433A (en) | Data processing method and system based on Internet of things technology | |
| JP2020526072A5 (en) | ||
| CN116859830A (en) | Production management control system for electronic grade ammonium fluoride production | |
| CN113190655B (en) | Spatial relation extraction method and device based on semantic dependency | |
| CN114745168A (en) | Cloud platform inlet real-time flow copying method and system and electronic equipment | |
| CN113656066B (en) | Clone code detection method based on feature alignment | |
| CN116340703A (en) | Soil sample parameter measurement method and system based on data | |
| CN116785967A (en) | Automatic batching system for electronic grade lithium hexafluorophosphate preparation | |
| CN116320459A (en) | Computer network communication data processing method and system based on artificial intelligence | |
| CN113963757A (en) | Oil-filled electrical equipment fault diagnosis method based on gas relation and graph neural network | |
| Guo et al. | Robust electronic nose in industrial cyber physical systems based on domain adaptive subspace transfer model | |
| Shibuya et al. | Reliable index for measuring information flow |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20230414 |
|
| WW01 | Invention patent application withdrawn after publication |