Smart city carbon monitoring method and system based on data feedback

Abstract

The invention relates to the technical field of carbon emission monitoring, in particular to a smart city carbon monitoring method and system based on data feedback, and the method comprises the following steps: building industrial characteristic correlation among carbon monitoring grids, and mapping the industrial characteristic correlation to a grid chart to obtain an industrial correlation snapshot; establishing carbon emission characteristic correlation among the carbon monitoring grids, and mapping the carbon emission characteristic correlation to the grid chart to obtain a carbon emission correlation snapshot; and establishing a carbon emission stability monitoring model for performing similarity measurement on the industrial correlation snapshots and the carbon emission correlation snapshots by using a twin neural network. According to the method, the carbon emission stability monitoring model is established by using the twin neural network, the carbon emission stability state of the target city is determined according to the feedback of industrial layout data such as the position coordinate data and the industrial economic data, underlying data support is provided for urban carbon monitoring, and the urban carbon emission monitoring accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission monitoring, and in particular to a smart city carbon monitoring method and system based on data feedback. Background Art

[0002] Carbon emissions refer to the general term for greenhouse gas emissions, among which carbon dioxide is the most important gas. Carbon emissions are mainly generated during product production, transportation, use and recycling. Therefore, carbon emissions in cities will change with changes in industrial layout. When monitoring urban carbon emissions, it is necessary to determine whether the city's carbon emissions are stable based on the feedback of urban industrial layout data. However, existing technologies usually ignore data feedback from industrial layout when monitoring carbon emissions, which leads to the city's carbon monitoring being separated from the underlying data support, resulting in the defect of insufficient monitoring accuracy. Summary of the invention

[0003] The purpose of the present invention is to provide a smart city carbon monitoring method and system based on data feedback, so as to solve the technical problem that the data feedback of industrial layout is ignored in the existing technology during carbon emission monitoring, resulting in the city carbon monitoring being separated from the underlying data support and causing insufficient monitoring accuracy.

[0004] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0005] A smart city carbon monitoring method based on data feedback includes the following steps:

[0006] Divide the target city into grids to form a grid map containing multiple carbon monitoring grids;

[0007] Based on the location coordinate data and industrial economic data of each carbon monitoring grid, establish the industrial characteristic correlation between each carbon monitoring grid, and map the industrial characteristic correlation to the grid map to obtain an industrial correlation snapshot;

[0008] Based on the carbon emission data of each carbon monitoring grid, establish the carbon emission characteristic correlation between each carbon monitoring grid, and map the carbon emission characteristic correlation to the grid map to obtain a carbon emission correlation snapshot;

[0009] A carbon emission stability monitoring model is established using twin neural networks to measure the similarity between industry correlation snapshots and carbon emission correlation snapshots;

[0010] The carbon emission stability state of the target city is determined based on the output of the carbon emission stability monitoring model.

[0011] As a preferred solution of the present invention, the grid graph is: G = [a1, a2, ..., a n ], where G is the identifier of the grid graph, a1, a2 and an They are the 1st, 2nd and nth carbon monitoring grids respectively.

[0012] As a preferred solution of the present invention, the method for establishing the correlation of industry characteristics includes:

[0013] The positional correlation of each carbon monitoring grid is quantified by the position coordinate data of each carbon monitoring grid, and the positional correlation is:

[0014]

[0015] Where P ij is the position correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, P i is the location coordinate data of the i-th carbon monitoring grid, P j is the location coordinate data of the jth carbon monitoring grid;

[0016] The industrial relevance of each carbon monitoring grid is quantified through the industrial economic data of each carbon monitoring grid. The industrial relevance is:

[0017]

[0018] In the formula, G ij is the industry correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, G i is the industrial economic data of the i-th carbon monitoring grid, G j is the industrial economic data of the jth carbon monitoring grid;

[0019] Multiplying the location correlation and the industry correlation to obtain the industry characteristic correlation;

[0020] The quantitative formula for the correlation of the industrial characteristics is:

[0021] PG ij =P ij *G ij ;

[0022] In the formula, PG ij is the correlation between the industrial characteristics of the i-th carbon monitoring grid and the j-th carbon monitoring grid, P ij is the position correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, G ij is the industry correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid.

[0023] As a preferred solution of the present invention, the industry correlation snapshot is: Where S is the identifier of the industry relevance snapshot, PG 11 is the correlation between the first carbon monitoring grid and the industry characteristics of the first carbon monitoring grid, PG1n is the correlation between the industry characteristics of the first carbon monitoring grid and the nth carbon monitoring grid, PG n1 is the correlation between the industrial characteristics of the nth carbon monitoring grid and the first carbon monitoring grid, PG nn is the correlation between the industry characteristics of the nth carbon monitoring grid and the nth carbon monitoring grid.

[0024] As a preferred embodiment of the present invention, the method for establishing the carbon emission characteristic correlation includes:

[0025] The carbon emission characteristic correlation of each carbon monitoring grid is quantified through the carbon emission data of each carbon monitoring grid. The carbon emission characteristic correlation is:

[0026]

[0027] In the formula, C ij is the correlation of carbon emission characteristics between the i-th carbon monitoring grid and the j-th carbon monitoring grid, C i is the carbon emission data of the i-th carbon monitoring grid, C j is the carbon emission data of the jth carbon monitoring grid.

[0028] As a preferred solution of the present invention, the carbon emission correlation snapshot is: Where H is the identifier of the carbon emission correlation snapshot, C 11 is the correlation between the carbon emission characteristics of the first carbon monitoring grid and the first carbon monitoring grid, C 1n is the correlation of carbon emission characteristics between the first carbon monitoring grid and the nth carbon monitoring grid, C n1 is the carbon emission characteristic correlation between the nth carbon monitoring grid and the first carbon monitoring grid, C nn is the correlation between the carbon emission characteristics of the n-th carbon monitoring grid and the n-th carbon monitoring grid.

[0029] As a preferred embodiment of the present invention, the method for constructing the carbon emission stability monitoring model includes:

[0030] Select a time period when carbon emissions are in a stable state as a positive sample time period, and obtain a carbon emission correlation snapshot and an industry correlation snapshot at each time in the positive sample time period;

[0031] Marking the similarity measurement values ​​of the carbon emission correlation snapshot and the industry correlation snapshot at each time point in the positive sample time period as 1;

[0032] Select a time period when carbon emissions are in an unstable state as a negative sample time period, and obtain a carbon emission correlation snapshot and an industry correlation snapshot at each time in the negative sample time period;

[0033] Marking the similarity measurement values ​​of the carbon emission correlation snapshot and the industry correlation snapshot at each time point in the negative sample time period as 0;

[0034] The carbon emission correlation snapshots and industry correlation snapshots and similarity measurement values ​​in the positive sample time period and the negative sample time period are mixed into a data set, and the data set is divided into a training set and a test set;

[0035] Using the training set, taking the carbon emission correlation snapshot and the industry correlation snapshot as two input quantities of the twin network, taking the similarity measurement value as the output quantity of the twin network, training the twin network, and generating a carbon emission stability monitoring model;

[0036] The performance of the carbon emission stability monitoring model was evaluated using the test set and the mean square error as the evaluation indicator.

[0037] As a preferred solution of the present invention, the carbon emission stability monitoring model is:

[0038]

[0039] Where d is the similarity measure between the carbon emission correlation snapshot and the industry correlation snapshot, H is the identifier of the carbon emission correlation snapshot, S is the identifier of the industry correlation snapshot, and CNN is the CNN neural network.

[0040] As a preferred embodiment of the present invention, the method for determining the stable state of carbon emissions includes:

[0041] When d = 1, the steady state of carbon emissions of the target city at the time corresponding to d is: stable;

[0042] When d=0, the stable state of carbon emissions of the target city at the time corresponding to d is: unstable.

[0043] As a preferred solution of the present invention, the present invention provides a smart city carbon monitoring system based on data feedback, which is applied to a smart city carbon monitoring method based on data feedback. The system includes:

[0044] Gridding units are used to divide the target city into grids to form a grid map containing multiple carbon monitoring grids;

[0045] A data preprocessing unit, for establishing the industrial characteristic correlation between each carbon monitoring grid based on the location coordinate data and industrial economic data of each carbon monitoring grid, and mapping the industrial characteristic correlation to the grid map to obtain an industrial correlation snapshot;

[0046] and for establishing carbon emission characteristic correlations between the carbon monitoring grids based on the carbon emission data of the carbon monitoring grids, and mapping the carbon emission characteristic correlations to the grid map to obtain a carbon emission correlation snapshot;

[0047] A model building unit, used for building a carbon emission stability monitoring model for measuring similarity between industry correlation snapshots and carbon emission correlation snapshots using a twin neural network;

[0048] The monitoring output unit is used to determine the carbon emission stability state of the target city based on the output of the carbon emission stability monitoring model.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] Based on the location coordinate data and industrial economic data of the city's carbon monitoring grid, as well as the carbon emission data of each carbon monitoring grid, the present invention uses a twin neural network to establish a carbon emission stability monitoring model, so as to determine the carbon emission stability state of the target city based on the feedback of such industrial layout data as location coordinate data and industrial economic data, provide underlying data support for urban carbon monitoring, and improve the accuracy of urban carbon emission monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0052] Figure 1 A flow chart of a smart city carbon monitoring method based on data feedback provided by an embodiment of the present invention;

[0053] Figure 2 A block diagram of a smart city carbon monitoring system based on data feedback provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] like Figure 1 As shown, the present invention provides a smart city carbon monitoring method based on data feedback, comprising the following steps:

[0056] Divide the target city into grids to form a grid map containing multiple carbon monitoring grids;

[0057] Based on the location coordinate data and industrial economic data of each carbon monitoring grid, the industrial characteristic correlation between each carbon monitoring grid is established, and the industrial characteristic correlation is mapped into the grid map to obtain an industrial correlation snapshot;

[0058] Based on the carbon emission data of each carbon monitoring grid, the carbon emission characteristic correlation between each carbon monitoring grid is established, and the carbon emission characteristic correlation is mapped to the grid map to obtain a carbon emission correlation snapshot;

[0059] A carbon emission stability monitoring model is established using twin neural networks to measure the similarity between industry correlation snapshots and carbon emission correlation snapshots;

[0060] The carbon emission stability state of the target city is determined based on the output of the carbon emission stability monitoring model.

[0061] When monitoring the stability of urban carbon emissions, the present invention adopts industrial layout data that is correlated with carbon emissions as underlying data support, and utilizes data feedback from the industrial layout to achieve follow-up monitoring of carbon emissions. It can perceive changes in the industrial layout under the feedback of the industrial layout data, and monitor whether carbon emissions follow the changes in the industrial layout and make adaptive changes when the industrial layout changes, and adapt to the changed industrial layout. If it can adapt to the changed industrial layout, it means that the city's carbon emissions are in a stable state. If it cannot adapt to the changed industrial layout, it means that the city's carbon emissions are in an unstable state, thereby realizing urban carbon emission monitoring.

[0062] The present invention perceives changes in industrial layout under the feedback of industrial layout data, monitors whether carbon emissions have stably followed changes in industrial layout, and realizes instant monitoring at the moment when carbon emissions are most likely to be disordered when the industrial layout changes dynamically, thereby ensuring the timeliness of carbon emission monitoring. In addition, the stability of carbon emissions obtained through such monitoring can further provide data reference for evaluating the rationality of industrial layout.

[0063] The urban industrial layout data in the present invention consists of two parts. The first part is location-related information, which is quantified using location correlation. The similar geological environment between adjacent regions in the city will inevitably lead to certain similarities in the industrial layout between adjacent regions. The similarity of industrial layout will produce similar carbon emissions. The proximity of location can indirectly reflect the industrial carbon emissions of each region. Therefore, location correlation is used as part of the industrial layout data. The second part is industrial economic information, which is quantified using industrial correlation to quantify the GDP value (industrial economic data) of each region in the city. The GDP value can directly reflect the industrial carbon emissions of each region. Therefore, industrial correlation is used as part of the industrial layout data.

[0064] The present invention combines indirect indicators of location-related information and direct indicators of industry relevance into industry layout data, and maps them to a grid map of urban divisions. The industry layout data can be displayed in a matrix form, and the traditional map form of the entire city can be abstracted into a matrix image containing industry layout data. When changes are found in the industry layout data, the corresponding city matrix image will change. Therefore, after the industrial layout data changes, the city will form new matrix images one by one, which are recorded. This is a data snapshot (a snapshot can record the system status and data at any point in time), and the present invention names it an industry relevance snapshot.

[0065] Similarly, the present invention monitors the actual carbon emission data of each area in the city (i.e., carbon emission data), maps it to the grid map of the city division, and can display the carbon emission data in a matrix form, so as to abstract the traditional map form of the entire city into a matrix image containing carbon emission data. When changes are found in the carbon emission data, the corresponding city matrix image will change. Therefore, the city will form new matrix images one by one after the carbon emission data changes, and recording them is also a data snapshot (the snapshot can record the system status and data at any point in time). The present invention names it as a carbon emission correlation snapshot.

[0066] In order to measure the similarity between the industry correlation snapshot and the carbon emission correlation snapshot, the present invention uses a twin network to generate a carbon emission stability monitoring model. The twin network includes two CNN neural networks, which can perform feature extraction on the industry correlation snapshot and the carbon emission correlation snapshot, compare two input samples and output the similarity between them, and can measure the similarity between different samples in an embedding space. Therefore, the carbon emission stability monitoring model constructed by the twin network in the present invention can measure the similarity between the industry correlation snapshot and the carbon emission correlation snapshot, and grasp the adaptability between the carbon emission correlation snapshot and the industry correlation snapshot. If the adaptability is high, the model output is 1, and if the adaptability is not high, the model output is 0. According to the output value, it can be monitored whether carbon emissions have completed a stable transition under changes in industrial layout.

[0067] The twin network in the present invention adopts a CNN neural network structure that is good at processing images. It can achieve high-precision similarity comparison of two images and discover subtle differences between the two images. It is used for the industry correlation snapshot and carbon emission correlation snapshot comparison of the present invention, and can also realize the perception of subtle differences. The model is used for automatic and high-precision perception. Compared with manual comparison, the efficiency is improved and no difference information is missed.

[0068] In order to adapt the image processing performance of the twin network, the present invention grids the city, constructs an image framework for the city, fills in the industrial layout data and carbon emission data to complete the city image construction, and forms city images of two data types. Since they all belong to the same city, the image similarity comparison performance of the twin network can be used to discover whether there are differences between the industry correlation snapshot and the carbon emission correlation snapshot. The existence of differences indicates that the city's carbon emissions cannot adapt to changes in the industrial layout, in order to achieve a stable transition.

[0069] The grid diagram is: G = [a1, a2, ..., a n ], where G is the identifier of the grid graph, a1, a2 and a n They are the 1st, 2nd and nth carbon monitoring grids respectively.

[0070] The present invention combines indirect indicators of location-related information and direct indicators of industry relevance into industry layout data, maps them to the grid map of city division, and can display the industry layout data in a matrix form, so as to abstract the traditional map form of the entire city into a matrix image containing industry layout data. When changes are found in the industry layout data, the corresponding city matrix image will change. Therefore, after the industrial layout data changes, the city will form new matrix images one by one, which are recorded. This is a data snapshot (the snapshot can record the system status and data at any time point). The present invention names it as an industry relevance snapshot, which is as follows:

[0071] The methods for establishing the correlation of industry characteristics include:

[0072] The positional correlation of each carbon monitoring grid is quantified through the position coordinate data of each carbon monitoring grid. The positional correlation is:

[0073]

[0074] Where P ij is the position correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, P i is the location coordinate data of the i-th carbon monitoring grid, P j is the location coordinate data of the jth carbon monitoring grid;

[0075] The industrial relevance of each carbon monitoring grid is quantified through the industrial economic data of each carbon monitoring grid. The industrial relevance is:

[0076]

[0077] In the formula, G ij is the industry correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, G iis the industrial economic data of the i-th carbon monitoring grid, G j is the industrial economic data of the jth carbon monitoring grid;

[0078] Multiply the location correlation and industry correlation to obtain the industry characteristic correlation;

[0079] The quantitative formula for the correlation of industry characteristics is:

[0080] PG ij =P ij *G ij ;

[0081] In the formula, PG ij is the correlation between the industrial characteristics of the i-th carbon monitoring grid and the j-th carbon monitoring grid, P ij is the position correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, G ij is the industry correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid.

[0082] The industry relevance snapshot is: Where S is the identifier of the industry relevance snapshot, PG 11 is the correlation between the first carbon monitoring grid and the industry characteristics of the first carbon monitoring grid, PG 1n is the correlation between the industry characteristics of the first carbon monitoring grid and the nth carbon monitoring grid, PG n1 is the correlation between the industrial characteristics of the nth carbon monitoring grid and the first carbon monitoring grid, PG nn is the correlation between the industry characteristics of the nth carbon monitoring grid and the nth carbon monitoring grid.

[0083] In order to convert the city grid map from one dimension to a two-dimensional image form that the CNN neural network is good at, so as to obtain more feature information in the processing process, the location information and industrial economic data of each carbon monitoring grid are converted into the correlation between each carbon monitoring grid, that is, converted into location correlation and industry correlation, which can characterize the industrial layout data of the city and reflect the correlation between the industrial layout data of each region. The constructed industrial correlation snapshot contains more information in a two-dimensional form for the CNN neural network to extract, thereby obtaining more detailed comparison information and ultimately obtaining more accurate comparison results.

[0084] The methods for establishing the correlation of carbon emission characteristics include:

[0085] The carbon emission characteristics correlation of each carbon monitoring grid is quantified through the carbon emission data of each carbon monitoring grid. The carbon emission characteristics correlation is:

[0086]

[0087] In the formula, Cij is the correlation of carbon emission characteristics between the i-th carbon monitoring grid and the j-th carbon monitoring grid, C i is the carbon emission data of the i-th carbon monitoring grid, C j is the carbon emission data of the jth carbon monitoring grid.

[0088] A snapshot of carbon emissions relevance is: Where H is the identifier of the carbon emission correlation snapshot, C 11 is the correlation between the carbon emission characteristics of the first carbon monitoring grid and the first carbon monitoring grid, C 1n is the correlation of carbon emission characteristics between the first carbon monitoring grid and the nth carbon monitoring grid, C n1 is the carbon emission characteristic correlation between the nth carbon monitoring grid and the first carbon monitoring grid, C nn is the correlation between the carbon emission characteristics of the n-th carbon monitoring grid and the n-th carbon monitoring grid.

[0089] In order to convert the city grid map from one dimension to a two-dimensional image form that the CNN neural network is good at, so as to obtain more feature information in the processing process, the present invention converts the carbon emission data of each carbon monitoring grid into the correlation between the carbon monitoring grids, that is, into carbon emission correlation, which can characterize the actual carbon emission situation of the city and reflect the correlation of carbon emissions between regions. The constructed carbon emission correlation snapshot contains more information in a two-dimensional form for the CNN neural network to extract, thereby obtaining more detailed comparison information and ultimately obtaining more accurate comparison results.

[0090] In order to measure the similarity between the industry correlation snapshot and the carbon emission correlation snapshot, the present invention uses a twin network to generate a carbon emission stability monitoring model. The twin network includes two CNN neural networks, which can extract features from the industry correlation snapshot and the carbon emission correlation snapshot, compare two input samples and output the similarity between them, and can measure the similarity between different samples in an embedding space. Therefore, the carbon emission stability monitoring model constructed by the twin network in the present invention can measure the similarity between the industry correlation snapshot and the carbon emission correlation snapshot, and grasp the adaptability between the carbon emission correlation snapshot and the industry correlation snapshot. If the adaptability is high, the model output is 1, and if the adaptability is not high, the model output is 0. According to the output value, it can be monitored whether the carbon emission has completed a stable transition under the change of industrial layout, as follows:

[0091] The construction method of the carbon emission stability monitoring model includes:

[0092] Select a time period when carbon emissions are in a stable state as a positive sample time period, and obtain carbon emission correlation snapshots and industry correlation snapshots at each time in the positive sample time period;

[0093] The similarity measurement values ​​of the carbon emission correlation snapshot and the industry correlation snapshot at each moment in the positive sample time period are marked as 1;

[0094] Select a time period when carbon emissions are in an unstable state as a negative sample time period, and obtain carbon emission correlation snapshots and industry correlation snapshots at each time in the negative sample time period;

[0095] The similarity measurement values ​​of the carbon emission correlation snapshot and the industry correlation snapshot at each moment in the negative sample time period are marked as 0;

[0096] The carbon emission correlation snapshots and industry correlation snapshots and similarity measurement values ​​in the positive sample time period and the negative sample time period are mixed into a data set, and the data set is divided into a training set and a test set;

[0097] Using the training set, taking the carbon emission correlation snapshot and the industry correlation snapshot as the two inputs of the twin network, and taking the similarity measurement value as the output of the twin network, the twin network is trained to generate a carbon emission stability monitoring model;

[0098] The performance of the carbon emission stability monitoring model was evaluated using the test set and the mean square error as the evaluation indicator.

[0099] The carbon emission stability monitoring model is:

[0100]

[0101] Where d is the similarity measure between the carbon emission correlation snapshot and the industry correlation snapshot, H is the identifier of the carbon emission correlation snapshot, S is the identifier of the industry correlation snapshot, and CNN is the CNN neural network.

[0102] The determination methods for the steady state of carbon emissions include:

[0103] When d = 1, the steady state of carbon emissions of the target city at the time corresponding to d is: stable;

[0104] When d=0, the stable state of carbon emissions of the target city at the time corresponding to d is: unstable.

[0105] like Figure 2 As shown, the present invention provides a smart city carbon monitoring system based on data feedback, which is applied to a smart city carbon monitoring method based on data feedback. The system includes:

[0106] Gridding units are used to divide the target city into grids to form a grid map containing multiple carbon monitoring grids;

[0107] A data preprocessing unit, used to establish the industrial characteristic correlation between each carbon monitoring grid based on the location coordinate data and industrial economic data of each carbon monitoring grid, and map the industrial characteristic correlation to the grid map to obtain an industrial correlation snapshot;

[0108] and for establishing the carbon emission characteristic correlation between each carbon monitoring grid based on the carbon emission data of each carbon monitoring grid, and mapping the carbon emission characteristic correlation to the grid map to obtain a carbon emission correlation snapshot;

[0109] A model building unit, used for building a carbon emission stability monitoring model for measuring similarity between industry correlation snapshots and carbon emission correlation snapshots using a twin neural network;

[0110] The monitoring output unit is used to determine the carbon emission stability state of the target city based on the output of the carbon emission stability monitoring model.

[0111] Based on the location coordinate data and industrial economic data of the city's carbon monitoring grid, as well as the carbon emission data of each carbon monitoring grid, the present invention uses a twin neural network to establish a carbon emission stability monitoring model, so as to determine the carbon emission stability state of the target city based on the feedback of such industrial layout data as location coordinate data and industrial economic data, provide underlying data support for urban carbon monitoring, and improve the accuracy of urban carbon emission monitoring.

[0112] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.

Claims

1. A smart city carbon monitoring method based on data feedback, characterized in that: The following steps are involved: Divide the target city into grids to form a grid map containing multiple carbon monitoring grids; Based on the location coordinate data and industrial economic data of each carbon monitoring grid, establish the industrial characteristic correlation between each carbon monitoring grid, and map the industrial characteristic correlation to the grid map to obtain an industrial correlation snapshot; Based on the carbon emission data of each carbon monitoring grid, establish the carbon emission characteristic correlation between each carbon monitoring grid, and map the carbon emission characteristic correlation to the grid map to obtain a carbon emission correlation snapshot; A carbon emission stability monitoring model is established using twin neural networks to measure the similarity between industry correlation snapshots and carbon emission correlation snapshots; The carbon emission stability state of the target city is determined based on the output of the carbon emission stability monitoring model.

2. According to claim 1, a smart city carbon monitoring method based on data feedback is characterized in that: The grid graph is: G = [a1, a2, ..., a n ], where G is the identifier of the grid graph, a1, a2 and a n They are the 1st, 2nd and nth carbon monitoring grids respectively.

3. The smart city carbon monitoring method based on data feedback according to claim 2 is characterized in that: The methods for establishing the correlation of industry characteristics include: The positional correlation of each carbon monitoring grid is quantified by the position coordinate data of each carbon monitoring grid, and the positional correlation is: Where P ij is the position correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, P i is the location coordinate data of the i-th carbon monitoring grid, P j is the location coordinate data of the jth carbon monitoring grid; The industrial relevance of each carbon monitoring grid is quantified through the industrial economic data of each carbon monitoring grid. The industrial relevance is: In the formula, G ij is the industry correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, G i is the industrial economic data of the i-th carbon monitoring grid, G j The industrial economic data for the ,th carbon monitoring grid; Multiplying the location correlation and the industry correlation to obtain the industry characteristic correlation; The quantitative formula for the correlation of the industrial characteristics is: PG ij =P ij *G ij ; In the formula, PG ij is the correlation between the industrial characteristics of the ith carbon monitoring grid and the ith carbon monitoring grid, P ij is the position correlation between the i-th carbon monitoring grid and the j-th carbon monitoring grid, G ij is the industry correlation between the ith carbon monitoring grid and the ith carbon monitoring grid.

4. The smart city carbon monitoring method based on data feedback according to claim 3 is characterized by: The industry relevance snapshot is: Where S is the identifier of the industry relevance snapshot, PG 11 is the correlation between the first carbon monitoring grid and the industry characteristics of the first carbon monitoring grid, PG 1n is the correlation between the industry characteristics of the first carbon monitoring grid and the nth carbon monitoring grid, PG n1 is the correlation between the industrial characteristics of the nth carbon monitoring grid and the first carbon monitoring grid, PG nn is the correlation between the industry characteristics of the nth carbon monitoring grid and the nth carbon monitoring grid.

5. The smart city carbon monitoring method based on data feedback according to claim 4 is characterized in that: The method for establishing the carbon emission characteristic correlation includes: The carbon emission characteristic correlation of each carbon monitoring grid is quantified through the carbon emission data of each carbon monitoring grid. The carbon emission characteristic correlation is: In the formula, C ij is the correlation of carbon emission characteristics between the i-th carbon monitoring grid and the j-th carbon monitoring grid, C i is the carbon emission data of the i-th carbon monitoring grid, C j is the carbon emission data of the jth carbon monitoring grid.

6. The smart city carbon monitoring method based on data feedback according to claim 5 is characterized in that: The carbon emission correlation snapshot is: Where H is the identifier of the carbon emission correlation snapshot, C 11 is the correlation between the carbon emission characteristics of the first carbon monitoring grid and the first carbon monitoring grid, C 1n is the correlation of carbon emission characteristics between the first carbon monitoring grid and the nth carbon monitoring grid, C n1 is the carbon emission characteristic correlation between the nth carbon monitoring grid and the first carbon monitoring grid, C nn is the correlation between the carbon emission characteristics of the n-th carbon monitoring grid and the n-th carbon monitoring grid.

7. The smart city carbon monitoring method based on data feedback according to claim 6 is characterized by: The method for constructing the carbon emission stability monitoring model includes: Select a time period when carbon emissions are in a stable state as a positive sample time period, and obtain a carbon emission correlation snapshot and an industry correlation snapshot at each time in the positive sample time period; Marking the similarity measurement values ​​of the carbon emission correlation snapshot and the industry correlation snapshot at each time point in the positive sample time period as 1; Select a time period when carbon emissions are in an unstable state as a negative sample time period, and obtain a carbon emission correlation snapshot and an industry correlation snapshot at each time in the negative sample time period; Marking the similarity measurement values ​​of the carbon emission correlation snapshot and the industry correlation snapshot at each time point in the negative sample time period as 0; The carbon emission correlation snapshots and industry correlation snapshots and similarity measurement values ​​in the positive sample time period and the negative sample time period are mixed into a data set, and the data set is divided into a training set and a test set; Using the training set, taking the carbon emission correlation snapshot and the industry correlation snapshot as two input quantities of the twin network, taking the similarity measurement value as the output quantity of the twin network, training the twin network, and generating a carbon emission stability monitoring model; The performance of the carbon emission stability monitoring model was evaluated using the test set and the mean square error as the evaluation indicator.

8. The smart city carbon monitoring method based on data feedback according to claim 7 is characterized by: The carbon emission stability monitoring model is: Where d is the similarity measure between the carbon emission correlation snapshot and the industry correlation snapshot, H is the identifier of the carbon emission correlation snapshot, S is the identifier of the industry correlation snapshot, and CNN is the CNN neural network.

9. The smart city carbon monitoring method based on data feedback according to claim 8 is characterized in that: The methods for determining the steady state of carbon emissions include: When d = 1, the steady state of carbon emissions of the target city at the time corresponding to d is: stable; When d=0, the stable state of carbon emissions of the target city at the time corresponding to d is: unstable.

10. A smart city carbon monitoring system based on data feedback, characterized in that: A smart city carbon monitoring method based on data feedback applied to any one of claims 1 to 9, the system comprising: Gridding units are used to divide the target city into grids to form a grid map containing multiple carbon monitoring grids; A data preprocessing unit, for establishing the industrial characteristic correlation between each carbon monitoring grid based on the location coordinate data and industrial economic data of each carbon monitoring grid, and mapping the industrial characteristic correlation to the grid map to obtain an industrial correlation snapshot; and for establishing carbon emission characteristic correlations between the carbon monitoring grids based on the carbon emission data of the carbon monitoring grids, and mapping the carbon emission characteristic correlations to the grid map to obtain a carbon emission correlation snapshot; A model building unit, used for building a carbon emission stability monitoring model for measuring similarity between industry correlation snapshots and carbon emission correlation snapshots using a twin neural network; The monitoring output unit is used to determine the carbon emission stability state of the target city based on the output of the carbon emission stability monitoring model.

Images