Regional carbon emission prediction method and system based on carbon emission monitoring, and electronic equipment

Abstract

The invention discloses a regional carbon emission prediction method and system based on carbon emission monitoring and electronic equipment, and relates to the technical field of carbon emission prediction, and the method comprises the steps: collecting regional discrete point emission data through a multi-modal sensing network, and generating a regional-level continuous flow field; performing feature decoupling and quantization on different emission source types to generate a source stream feature vector matrix; reconstructing a continuous space-time emission field based on the source stream feature vector matrix, identifying a conduction path, and dynamically binding with field features along the path to generate a dynamic field data set; wherein the field characteristics comprise emission intensity, diffusion rate and regional coupling coefficient; constructing a field domain succession rule model based on the dynamic field data set, identifying a dominant evolution path based on an evolution baseline library of the model, injecting external disturbance along a conduction path of the dominant evolution path, and generating a multi-branch evolution scene tree; wherein the external disturbance comprises policy regulation and control intensity and a meteorological anomaly index. According to the invention, accurate prediction of regional carbon emission is realized.

Description

Technical Field

[0001] This invention relates to the field of carbon emission prediction technology, and in particular to a regional carbon emission prediction method, system and electronic equipment based on carbon emission monitoring. Background Technology

[0002] Precise carbon emission control is a crucial issue for both nations and industries. Regional carbon emission forecasting technologies, through quantitative analysis of emission trends, can provide critical decision support for carbon quota allocation, emission reduction pathway optimization, and policy formulation. However, because carbon emissions are influenced by various factors, such as energy structure, industrial activities, traffic flow, and climate conditions, traditional forecasting methods often struggle to balance high accuracy, real-time performance, and dynamic adaptability, leading to significant discrepancies between forecast results and actual emissions.

[0003] In related technologies, patent application CN118484663A discloses a regional carbon emission prediction method based on socioeconomic data. This method divides regions based on socioeconomic statistical data and normalizes the carbon emission processing dataset and the initial influencing factor processing dataset through standard energy conversion coefficients. Although it improves the ability to understand trends in different regions through sub-regional analysis and selection of corresponding initial influencing factor datasets, it relies on a single factor of socioeconomic statistical data, resulting in the static accounting defect of historical carbon emission data.

[0004] The patent application with patent number CN119441802A discloses a carbon emission prediction method and system. This technology improves the prediction reliability by constructing a life cycle carbon emission model, but its regional division ignores the carbon flow and transmission mechanism between adjacent regions, and the spatial discretization error caused by the regional division is relatively large.

[0005] The patent application with patent number CN118710222A discloses a carbon emission monitoring system and method. The scheme achieves carbon emission regulation through regional classification and profit and loss complementarity, but fails to quantify the nonlinear coupling effect of multiple factors.

[0006] The related technologies suffer from several drawbacks, including the fragmentation of carbon flow continuity by discretization of regions, the amplification of errors due to the lack of modeling of cross-regional conduction effects, the lack of dynamic decoupling between static data and real-time monitoring signals, and insufficient coupling prediction of multi-factor nonlinear interactions, thus limiting the accuracy of the prediction results. Summary of the Invention

[0007] This invention provides a regional carbon emission prediction method, system, electronic device, and storage medium based on carbon emission monitoring, which solves the problem of how to improve the accuracy of regional carbon emission prediction.

[0008] To achieve the above objectives, this application adopts the following technical solution: Firstly, a regional carbon emission prediction method based on carbon emission monitoring is provided, including: Regional discrete point emission data are collected through a multimodal sensor network to generate a regional continuous flow field; based on orthogonal constraints, different emission source types are decoupled and quantized to generate a source flow feature vector matrix with three-dimensional geocoding and dual timestamps. Based on the source-flow feature vector matrix, a continuous spatiotemporal emission field is reconstructed, the conduction path is identified, and the field features are dynamically bound along the path to generate a dynamic field dataset; wherein, the field features include emission intensity, diffusion rate, and regional coupling coefficient. Based on the dynamic field dataset, a field succession law model is constructed. Based on the field succession law model, an evolutionary baseline library is constructed. Based on the evolutionary baseline library, the dominant evolutionary path is identified, and external disturbances are injected along its transmission path to generate a multi-branch evolutionary scenario tree. The external disturbances include the intensity of policy regulation and meteorological anomaly index.

[0009] Secondly, a regional carbon emission prediction system based on carbon emission monitoring is provided, including: The dynamic source flow analysis module is used to collect regional discrete point emission data through a multimodal sensor network and generate a regional continuous flow field; based on orthogonal constraints, it performs feature decoupling and quantization on different emission source types to generate a source flow feature vector matrix with three-dimensional geocoding and dual timestamps. The spatiotemporal field reconstruction module is used to reconstruct a continuous spatiotemporal emission field based on the source-flow feature vector matrix, identify the conduction path, dynamically bind the field features along the path, and generate a dynamic field dataset; wherein, the field features include emission intensity, diffusion rate, and regional coupling coefficient; The adaptive evolutionary deduction module is used to construct a field succession law model based on the dynamic field dataset, construct an evolutionary baseline library based on the field succession law model, identify the dominant evolutionary path based on the evolutionary baseline library, inject external disturbances along its transmission path, and generate a multi-branch evolutionary scenario tree; wherein, the external disturbances include the intensity of policy regulation and meteorological anomaly index.

[0010] Thirdly, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the regional carbon emission prediction method based on carbon emission monitoring as described in the first aspect.

[0011] Fourthly, a readable storage medium is provided, on which a program or instructions are stored, which, when executed by a processor, implement the steps of the regional carbon emission prediction method based on carbon emission monitoring as described in the first aspect.

[0012] The regional carbon emission prediction method based on carbon emission monitoring of the present invention has the following beneficial effects: By leveraging real-time sensor data and dynamic decoupling algorithms, the system addresses the challenge of dynamic decoupling from multiple sources. It constructs a continuous flow field based on discrete monitoring points, thereby reconstructing the spatiotemporal emission field and achieving seamless modeling of cross-regional carbon flows. This eliminates the boundary effects of rigid regional divisions and significantly reduces spatial discretization errors. Furthermore, it generates scenario trees by injecting perturbations, integrating the coupled influences of multiple factors such as policy and meteorology. This enables accurate carbon emission prediction and provides a highly reliable simulation tool for carbon quota allocation and emission reduction pathway optimization.

[0013] The system, electronic equipment, and readable storage medium of this invention, which are the same as the regional carbon emission prediction method based on carbon emission monitoring, can achieve the same technical effect. To avoid duplication, they will not be described in detail here. Attached Figure Description

[0014] Figure 1 A schematic flowchart illustrating a regional carbon emission prediction method based on carbon emission monitoring, provided for embodiments of this application; Figure 2 A schematic flowchart illustrating another regional carbon emission prediction method based on carbon emission monitoring provided in this application embodiment; Figure 3 A schematic diagram of a regional carbon emission prediction system based on carbon emission monitoring is provided as an embodiment of this application; Figure 4 A schematic diagram of a regional carbon emission prediction device based on carbon emission monitoring is provided in this application embodiment; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0015] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions in the embodiments of this application are clearly described. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art are within the scope of protection of this application.

[0016] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0017] The steps described in the specification and the flowcharts in the accompanying drawings of this invention are not necessarily strictly executed according to the step numbers; the execution order of the method steps can be changed. Furthermore, certain steps can be omitted, multiple steps can be combined into one step, and / or one step can be broken down into multiple steps.

[0018] The existing technologies suffer from drawbacks such as fragmented carbon flow continuity due to discretized regional divisions, amplified errors due to unmodeled cross-regional transmission effects, lack of dynamic decoupling between static data and real-time monitoring signals, and insufficient coupling prediction of multi-factor nonlinear interactions. There is an urgent need to develop a regional carbon emission prediction method that integrates dynamic monitoring, spatial continuous modeling, and multi-factor coupling analysis to overcome the current technological bottlenecks and provide a high-precision decision-making tool for carbon neutrality strategies.

[0019] This specification provides a regional carbon emission prediction method based on carbon emission monitoring, and also relates to a regional carbon emission prediction system based on carbon emission monitoring, a regional carbon emission prediction device based on carbon emission monitoring, a computer device, and a computer-readable storage medium. These will be described in detail below with reference to the accompanying drawings and preferred embodiments.

[0020] Please see Figure 1-2 This application provides a regional carbon emission prediction method based on carbon emission monitoring, as shown in the figure, including: Step S1: Collect regional discrete point emission data through a multimodal sensor network to generate a regional continuous flow field; perform feature decoupling and quantization on different emission source types based on orthogonal constraints to generate a source flow feature vector matrix with three-dimensional geocoding and dual timestamps.

[0021] This step establishes a dynamic carbon emission source flow tracking program based on multimodal sensor data, transforming instantaneous emissions from discrete monitoring points into a regional continuous flow field distribution. From the regional continuous flow field distribution of mixed emission sources, the fluctuation characteristics of different emission types, such as industrial, transportation, and biological sources, are separated, forming a source flow feature vector matrix with three-dimensional geocoding and dual timestamps. Further, step S1 includes: Step S11: Obtain the regional continuous flow field distribution; wherein, the regional continuous flow field distribution includes the spatiotemporally continuous emission intensity, diffusion direction and emission waveform characteristics.

[0022] Step S12: Based on the emission waveform in the flow field, perform spatiotemporal fluctuation pattern matching to construct feature dimensions including steady-state fluctuation axis, pulse fluctuation axis and environmental coupling axis.

[0023] This step constructs feature dimensions based on the propagation characteristics of emission waveforms in the flow field; specifically: steady-state fluctuation axis, capturing high-frequency regular vibrations of industrial source emissions, such as the fixed-period waveform of continuous emissions from factory chimneys; pulse fluctuation axis, extracting short-term high-intensity spikes of traffic source emissions, such as the sudden emission waveform generated by road vehicle acceleration; environmental coupling axis, analyzing the correlation patterns between low-frequency fluctuations of biological source emissions and surface temperature and vegetation index, such as the diurnal cycle waveform of forest respiration.

[0024] Step S13: Perform feature dimension unwrapping based on the orthogonal constraint conditions between feature dimension axes.

[0025] This step achieves feature separation through orthogonal constraints between fluctuation axes; specifically: applying time continuity constraints to the steady-state fluctuation axis to eliminate the interference of traffic pulses on industrial source features; applying spatial path constraints to the pulse fluctuation axis to retain only transient features that coincide with the traffic network topology; and applying ecological parameter correlation constraints to the environmental coupling axis to filter out random fluctuations unrelated to natural factors.

[0026] Step S14: Perform dynamic feature vectorization for each feature dimension: For the steady-state fluctuation axis, obtain the amplitude envelope and frequency stability index of the steady-state fluctuation; For the pulse wave axis, the peak density and spatial propagation attenuation rate of the pulse waveform are statistically analyzed; For the environmental coupling axis, evaluate the time-delay correlation coefficient between low-frequency fluctuations and environmental parameters.

[0027] This step implements a differentiated feature extraction strategy for each emission source type; that is, industrial source quantification obtains the amplitude envelope and frequency stability index of steady-state fluctuations; traffic source quantification calculates the peak density and spatial propagation attenuation rate of pulse waveforms; and biological source quantification assesses the time-delay correlation coefficient between low-frequency fluctuations and environmental parameters.

[0028] Step S15: Dynamically associate the quantization results with the spatiotemporal coordinates of the regional continuous flow field, add a three-dimensional geocoding of longitude, latitude and elevation layers to each spatial unit, embed a dual timestamp of absolute timestamp and emission cycle phase into each time slice, and output the source flow feature vector matrix.

[0029] Specifically, the matrix row vector structure is: [spatial encoding, timestamp, industrial features, traffic features, biological features].

[0030] In practice, this step involves collecting discrete emission data through multimodal sensor networks, such as satellite remote sensing, ground sensors, and mobile monitoring equipment. A Gaussian plume diffusion model is established using flow field inversion algorithms to reconstruct the point source data into a two-dimensional planar flow field distribution. Spectral clustering is used to decouple the features of mixed emission sources, separating the fluctuation characteristics of different emission types, such as industrial, traffic, and biological sources, from the regional continuous flow field distribution of mixed emission sources. The output is a time-resolution feature vector matrix, including the feature parameters of each source category.

[0031] The above process achieves refined separation and structured characterization of multi-source carbon emissions in the region. Multi-source emission characteristics are decoupled by forming an orthogonal feature space through spatiotemporal fluctuation pattern matching, constructing a steady-state fluctuation axis, a pulse fluctuation axis, and an environmental coupling axis. Based on the physical differences between high-frequency vibrations from industrial sources, transient pulses from transportation sources, and low-frequency fluctuations from biological sources, the feature dimensions are untangled and separated. Applied temporal continuity constraints, spatial path constraints, and ecological parameter correlation constraints ensure the independence of each emission source's characteristics in the time, spatial, and frequency domains. Differential feature quantification is achieved by implementing feature extraction strategies for the physical characteristics of different emission sources: industrial sources are characterized by their periodic regularity using amplitude envelopes and frequency stability indices; transportation sources are described by their transient propagation characteristics using peak density and spatial attenuation rate; and biological sources are quantified using time-delay correlation coefficients to quantify their dynamic response relationship with environmental parameters. This quantification method preserves the typical characteristic patterns of various emission sources. Spatiotemporal structured representation, three-dimensional geocoding, longitude, latitude, and elevation layers, dual timestamps, and a binding mechanism for absolute time and periodic phases transform the decoupled multi-source characteristic flow field into a feature vector matrix with a clear spatiotemporal reference system. The row vector structure of the matrix [spatial coding, timestamp, industrial features, traffic features, biological features] realizes the systematic correlation between emission features and spatiotemporal coordinates. The output source flow feature vector matrix provides physically interpretable structured input data for subsequent spatiotemporal field reconstruction. Its technical value lies in simultaneously meeting three core requirements: the purity of emission source separation, the representativeness of feature quantification, and the accuracy of spatiotemporal positioning.

[0032] Step S2: Based on the source flow feature vector matrix, a continuous spatiotemporal emission field is reconstructed, the conduction path is identified, and the field features are dynamically bound along the path to generate a dynamic field dataset; wherein, the field features include emission intensity, diffusion rate and regional coupling coefficient.

[0033] This step reconstructs the discrete source-flow features of the source-flow feature vector matrix into a continuous spatiotemporal emission field; during the reconstruction process, potential channels for emission conduction between regions are automatically identified, forming a dynamic field mechanism including emission intensity, diffusion rate, and regional coupling coefficient. Further, step S2 includes: Step S21: Construct a three-dimensional spatiotemporal convolution kernel group based on the source-flow feature vector matrix, and scan the source-flow feature vector matrix through the three-dimensional spatiotemporal convolution kernel group to generate a field correlation map.

[0034] Step S21 specifically includes: Step S211: Analyze the source-flow feature vector matrix, and based on its spatial resolution and temporal density dynamics, generate a convolutional kernel structure including a spatial kernel and a temporal kernel, and define a multi-scale spatiotemporal coverage range for each convolutional kernel; thus obtaining a three-dimensional spatiotemporal convolutional kernel group including a spatial nested kernel, a temporally adaptive kernel, and multi-scale receptive field parameters. Among them, the spatial core forms a pyramid-shaped nested structure along the latitude and longitude direction, with the bottom core covering high-resolution micro-regions and the upper core covering low-resolution macro-regions; Time core: The time window length is automatically adjusted according to the emission cycle phase. For example, industrial sources use a long-cycle core to capture production rhythms, while traffic sources use a short-cycle core to match peak fluctuations. Microscale receptive field: focuses on high-frequency changes in local emission hotspots, such as the instantaneous emission interactions of factory chimney clusters; macroscale receptive field: captures low-frequency trends in emission transmission between regions, such as atmospheric diffusion processes across urban clusters.

[0035] Step S212: Based on the three-dimensional spatiotemporal convolution kernel group scanning source flow feature vector matrix, generate a field correlation map including spatial correlation and temporal correlation.

[0036] Step S22: Identify potential transmission channels based on the field correlation degree distribution of the field correlation map, and dynamically bind the path of the transmission channel with field characteristics including emission intensity gradient, diffusion rate and regional coupling coefficient.

[0037] Among them, emission intensity gradient: the intensity decay rate is calculated along the transmission path; Diffusion rate: Atmospheric diffusion efficiency is inverted based on spatial correlation and time delay; Regional coupling coefficient: quantifies the intensity of mutual influence between regions by using the entropy value of the transmission path.

[0038] In some possible implementations, step S22 further includes: Step S221: Construct a conduction channel topology network including channel levels, spatial coordinate chains, and entropy labels. Specifically, this includes: Step S2210: Hierarchical fusion is performed on the output of the multi-scale receptive field to obtain a preliminary continuous spatiotemporal emission field containing intensity distribution, spatial correlation degree and time delay parameters; Step S2211: Select adjacent region chains that have a spatial correlation degree that is consistently higher than a set dynamic threshold and a temporal correlation degree that fluctuates below a set tolerance range as the main transmission path; Step S2212: Identify the abnormal path where the spatial correlation suddenly increases but the temporal correlation shows periodic oscillations as the secondary transmission path; Step S2213: Based on the preliminary continuous spatiotemporal emission field, construct the transmission path entropy index based on the field correlation distribution; Step S2214: Classify channel levels based on the propagation path entropy value: The main transmission path with an entropy value below a set threshold is designated as a primary channel; such as a stable diffusion corridor along the prevailing wind direction. Secondary conduction paths with entropy values ​​within a set threshold range are classified as secondary channels; such as intermittent diffusion channels formed by topographic uplift.

[0039] Step S222: Based on the conduction channel topology network, the dynamic quantization parameter set is output for the quantization field characteristics, including: Emission intensity gradient field: The decay rate per unit distance is calculated by the intensity difference between adjacent nodes, such as the emission intensity reduction rate from industrial area to residential area; Diffusion rate field: Combining spatial correlation, waveform similarity and time delay, atmospheric diffusion efficiency is inferred from the response hysteresis; The regional coupling matrix includes direct coupling coefficients and indirect coupling coefficients; direct coupling coefficients: the assessment value of the influence intensity between adjacent regions connected by the main transmission path; indirect coupling coefficients: the assessment value of the cross-regional cascade effect formed by the secondary transmission path.

[0040] Step S223: Dynamically associate the dynamic quantization parameter set with the physical characteristics of the conduction channel.

[0041] Specifically, it includes: Step S2231: Dynamically match the emission intensity gradient vector with the direction of the transmission channel, and the angle between the spatial direction of the emission intensity gradient vector and the extension direction of the transmission path is less than a set angle threshold.

[0042] Step S2232: Automatically adjust the gradient to obtain the step size according to the conduction channel level. The first-level channel uses a large step size, and the second-level channel uses a fine step size.

[0043] Step S2233: Based on the dynamic allocation rate calculation rules of the conduction channel level, the first-level channel uses the global baseline value of the diffusion rate field, while the second-level channel applies a terrain shading coefficient to the baseline rate in combination with the conduction path entropy value, such as reducing diffusion efficiency in mountainous areas. In other words, the first-level channel uses the global rate, while the second-level channel uses a locally corrected rate.

[0044] The above steps aim to achieve parameter space mapping, embedding intensity gradient vectors at each node of the conduction path. The magnitude represents the attenuation rate, and the direction indicates the diffusion trend. Differential diffusion rate calculation weights are assigned based on the channel level. Dynamic adaptation between the conduction path and the gradient field is achieved through parameter space mapping: the angle constraint between the intensity gradient vector and the direction of the conduction channel ensures that the field force direction matches the medium structure. A hierarchical step-size strategy, with large-scale step sizes for first-level channels and fine-scale step sizes for second-level channels, optimizes parameter acquisition at different spatial scales. A rate allocation mechanism based on channel level, using a baseline value for first-level channels and adding a terrain coefficient for second-level channels, achieves differentiated control of diffusion efficiency. The final output spatialized parameter correlator includes gradient path constraints, hierarchical rate weights, and a coupling matrix, providing structured input for subsequent processing.

[0045] Through structured analysis of the conduction path, discrete feature fields are transformed into hierarchical networks with entropy quantification characteristics, forming a spatial conduction system including primary / secondary channel classification. Channel level division, with first / secondary levels determined by strict entropy thresholds, ensures path stability and the physical interpretability of anomalous conduction. Dynamic coupling analysis of field features and parallel calculation of the intensity gradient field and diffusion rate field enable the synchronous quantification of the spatial attenuation characteristics of emission intensity and atmospheric diffusion efficiency. The introduction of direct / indirect coupling coefficients establishes a mathematical representation model of the regional impact of primary and secondary conduction paths. By constructing a multi-parameter collaborative optimization mechanism, physical features, gradient vector direction, channel level, and dynamic parameters, rate weights, and coupling coefficients are spatially correlated to form an adaptive field mechanism adjustment capability. Specifically, this manifests as: entropy-based channel level difference processing, global / local rate calculation, dynamic adjustment of coupling coefficients, primary channel enhancement / secondary channel attenuation, and an anomalous conduction path suppression mechanism controlled by attenuation factors. The final output is an organic fusion of the gradient field, rate field, regional coupling matrix, and conduction topology—a dynamic field dataset.

[0046] A dynamic field mechanism was constructed, forming a complete analytical framework from the original emission field to the dynamic transmission mechanism. Through a three-stage process of hierarchical channel network, parameterized field characteristics, and dynamic fusion mechanism, a regional interaction model with clear physical meaning and quantifiable verification was output. It also includes the static topological properties and dynamic interaction parameters of spatial transmission paths, providing a structured analysis tool for the study of the mechanism of environmental processes such as atmospheric diffusion.

[0047] In some possible implementations, multi-parameter field mechanism fusion optimization is also performed, including: Step S224: Dynamically adjust the region coupling matrix based on the propagation path entropy value.

[0048] Specifically, the direct coupling coefficient of the connection region of the main transmission path with low entropy is enhanced; a coupling attenuation factor is applied to the secondary transmission path with entropy fluctuations; and gradient continuity constraints are implemented along the transmission path.

[0049] In practical implementation, the coupling matrix is ​​dynamically optimized based on the entropy value of the conduction path. For conduction paths with entropy values ​​below the path threshold, the direct coupling coefficient of their connecting regions increases exponentially. When the entropy value of a secondary conduction path fluctuates beyond the tolerance range, a dynamic adjustment mechanism for the coupling attenuation factor is triggered. For example, when the entropy value suddenly increases, the indirect coupling coefficient is automatically reduced to suppress the excessive influence of abnormal conduction. Gradient continuity is guaranteed along the conduction path, ensuring that the difference in intensity attenuation between adjacent nodes does not exceed the maximum allowable gradient mutation threshold of the conduction path. For nodes that violate conservation constraints, the gradient recalculation process is automatically triggered. The output optimized field parameter system includes a smoothed gradient field, an entropy-controlled coupling matrix, and a graded velocity field.

[0050] Furthermore, to further optimize the field mechanism dynamically, it also includes: Step S2241, conduction stability assessment: Step S2241a: Perform stability verification on the transmission path with entropy value lower than the path threshold to verify whether the temporal correlation fluctuation of the low entropy channel remains in a stable range. Step S2241b: Check whether the entropy gradient of adjacent nodes on the conduction path conforms to the physical laws of atmospheric diffusion. Step S2241c involves monitoring entropy fluctuations in the secondary conduction path, capturing the spatiotemporal coupling relationship between entropy surge events and meteorological parameters, such as sudden wind speed changes and the formation of inversion layers, and detecting whether entropy fluctuations exhibit regular periodic characteristics, such as conduction enhancement phenomena occurring at fixed times each day.

[0051] Finally, a channel stability assessment report is output, including the main channel verification mark, secondary channel anomaly label, and related environmental parameters.

[0052] Step S2242, Dynamic adjustment of coupling strength: Step S2242a: Implement exponential enhancement on the verified main conduction path. Dynamically calculate the enhancement factor based on the extent to which the entropy value is lower than a set threshold. For example, when the entropy value is 20% lower than the threshold, the enhancement factor is 1.5 times. Step S2242b: Implement coupling strength gradient diffusion for adjacent nodes in the main transmission path connection area, such as simultaneously increasing the coefficient of surrounding residential areas when strengthening the central industrial area; Step S2242c: Initiate dynamic attenuation for the marked abnormal secondary conduction path: Attenuation trigger condition: when the entropy value fluctuation of the secondary channel exceeds the tolerance range and the duration is greater than the minimum response time, such as fluctuation exceeding 30% for 2 hours; Attenuation intensity calculation: generate an attenuation factor based on the entropy value fluctuation rate and historical stability record, such as applying instantaneous attenuation for sudden fluctuations and implementing phased attenuation for periodic fluctuations.

[0053] Finally, the optimized coupling matrix is ​​output, which includes update parameters to strengthen the main channel connection and suppress the propagation of secondary anomalies.

[0054] Step S2243, parameter feedback optimization: Step S2243a: Correct the calculation rule for the propagation path entropy value based on the coupling strength adjustment result. Entropy weight of main transmission path: The region that enhances the coupling coefficient automatically reduces the time fluctuation sensitivity parameter in its entropy calculation; Secondary conduction path entropy tolerance: The tolerance range for entropy fluctuation of the path with applied attenuation factor synchronously expands. Step S2243b: Optimize channel hierarchy partitioning based on the new coupling matrix: Main transmission path extension: If a continuous high-correlation chain is formed in the enhanced coupling region, it is upgraded to a primary channel; Secondary conduction path degradation: If a persistently suppressed abnormal path fails to recover its entropy, it is downgraded to a temporary conduction path.

[0055] The above process ensures the reliability of the low-entropy path through multi-dimensional verification: temporal correlation fluctuation detection confirms the continuity of conduction; entropy gradient verification of adjacent nodes confirms compliance with atmospheric diffusion laws; monitoring of entropy abrupt changes in secondary channels is linked to changes in meteorological parameters; periodic feature identification distinguishes regular conduction fluctuations; the output channel stability report, main channel verification / secondary anomaly labels provide a basis for coupling adjustment decisions. Coupling strength is adaptively controlled, and the dynamic coupling strength adjustment unit performs differentiated parameter updates: exponential enhancement is implemented for stable main channels, with the entropy deviation determining the enhancement factor, which diffuses to adjacent nodes to form a gradient transition; attenuation is triggered by abnormal secondary channel initiation conditions, with tolerance exceeding limits and continuous timeouts, and the attenuation factor is dynamically generated based on fluctuation characteristics, either suddenly or periodically; the output optimized coupling matrix achieves selective enhancement and suppression of the conduction network. The parameter system undergoes closed-loop iteration, and the parameter feedback optimization unit establishes a dynamic reverse correction mechanism: the entropy calculation of the main path reduces time sensitivity to consolidate the reinforcement effect, and the tolerance range of the secondary path is expanded to buffer abnormal fluctuations; changes in the coupling matrix trigger channel-level reconstruction, strengthening the continuous chain to upgrade to a first-level channel, and continuously suppressing the path to be downgraded to a temporary path; finally, the dynamic field mechanism of closed-loop optimization is output, including the updated coupling matrix, transmission topology, and entropy calculation parameters; realizing the synchronous evolution of transmission topology, entropy rules, and coupling parameters.

[0056] Reliable transmission paths are selected through stability assessment, and network structure is optimized through dynamic adjustment of coupling strength. Finally, a self-correcting field mechanism iterative loop is formed through parameter feedback, thus constructing a dynamic transmission control system with the capabilities of anomaly suppression, main path reinforcement, and hierarchical adaptation.

[0057] Step S225, construct the dynamic influence network between parameters, and configure it as follows: Changes in diffusion rate automatically trigger gradient field step size recalibration; The regional coupling coefficient is adjusted to update the propagation path entropy calculation parameters in reverse. Injecting the optimized parameter system into the physical structure of the conduction channel includes: Gradient vector field embedding: Storing three-dimensional vectors of gradient magnitude and direction at the nodes of the propagation path; Rate weight label: Attach a differentiated rate calculation rule label to each channel level; Coupled matrix topology: Mapping a two-dimensional coupling matrix to the connection properties in a conduction channel topology network.

[0058] The above process achieves dynamic optimization of the field mechanism through a triple guarantee mechanism: entropy-driven coupling coefficient adjustment, low-entropy path exponential enhancement / high-entropy fluctuation-triggered decay, gradient continuity constraints, threshold control of decay difference between adjacent nodes, and automatic recalculation for violations of constraints, ensuring the physical rationality of field parameters in spatiotemporal evolution; the output optimized field parameter system, a smooth gradient field + entropy-controlled coupling matrix + hierarchical velocity field constitutes a dynamic equilibrium system that conforms to the conservation law. The organic field mechanism synthesis submodule establishes a closed-loop feedback network between parameters: velocity change triggers gradient field calibration → coupling coefficient adjustment feeds back into entropy calculation → finally, the gradient vector, velocity weight, and coupling matrix are mapped to the topological attributes of the transmission network, respectively, with three-dimensional vector storage / rule labels / connector attributes; the formed spatiotemporal unified entity realizes the synchronous dynamic update and mutual constraints of four core elements: gradient field, velocity field, coupling matrix, and transmission topology.

[0059] Based on this, step S2, hierarchical parameter mapping, ensures that the initial field structure matches the medium, maintains evolutionary stability through dynamic optimization, and finally synthesizes a multi-physics coupling system with self-consistent feedback capability, providing a computational framework for complex conduction processes that combines spatial resolution accuracy and time-varying stability.

[0060] Step S3: Construct a field succession law model based on the dynamic field dataset, construct an evolutionary baseline library based on the field succession law model, identify the dominant evolutionary path based on the evolutionary baseline library, inject external disturbances along its transmission path, and generate a multi-branch evolutionary scenario tree; wherein, the external disturbances include the intensity of policy regulation and meteorological anomaly index.

[0061] This step uses the output of the dynamic field mechanism in step S2 as initial conditions to establish a prediction framework based on the field succession law, identify the dominant evolutionary path, and, combined with a parameterized perturbation injection mechanism, generate an evolutionary scenario tree that includes the influence of multiple factors such as policy regulation and meteorological anomalies. Specifically, it includes: Step S31: Construct a field succession model including an evolutionary baseline library and a memory unit set. Specifically, this includes: Step S311: Extract the historical change pattern of gradient vectors from the primary channel to form the main path evolution baseline; In practice, the original gradient sequence of the primary channel is decomposed into fluctuation characteristics, the diurnal periodic component of the industrial area intensity fluctuation baseline is removed, the confidence domain of the dominant diffusion direction in the stable direction angle interval is extracted, and the periodic component and direction constraint are fused to form the main path evolution baseline. Step S312: Extract intermittent active features from the secondary channels to form the secondary path evolution baseline; In practice, during the active period of the secondary channel, the peak pulse intensity and traffic peak diffusion enhancement threshold are captured, the pulse interval pattern and intermittent active cycle are statistically analyzed, and the pulse characteristics and periodicity are integrated to form the secondary path evolution baseline. Step S313: Store the intensity decay features of key nodes in the intensity memory chain according to the time series. In practice, the extreme points of intensity fluctuations in the main path evolution baseline are automatically marked as key nodes of the gradient field; the decay characteristics, the valley values ​​of intensity fluctuations during the day and night, and the decay rate of the same period on the next day are stored according to the nodes. Step S314: Store the response pattern of the diffusion rate corresponding to the meteorological scene into the rate memory ring.

[0062] In practice, the pulse events in the secondary path evolution baseline are associated with meteorological scene IDs; Step S315: Store the mapping relationship between pulse intensity and meteorological scene into the evolution baseline library of the field succession law model.

[0063] In some possible implementations, the propagation path topology decomposition step is performed before generating the path evolution baseline, as follows: Step S30: Based on the regional coupling coefficient, filter strongly correlated paths with spatial continuity not less than a set threshold; extract gradient vectors containing intensity and direction angle along the path, and divide them into original gradient sequences according to time windows; identify weakly correlated paths with regional coupling coefficients lower than the set threshold but with pulse-like transitions, and mark the active periods of the channels with fixed timestamps (such as the active periods of the channels with fixed timestamps corresponding to traffic peaks); output the set of primary and secondary channels with spatial topology labels.

[0064] By employing spatial topological decomposition, gradient pattern refinement, and baseline validation and storage, a classification system for primary / secondary channel transmission paths based on historical observation data is constructed, along with the dynamic evolution patterns of their intensity fluctuations, pulse periods, and meteorological correlations. The output baseline library provides quantifiable and traceable benchmark data support for subsequent steady-state / variant path identification, path selection, and disturbance simulation.

[0065] The above process binds and stores dynamic field mechanism parameters with spatiotemporal evolution characteristics. The intensity memory chain stores the intensity decay characteristics of key nodes in the gradient field in a time series, and the rate memory loop records the response patterns of the diffusion rate field under different meteorological conditions. The output includes a field succession law model comprising an evolutionary baseline library and a set of memory units. In specific implementation, the main path evolutionary baseline extracts the historical change patterns of the gradient vector of the primary conduction channel, such as the diurnal fluctuation pattern of emission intensity in industrial areas; the secondary path evolutionary baseline records the intermittent activity characteristics of the secondary conduction channel, such as the pulsed diffusion enhancement during peak traffic hours. Among these, key nodes include factory clusters, and meteorological scenarios include stable weather and strong winds.

[0066] Step S32: Identify steady-state paths and variant paths by comparing the deviation of the real-time field strength gradient from the historical evolution baseline. Correct non-physical evolution paths through intensity inertia constraints and topological inertia verification, and output a set of dominant evolutionary paths, including steady-state paths, variant paths, and inertia correction records.

[0067] Specifically: Steady-state path screening: Select transmission paths whose gradient change trend matches the historical baseline better than a set threshold, such as paths that conform to the daily cycle of industrial emissions; their intensity decay characteristics are directly verified by calling historical decay slopes from the intensity memory chain. Path mutation detection: Captures main transmission paths with abrupt changes in gradient vector direction or abnormal decay rates, such as path intensity reversal caused by sudden pollution events; its abnormal rate values ​​are verified for rationality by matching meteorological scene IDs with rate memory loops. Intensity inertia constraint: Using the historical maximum fluctuation threshold stored in the intensity memory chain, the gradient change amplitude of adjacent time slices is limited; the gradient change amplitude of adjacent time slices is limited to not exceeding the historical maximum fluctuation threshold. Topology inertia verification: Based on the physical connectivity rules of the secondary transmission path evolution baseline, the generation of non-connected paths is prohibited; ensure the physical connectivity of newly identified paths with the transmission network topology, such as prohibiting the generation of paths that cross mountains without diffusion channels.

[0068] Step S33: Implement multi-dimensional perturbation simulation and generate an evolution scenario tree: convert the intensity of emission reduction policies into gradient field decay correction coefficients, superimpose asymptotic perturbations on the steady-state path, and call the historical decay slope of the corresponding node in the intensity memory chain as a reference; map the extreme weather index into diffusion rate perturbation weights, inject instantaneous perturbations on the variation path, and call the response mode of the same weather in the rate memory loop as a reference.

[0069] This step generates a multi-factor scenario tree, quantifying external factors into field mechanism correction parameters; policy regulation injection: converting the intensity of emission reduction policies into gradient field decay rate correction coefficients, such as the intensity gradient multiplication factor of industrial zones corresponding to production restriction orders; meteorological anomaly injection: mapping extreme weather indices into diffusion rate field disturbance weights, such as stable weather triggering global decay of the rate field; implementing multi-dimensional disturbance simulation along the active programming path, with steady-state scenario branches: superimposing progressive disturbances on the steady-state path, such as the gradually increasing impact of carbon tax policies; abrupt scenario branches: injecting sudden disturbances into the abrupt path, such as the instantaneous change in diffusion patterns caused by typhoon passage; outputting an evolutionary scenario tree, including multi-level disturbance branches, path impact intensity labels, and spatiotemporal evolution trajectories, where the steady-state scenario branches are based on the progressive correction of decay characteristics based on the intensity memory chain; and the abrupt scenario branches are based on the instantaneous reconstruction of response patterns based on the rate memory loop.

[0070] The above process extracts spatiotemporal features such as the diurnal fluctuations of the main / secondary path gradient vectors and dynamic field parameters (e.g., traffic peak pulses) to construct a field dynamic law knowledge base with dual baselines for the main / secondary paths and dual memories of the intensity chain / rate loop. This forms a spatiotemporal benchmark framework for modeling, whose output provides a historical model reference system for subsequent path selection and disturbance simulation. Based on multi-dimensional matching of real-time gradient fields with evolutionary baseline trends / decay / meteorological responses, it achieves: inertial constraint verification of steady-state paths and retrospective verification of outlier meteorological scenarios for mutated paths; the output path set also retains historical compliance records, ensuring that the evolution process conforms to physical laws and historical experience. The path evolution scenario tree generation transforms external intervention variables such as emission reduction policies and extreme weather into two types of disturbance mechanisms: gradual disturbance: simulating the cumulative impact of policies on steady-state paths through the attenuation slope benchmark value of the intensity memory chain; and instantaneous disturbance: reconstructing path mutations under extreme conditions through the meteorological response model of the rate memory loop. The generated scenario tree includes spatiotemporal trajectory labels, enabling path evolution prediction under multi-level disturbances.

[0071] By constraining real-time path selection through historical patterns and realizing the quantitative mapping of disturbance effects based on memory chains / loops, the final output is a field evolution prediction system that combines physical rationality and policy responsiveness, supporting dynamic simulation and intervention assessment of transmission paths in complex environmental systems.

[0072] Based on the above technical solution, this application first establishes a dynamic carbon emission source-stream tracking program based on multimodal sensor data, transforming the instantaneous emissions from discrete monitoring points into a regional continuous flow field distribution. From this regional continuous flow field distribution, fluctuation characteristics of different emission types, such as industrial, transportation, and biological sources, are separated to form a source-stream feature vector matrix with timestamps. Secondly, a three-dimensional spatiotemporal convolution kernel group is constructed based on the source-stream feature vector matrix. Through a multi-scale field matching program, the discrete source-stream features of the source-stream feature vector matrix are reconstructed into a continuous spatiotemporal emission field. During the reconstruction process, potential channels for emission transmission between regions are automatically identified, forming a dynamic field mechanism including emission intensity, diffusion rate, and regional coupling coefficient. Finally, using the dynamic field mechanism as initial conditions, a prediction framework based on field succession laws is established, identifying the dominant evolutionary path. Combined with a parameterized perturbation injection mechanism, an evolutionary scenario tree influencing multiple factors, including policy regulation and meteorological anomalies, is generated. This solution, through the synergistic effect of dynamic analysis of multimodal sensor data, intelligent spatiotemporal field reconstruction, and adaptive evolutionary deduction, constructs a regional carbon emission prediction system that integrates physical mechanisms and data-driven approaches. The system possesses comprehensive emission characterization capabilities. Through cascaded processing of dynamic source-flow analysis and spatiotemporal field reconstruction, it achieves end-to-end data fusion encompassing point monitoring, flow field distribution, multi-source characteristics, and transmission mechanisms, overcoming the contradiction between the discreteness of monitoring data and the continuity of the emission process in traditional methods. The collaborative computation of three-dimensional spatiotemporal convolution kernel groups and multi-scale field matching procedures enables the system to simultaneously analyze spatial heterogeneity and temporal dynamics, forming a physically interpretable continuous emission field characterization. The cross-scale coupled prediction mechanism, using a dynamic field mechanism as the initial condition for the prediction framework, organically integrates physical parameters such as emission intensity, diffusion rate, and regional coupling coefficient. It establishes a mathematical mapping relationship between microscopic emission behavior and macroscopic evolutionary trends through field succession laws. The parameterized perturbation injection mechanism, while maintaining core physical constraints, achieves systematic coupling between social factors such as policy regulation and natural factors such as meteorological conditions. The adaptive inference architecture constructs a predictive backbone based on field dynamics through authoritative path identification, and forms an evolutionary scenario tree with branching structure by combining multi-factor perturbation injection. This architecture maintains the coherence of the core physical processes and can flexibly respond to the influence of external intervention variables, so that the prediction results have both physical rationality and policy guidance value.

[0073] In summary, this embodiment establishes a closed-loop prediction paradigm of "data assimilation-field reconstruction-evolutionary deduction". Through feature learning under physical constraints and multi-factor coupled modeling, it realizes the technological leap from real-time monitoring to advanced prediction of regional carbon emissions. The dynamic field mechanism and evolutionary scenario tree output by the system can directly serve the decision support needs for precise carbon emission control.

[0074] See Figure 3Corresponding to the above-described embodiments of regional carbon emission prediction methods based on carbon emission monitoring, this application provides a regional carbon emission prediction system based on carbon emission monitoring, comprising: The dynamic source flow analysis module 1001 is used to collect regional discrete point emission data through a multimodal sensor network and generate a regional continuous flow field; based on orthogonal constraints, it performs feature decoupling and quantization on different emission source types to generate a source flow feature vector matrix with three-dimensional geocoding and dual timestamps. The spatiotemporal field reconstruction module 1002 is used to reconstruct a continuous spatiotemporal emission field based on the source flow feature vector matrix, identify the conduction path, dynamically bind the field features along the path, and generate a dynamic field dataset; wherein, the field features include emission intensity, diffusion rate, and regional coupling coefficient. The adaptive evolutionary deduction module 1003 is used to construct a field succession law model based on the dynamic field dataset, construct a historical evolutionary baseline library based on the field succession law model, identify the dominant evolutionary path based on the historical evolutionary baseline library, inject external disturbances along its transmission path, and generate a multi-branch evolutionary scenario tree; wherein, the external disturbances include the intensity of policy regulation and meteorological anomaly index.

[0075] The aforementioned regional carbon emission prediction system based on carbon emission monitoring implements the steps and processes of the aforementioned regional carbon emission prediction method based on carbon emission monitoring, and achieves the same technical effect. To avoid repetition, these steps will not be repeated here.

[0076] For illustrative implementation examples, see [link to specific implementation details]. Figure 4 Corresponding to the above-described embodiments of regional carbon emission prediction methods based on carbon emission monitoring, this application provides a regional carbon emission prediction device based on carbon emission monitoring, comprising: The multimodal sensor array unit includes a distributed gas concentration sensor group with a spatially discrete layout, containing multi-parameter probes such as CO2 and CH4; a dynamic flow monitor equipped with a Doppler flow velocity sensor and a thermal radiation spectrometer; an environmental parameter acquisition module integrating temperature, humidity, wind speed and direction, and vegetation index sensors; and a star topology connection established with the dynamic source and flow preprocessing controller via a wireless sensor network, with real-time data streams transmitted to edge computing nodes through an encrypted channel. The dynamic source-stream preprocessing controller includes a non-uniform grid generator, a spatial density adaptive adjustment circuit, a time-frequency feature decoupling processor, a dual-channel arithmetic unit: a time-domain pulse capture module and a frequency-domain steady-state separation module, a feature vector encapsulation unit, and a matrix generation circuit with timestamp encoding; it receives the raw data stream from the sensor array and outputs it to the spatiotemporal field reconstruction arithmetic unit; it establishes a bidirectional data channel with the dynamic field storage repeater for historical feature comparison. The spatiotemporal field reconstruction operator includes a 3D convolutional kernel group generation device, containing a spatially nested kernel building unit and a time-adaptive kernel regulator; a multi-scale field fusion processor with dual channels for micro / macro-scale computation, including a correlation calculation array; and a dynamic field mechanism synthesizer, in which a gradient-rate-coupling parameter binding circuit receives the feature matrix of the preprocessing controller, outputs dynamic field parameters to the evolutionary inference master control unit, and is directly connected to the transmission topology memory to store / recall transmission path entropy data. The evolutionary simulation main control unit includes a field succession law modeling module, which contains a field strength baseline tracker and a memory unit storage array; a path weight migration analyzer; a real-time field and historical baseline comparison and calculation unit; a multi-factor perturbation injection interface; a policy parameter converter and a meteorological anomaly mapper; it receives dynamic field parameters from the spatiotemporal field calculator and outputs them to the scenario tree generation device; it connects to the policy control input terminal through an external policy bus and establishes a real-time data stream channel with the meteorological data gateway. The auxiliary function device includes a dynamic field storage repeater, which contains gradient field / rate field / coupling matrix partitioned storage units. It forms a triangular data path with the preprocessing controller, reconstruction arithmetic unit, and main control unit, and transmits topology memory. It is a three-dimensional channel network dedicated storage array and is connected to the reconstruction arithmetic unit and main control unit through a high-speed data bus. The human-computer interaction terminal includes a scenario tree visualization engine and a parameter control interface. It interacts with the main control unit through a control bus and receives evolution scenario data.

[0077] The aforementioned regional carbon emission prediction device based on carbon emission monitoring implements the steps and processes of the aforementioned regional carbon emission prediction method based on carbon emission monitoring, and achieves the same technical effect. To avoid repetition, these will not be repeated here.

[0078] See Figure 5 Corresponding to the above-described embodiments of the regional carbon emission prediction method based on carbon emission monitoring, this application provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps and processes of the above-described embodiments of the regional carbon emission prediction method based on carbon emission monitoring, and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0079] The memory 1009 can be used to store software programs and various data. The memory 1009 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback function, image playback function, etc.). Furthermore, the memory 1009 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 1009 in the embodiments of this application includes, but is not limited to, these and any other suitable types of memory.

[0080] The processor 1010 may include one or more processing units; optionally, the processor 1010 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into the processor 1010.

[0081] Corresponding to the above-described embodiments of the regional carbon emission prediction method based on carbon emission monitoring, this application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the steps and processes of the above-described embodiments of the regional carbon emission prediction method based on carbon emission monitoring, and achieve the same technical effect. To avoid repetition, these will not be described again here.

[0082] The processor is the processor in the electronic device described in the above embodiments of this application. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk.

[0083] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0084] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0085] It is understood that the embodiments of this application have been described above in conjunction with the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. As those skilled in the art will know, various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, those skilled in the art, under the guidance or instruction of this application, can modify these features and embodiments to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, this invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of this invention.

Claims

1. A regional carbon emission prediction method based on carbon emission monitoring, characterized in that, include: Regional continuous flow field is generated by collecting emission data from discrete points in the region through a multimodal sensor network. Based on orthogonal constraints, feature decoupling and quantification are performed on different emission source types to generate source flow feature vector matrices with three-dimensional geocoding and dual timestamps; Based on the source-flow feature vector matrix, a continuous spatiotemporal emission field is reconstructed, the conduction path is identified, and the field features are dynamically bound along the path to generate a dynamic field dataset; wherein, the field features include emission intensity, diffusion rate, and regional coupling coefficient. Based on the dynamic field dataset, a field succession law model is constructed. Based on the field succession law model, an evolutionary baseline library is constructed. Based on the evolutionary baseline library, the dominant evolutionary path is identified, and external disturbances are injected along its transmission path to generate a multi-branch evolutionary scenario tree. The external disturbances include the intensity of policy regulation and meteorological anomaly index.

2. The regional carbon emission prediction method based on carbon emission monitoring according to claim 1, characterized in that, The process involves collecting emission data from discrete points in the region using a multimodal sensor network to generate a regional continuous flow field. Based on orthogonal constraints, feature decoupling and quantification are performed on different emission source types to generate a source flow feature vector matrix with three-dimensional geocoding and dual timestamps, including: Obtain the regional continuous flow field distribution; wherein, the regional continuous flow field distribution includes spatiotemporally continuous emission intensity, diffusion direction and emission waveform characteristics; Spatiotemporal fluctuation pattern matching is performed based on the emission waveform in the flow field to construct feature dimensions including steady-state fluctuation axis, pulse fluctuation axis and environmental coupling axis; Feature dimension unwrapping is performed based on orthogonal constraints between feature dimension axes. Dynamic feature vectorization is performed for each feature dimension: for the steady-state fluctuation axis, the amplitude envelope and frequency stability index of the steady-state fluctuation are obtained; for the pulse fluctuation axis, the peak density and spatial propagation attenuation rate of the pulse waveform are statistically analyzed; for the environmental coupling axis, the time-delay correlation coefficient between low-frequency fluctuations and environmental parameters is evaluated. The quantification results are dynamically correlated with the spatiotemporal coordinates of the regional continuous flow field. A three-dimensional geocoding of longitude, latitude, and elevation layers is added to each spatial unit. A dual timestamp of absolute timestamp and emission cycle phase is embedded in each time slice, and the source flow feature vector matrix is ​​output.

3. The regional carbon emission prediction method based on carbon emission monitoring according to claim 1, characterized in that, The process involves reconstructing a continuous spatiotemporal emission field based on the source-flow feature vector matrix, identifying the transmission path, dynamically binding the field features along the path, and generating a dynamic field dataset, including: A three-dimensional spatiotemporal convolution kernel group is constructed based on the source-flow feature vector matrix. The source-flow feature vector matrix is ​​scanned through the three-dimensional spatiotemporal convolution kernel group to generate a field correlation map. Based on the field correlation degree distribution of the field correlation map, potential transmission channels are identified, and the paths of the transmission channels are dynamically bound to field characteristics including emission intensity gradient, diffusion rate and regional coupling coefficient. Wherein, the emission intensity gradient is the intensity decay rate calculated along the conduction path; The diffusion rate is based on the atmospheric diffusion efficiency derived from spatial correlation and time delay. The regional coupling coefficient is a measure of the intensity of mutual influence between regions by using the entropy value of the transmission path.

4. The regional carbon emission prediction method based on carbon emission monitoring according to claim 3, characterized in that, The step of constructing a three-dimensional spatiotemporal convolutional kernel group based on the source-flow feature vector matrix, and scanning the source-flow feature vector matrix through the three-dimensional spatiotemporal convolutional kernel group to generate a field correlation map includes: The source-flow feature vector matrix is ​​analyzed, and based on its spatial resolution and temporal density dynamics, a convolutional kernel structure containing spatial kernels and temporal kernels is generated. A multi-scale spatiotemporal coverage range is defined for each convolutional kernel; thus, the three-dimensional spatiotemporal convolutional kernel group containing spatial nested kernels, temporally adaptive kernels, and multi-scale receptive field parameters is obtained. The generated field correlation map includes spatial correlation and temporal correlation.

5. The regional carbon emission prediction method based on carbon emission monitoring according to claim 3, characterized in that, The identification of potential conduction channels based on the field correlation distribution of the field correlation map, and the dynamic binding of the path of the conduction channel with field characteristics including emission intensity gradient, diffusion rate, and regional coupling coefficient, includes: Construct a conduction channel topology network that includes channel levels, spatial coordinate chains, and entropy labels; Based on the conduction channel topology network, the quantization field characteristic outputs a dynamic quantization parameter set, including: emission intensity gradient field: calculating the attenuation rate per unit distance through the intensity difference between adjacent nodes; diffusion rate field: combining spatial correlation, waveform similarity, and time delay, and using response lag to infer atmospheric diffusion efficiency; regional coupling matrix, including direct coupling coefficient and indirect coupling coefficient; direct coupling coefficient: an assessment value of the influence intensity between adjacent regions connected by the main conduction path; indirect coupling coefficient: an assessment value of the cross-regional cascade effect formed through secondary conduction paths. The dynamic quantization parameter set is dynamically associated with the physical characteristics of the conduction channel; Dynamically adjust the region coupling matrix based on the entropy value of the transmission path; A dynamic influence network among parameters is constructed, configured as follows: changes in diffusion rate automatically trigger gradient field step size recalibration; regional coupling coefficient adjustment updates the propagation path entropy calculation parameters in reverse; the optimized parameter system is injected into the physical structure of the propagation channel, including: Gradient vector field embedding: Storing three-dimensional vectors of gradient magnitude and direction at the nodes of the propagation path; Rate weight label: Attach a differentiated rate calculation rule label to each channel level; Coupled matrix topology: Mapping a two-dimensional coupling matrix to the connection properties in a conduction channel topology network.

6. The regional carbon emission prediction method based on carbon emission monitoring according to claim 5, characterized in that, The construction of the conduction channel topology network, including channel levels, spatial coordinate chains, and entropy labels, includes: The outputs of the multi-scale receptive fields are hierarchically fused to obtain a preliminary continuous spatiotemporal emission field containing intensity distribution, spatial correlation degree and time delay parameters; The primary transmission path is the adjacent region chain whose spatial correlation consistently exceeds a set dynamic threshold and whose temporal correlation fluctuation is below a set tolerance range. Secondary transmission paths are identified by capturing abnormal paths where spatial correlation suddenly increases but temporal correlation exhibits periodic oscillations. Based on the preliminary continuous spatiotemporal emission field, a transmission path entropy index is constructed based on the field correlation distribution. Channel levels are classified based on the entropy value of the conduction path: Configure primary conduction paths with entropy values ​​below a set threshold as primary channels; configure secondary conduction paths with entropy values ​​within a set threshold range as secondary channels.

7. The regional carbon emission prediction method based on carbon emission monitoring according to claim 5, characterized in that, The dynamic association of the dynamic quantization parameter set with the physical characteristics of the conduction channel includes: The emission intensity gradient vector is dynamically matched with the direction of the transmission channel, and the angle between the spatial direction of the emission intensity gradient vector and the extension direction of the transmission path is less than a set angle threshold. The gradient is automatically adjusted to obtain the step size based on the conduction channel level. A large step size is used for the first-level channel and a fine step size is used for the second-level channel. The calculation rule for the dynamic allocation rate based on the conduction channel level is as follows: the first-level channel uses the global rate, and the second-level channel uses the local correction rate.

8. The regional carbon emission prediction method based on carbon emission monitoring according to claim 1, characterized in that, The process of constructing a field succession law model based on the dynamic field dataset, constructing an evolutionary baseline library based on the field succession law model, identifying the dominant evolutionary path based on the evolutionary baseline library, injecting external perturbations along its propagation path, and generating a multi-branch evolutionary scenario tree includes: A field succession law model is constructed, including an evolutionary baseline library and a memory unit set. This includes: extracting historical change patterns of gradient vectors from primary channels to form the main path evolutionary baseline; extracting intermittent active features from secondary channels to form secondary path evolutionary baselines; storing the intensity decay features of key nodes in an intensity memory chain according to time series; storing the response patterns of diffusion rates corresponding to meteorological scenarios in a rate memory loop; and storing the mapping relationship between pulse intensity and meteorological scenarios in the evolutionary baseline library. By comparing the deviation of the real-time field strength gradient from the historical evolution baseline, steady-state paths and variant paths are identified. Non-physical evolution paths are corrected through intensity inertia constraints and topological inertia verification, and a set of dominant evolutionary paths is output, including steady-state paths, variant paths and inertia correction records. Implement multi-dimensional perturbation simulation to generate an evolution scenario tree: convert the intensity of emission reduction policies into gradient field decay correction coefficients, superimpose asymptotic perturbations on the steady-state path, and call the historical decay slope of the corresponding node in the intensity memory chain as a reference; map the extreme weather index into diffusion rate perturbation weights, inject instantaneous perturbations on the variation path, and call the response patterns of the same weather in the rate memory loop as a reference.

9. A regional carbon emission prediction system based on carbon emission monitoring, characterized in that, include: The dynamic source flow analysis module is used to collect regional discrete point emission data through a multimodal sensor network and generate a regional continuous flow field; Based on orthogonal constraints, feature decoupling and quantification are performed on different emission source types to generate source flow feature vector matrices with three-dimensional geocoding and dual timestamps; The spatiotemporal field reconstruction module is used to reconstruct a continuous spatiotemporal emission field based on the source-flow feature vector matrix, identify the conduction path, dynamically bind the field features along the path, and generate a dynamic field dataset; wherein, the field features include emission intensity, diffusion rate, and regional coupling coefficient; The adaptive evolutionary deduction module is used to construct a field succession law model based on the dynamic field dataset, construct an evolutionary baseline library based on the field succession law model, identify the dominant evolutionary path based on the evolutionary baseline library, inject external disturbances along its transmission path, and generate a multi-branch evolutionary scenario tree; wherein, the external disturbances include the intensity of policy regulation and meteorological anomaly index.

10. An electronic device, characterized in that, The electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the regional carbon emission prediction method based on carbon emission monitoring as described in any one of claims 1 to 6.

Images