Highway carbon emission dynamic monitoring method and system based on multi-source data fusion

Abstract

The invention discloses a highway carbon emission dynamic monitoring method and system based on multi-source data fusion. The method comprises the following steps: S1, monitoring carbon emission data in a construction period; the method comprises the following steps: step S101, setting a monitoring target; step S102, making a carbon emission monitoring plan; s103, carbon emission in the construction period is monitored in real time; step S104, establishing a data collection system; s2, collecting and monitoring carbon emission data in an operation and maintenance period; wherein in the step S201, information of running vehicles is monitored in real time, monitoring data is associated with geographic information of a highway in combination with a geographic information system, and monitoring of traffic flow carbon emission in different geographic areas is achieved; and S202, docking with the target highway toll station data to realize all-directional monitoring of the traffic flow carbon emission attached data. The carbon emission in the construction period is counted, and the carbon emission in the operation and maintenance period is subjected to multi-source monitoring on the toll intersection in the driving process, so that the carbon emission monitoring is more comprehensive.

Description

Technical Field

[0001] The present invention relates to the technical field of highway carbon emissions, and in particular to a method and system for dynamic monitoring of highway carbon emissions based on multi-source data fusion. Background Art

[0002] During economic development, humans have released large amounts of greenhouse gases (GHGs) into the atmosphere, contributing to global warming. Climate change and the continued emission of greenhouse gases (GHGs) are widely recognized as one of the major environmental issues of the current decade. Countries around the world have responded to these environmental challenges. Transportation, construction, and industry are the three major sources of CO2 emissions. Among these, CO2 emissions from the transportation sector account for 27% of total global CO2 emissions related to energy activities. The transportation sector is a significant contributor to climate change, accounting for 24% of global carbon emissions. Within the transportation sector, road transport is the absolute majority of transportation CO2 emissions and a key focus for emission reduction. Analyzing the development of the transportation industry, my country's transportation infrastructure density, per capita infrastructure, and per capita vehicle ownership still lag significantly behind other countries, and are expected to increase significantly in the future. Road transport, in particular, will remain the primary source of carbon emissions in my country's transportation sector and will continue to strongly drive its growth. Controlling and reducing road transport carbon emissions is a key approach to reducing transportation carbon emissions and achieving global emission reduction commitments. Quantifying and evaluating the environmental impacts of highways is crucial in the context of developing green and low-carbon transportation projects.

[0003] Numerous factors are involved in the life cycle of a highway. Influenced by scientific and technological advancements, technological innovations, and policy incentives, the vehicle composition, fuel types, and sources of vehicles operating during a highway's life cycle are complex and diverse. New materials, technologies, processes, and equipment are constantly emerging during construction and maintenance. Given the coexistence of multiple technologies, it is crucial to comprehensively consider all aspects and elements to accurately, comprehensively, and comprehensively assess the global low-carbon effects of a particular technology or policy. This requires that carbon emission monitoring, accounting, and management throughout the life cycle of highways must achieve digitalization and intelligence while meeting the precision requirements of complex systems. Therefore, a dynamic highway carbon emission monitoring method and system based on multi-source data fusion is urgently needed. Summary of the Invention

[0004] The technical problem solved by the present invention is how to monitor carbon emissions over the entire life cycle of a highway. To solve the above technical problem, the technical solution adopted by the present invention is:

[0005] A method for dynamic monitoring of highway carbon emissions based on multi-source data fusion includes the following steps:

[0006] Step S1: Monitoring carbon emission data during the construction period; wherein:

[0007] Step S101, setting monitoring targets;

[0008] Step S102: Formulate a carbon emission monitoring plan;

[0009] Step S103: Real-time monitoring of carbon emissions during the construction period;

[0010] Step S104: establishing a data collection system;

[0011] Step S2: Collect and monitor carbon emission data during the operation and maintenance period;

[0012] Step S201: Real-time monitoring of vehicle information is performed. In combination with a geographic information system, the monitoring data is associated with geographic information of highways to monitor carbon emissions from traffic flows in different geographic areas.

[0013] Step S202: Connect with the target highway toll station data to fully grasp the traffic flow characteristics of the highway driving section and achieve all-round monitoring of the traffic flow carbon emission data.

[0014] In some embodiments, step S2 further includes step S203 , modeling and simulating highway traffic flow, and estimating and predicting the carbon emissions of the traffic flow.

[0015] In some embodiments, in step S201 , the information of the moving vehicle includes speed and vehicle type.

[0016] In some embodiments, the total carbon emissions of the vehicle are calculated using the monitoring information in step S201, where CO2 emissions (g / km) = basic fuel consumption rate (L / 100km)*[1+a+b*(|actual speed-economic speed|)+c*slope percentage]*d / 100;

[0017] a is the vehicle weight correction factor, b is the speed correction factor, c is the slope correction factor; d is the fuel emission factor;

[0018] Total vehicle carbon emissions (g) = CO2 emissions (g / km) * driving distance (km).

[0019] In some embodiments, the vehicle weight correction coefficient of sedans / SUVs is 0.4, the vehicle weight correction coefficient of trucks / buses is 0.55, the speed correction coefficient of sedans / SUVs is 0.008, the speed correction coefficient of trucks / buses is 0.012, the slope correction coefficient is 0.1, the fuel emission factor of gasoline is 2320gCO2 / L, and the fuel emission factor of diesel is 2660gCO2 / L.

[0020] In some embodiments, in step S202 , the traffic flow characteristics of the highway driving section include the number of vehicles, vehicle types, and mileage.

[0021] In some embodiments, different unit-mileage carbon emission factors are matched according to vehicle types, and the unit-mileage carbon emission factors of each vehicle are multiplied by the mileage to obtain the carbon emissions of each vehicle, and the total vehicle carbon emissions are obtained based on the number of vehicles.

[0022] In some embodiments, step S203 includes: a graph structure construction layer: abstracting the traffic road network into a graph structure G = (V, E, A);

[0023] V: represents the key point in the transportation network;

[0024] E: represents the connection relationship between nodes;

[0025] A: Encodes the spatial relationship and connection strength between nodes;

[0026] Input: geographic information data, road network topology data, sensor location data;

[0027] Output: The defined graph structure G serves as the basis for subsequent graph convolution operations;

[0028] Spatiotemporal graph convolution module: Simultaneously captures the spatial dependencies and temporal dynamics of traffic carbon emission data;

[0029] Structure: Spatial graph convolution layer: operates on the adjacency matrix A of the graph and aggregates the carbon emission-related feature information of neighboring nodes;

[0030] Gated temporal convolution layer: At each node, a one-dimensional convolution is applied along the time axis; using a gating mechanism or simple convolution + nonlinear activation, the time evolution pattern of the node itself is learned;

[0031] Stacking: Multiple ST-GCN modules can be stacked to learn deeper spatiotemporal feature representations; each layer operates on the updated node features;

[0032] Input: graph structure G and the carbon emission-related feature vector of each node at the historical time step;

[0033] Output: The high-level feature representation of each node learned at the latest time step, which integrates spatiotemporal information;

[0034] Node-level feature fusion and pooling layer: If you need to integrate information from different ST-GCN layers or other sources, you can perform fusion here;

[0035] Pooling: If the final prediction target is regional or global carbon emissions, it is necessary to aggregate the features of all nodes or key node groups. Common methods include global average pooling, global maximum pooling, or attention-weighted pooling.

[0036] Output: The high-level feature representation of all nodes output by the last ST-GCN module;

[0037] Output: A fused global / regional feature vector, or still a feature vector for each node;

[0038] Fully connected regression layer: Receives fused / pooled feature vectors, learns the complex nonlinear mapping relationship between features and target carbon emissions, and completes the final regression prediction;

[0039] Input: feature vector from the previous layer;

[0040] Output: predicted future carbon emissions;

[0041] Output layer: Depending on the task requirements, it may include a linear layer or directly use the output of the fully connected layer; calculate the final carbon emissions prediction value;

[0042] Output: Final carbon emissions forecast.

[0043] A dynamic monitoring system for carbon emissions of highways based on multi-source data fusion, using any of the above methods, comprising a construction period monitoring module and a construction period statistics module, wherein the construction period statistics module performs statistics based on the data provided by the construction period detection module;

[0044] It also includes an operation and maintenance period monitoring module, a data docking module, a calculation module and a display module; the operation and maintenance period monitoring module is used to monitor the information of traveling vehicles; the data docking module is used to dock with the data of the target highway toll station; the calculation module is used to calculate the data provided by the operation and maintenance period monitoring module and the data docking module; the display module is used to display the calculation results of the calculation module and the statistical results of the construction period statistics module.

[0045] In some embodiments, a prediction module is further included, and the display module also displays the prediction results of the prediction module.

[0046] The beneficial effects of adopting the above technical solution are: statistics on carbon emissions during the construction period, and multi-source monitoring of carbon emissions during the operation and maintenance period during driving and at toll intersections, making carbon emission monitoring more comprehensive and providing factual basis for timely adjustments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] Figure 1 This is an overall flow chart of a method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to the present invention;

[0049] Figure 2 This is a flow chart of S1 of a method for dynamic monitoring of carbon emissions of highways based on multi-source data fusion according to the present invention;

[0050] Figure 3 This is a flow chart S2 of a method for dynamic monitoring of carbon emissions of highways based on multi-source data fusion according to the present invention;

[0051] Figure 4 This is a flow chart of step S203 in a method for dynamic monitoring of carbon emissions from highways based on multi-source data fusion according to the present invention. DETAILED DESCRIPTION

[0052] The following is a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts are within the scope of protection of the present invention.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] like Figure 1-4 As shown, a dynamic monitoring method for highway carbon emissions based on multi-source data fusion includes the following steps:

[0055] Step S1: Monitoring carbon emission data during the construction period; wherein:

[0056] Step S101, setting monitoring targets;

[0057] Step S102: Formulate a carbon emission monitoring plan;

[0058] Step S103: Real-time monitoring of carbon emissions during the construction period;

[0059] Step S104: establishing a data collection system;

[0060] The carbon emission factor database for the highway construction phase includes material production, transportation, and construction. The life cycle assessment method is used to calculate the carbon emission factors required for each construction phase through the mining and transportation of building materials, the energy consumption of construction machinery, etc. For material carbon emission calculations, the building material carbon emission factor refers to the sum of the carbon emissions generated during the production process of building materials with unit mass or unit volume and the carbon emissions generated through chemical reactions. It is an important parameter to characterize the carbon emissions of a certain building material and an important data for calculating the carbon footprint of building materials. The determination of the building material carbon emission factor mainly comes from three aspects: carbon emissions from energy consumption, carbon emissions from chemical reactions of materials, and the recovery coefficient of building materials. The measurement process is as follows: the first step is to estimate the carbon dioxide emissions of building material products using the carbon content, energy usage, and carbon content of the original building materials; the second step is to consider the recyclability of building materials and the material recovery coefficient based on the full life cycle theory. The carbon dioxide emissions of building materials are calculated using the table of renewable material recovery coefficients. Due to the wide variety of building materials and the huge differences in the production processes of building materials, the carbon emission factors of the main building materials are calculated and sorted. See the table below.

[0061] Material Type <![CDATA[Building material carbon emission factor (kgCO2e / t)]]> Straight-run asphalt 586.52 polymer-modified asphalt 895.05 Emulsified asphalt 636.49 Road petroleum asphalt 439.80 SBS modified asphalt 613 Rubber modified asphalt 220 Waterproof bitumen (average) 480 Liquid petroleum asphalt 390 coal tar 270 Modified emulsified asphalt 285.83 Emulsified asphalt (65% asphalt content) 246.62 Emulsified asphalt (60% asphalt content) 228.76 Emulsified asphalt (50% asphalt content) 193.04 Ordinary Portland cement (market average) 735 C30 concrete 295

[0062] Mechanical equipment used during the construction period refers to the mechanical equipment used during highway construction, such as excavators, lifting machinery, and aerial work machinery. The carbon emission factor of a construction machinery shift is calculated based on the resources and energy consumed per construction machinery shift. The energy consumed per shift and the carbon emission factor of energy are found through the standard. The carbon emission factor of each piece of mechanical equipment is obtained by multiplying the energy consumed per shift by the carbon emission factor of energy consumed. The energy consumption per shift of commonly used construction machinery on major highways is calculated and sorted. The following table is shown:

[0063] Machine name Performance specifications Energy consumption (diesel kg) crawler bulldozers Power 60KW 40.86 crawler bulldozers Power 90KW 65.37 crawler bulldozers Power 105KW 76.52 crawler bulldozers Power 120KW 89.14 crawler bulldozers Power 240KW 174.57 crawler bulldozers Power 320KW 234.75 wheel bulldozer Power 135KW 98.06 wheel bulldozer Power 160KW 114.40 Crawler single-bucket excavator <![CDATA[Bucket capacity 0.6m 3 > 37.45 Crawler single-bucket excavator <![CDATA[Bucket capacity 1.0m 3 > 74.91 Crawler single-bucket excavator <![CDATA[Bucket capacity 1.6m 3 > 89.89 Crawler single-bucket excavator <![CDATA[Bucket capacity 2.0m 3 > 91.93 Wheeled single-bucket excavator <![CDATA[Bucket capacity 0.2m 3 > 21.79 Wheeled single-bucket excavator <![CDATA[Bucket capacity 0.4m 3 > 29.28 Wheeled single-bucket excavator <![CDATA[Bucket capacity 0.6m 3 > 37.45 crawler loaders <![CDATA[Bucket capacity 1.5m 3 > 55.54 crawler loaders <![CDATA[Bucket capacity 2.0m 3 > 78.17

[0064] The fuel emission factor of diesel is 3186gCO2 / kg; the carbon emissions per shift of each on-site machine is equal to the energy consumption of each on-site machine multiplied by the energy emission factor. The total carbon emissions of each on-site machine are obtained by multiplying the total number of shifts collected by each on-site machine by the carbon emissions per shift of each on-site machine, thereby obtaining the carbon emissions of the machinery during the construction period.

[0065] Step S2: Collect and monitor carbon emission data during the operation and maintenance period;

[0066] Step S201: Real-time monitoring of information on moving vehicles, including speed and vehicle type, is performed. By combining the geographic information system, the monitoring data is associated with geographic information of highways to monitor carbon emissions from traffic flows in different geographic areas.

[0067] In step S201, the total carbon emissions of the vehicle are calculated using the monitoring information, where CO2 emissions (g / km) = basic fuel consumption rate (L / 100km)*[1+a+b*(|actual speed - economic speed|)+c*slope percentage]*d / 100;

[0068] a is the vehicle weight correction coefficient, b is the speed correction coefficient, c is the slope correction coefficient; d is the fuel emission factor; the vehicle weight correction coefficient for cars / SUVs is 0.4, the vehicle weight correction coefficient for trucks / buses is 0.55, the speed correction coefficient for cars / SUVs is 0.008, the speed correction coefficient for trucks / buses is 0.012, the slope correction coefficient is 0.1, the fuel emission factor for gasoline is 2320gCO2 / L, and the fuel emission factor for diesel is 2660gCO2 / L.

[0069] Total vehicle carbon emissions (g) = CO2 emissions (g / km) * driving distance (km).

[0070] Step S202: Connect with the target highway toll station data to fully understand the traffic flow characteristics of the highway driving section. The traffic flow characteristics of the highway driving section include the number of vehicles, vehicle type and mileage, so as to achieve all-round monitoring of the traffic flow carbon emission ancillary data.

[0071] Different unit-mileage carbon emission factors are matched according to vehicle types, and the unit-mileage carbon emission factors of each vehicle are multiplied by the mileage to obtain the carbon emissions of each vehicle, and the total vehicle carbon emissions are obtained based on the number of vehicles.

[0072] The method further includes step S203 of modeling and simulating the highway traffic flow, and estimating and predicting the carbon emissions of the traffic flow.

[0073] Graph Construction Layer: abstracts the traffic network into a graph structure G = (V, E, A);

[0074] V (node): represents key points in the transportation network, such as intersections, sensor locations, transportation hubs, and regional centroids;

[0075] E (edge): represents the connection relationship between nodes, such as road connection and spatial proximity;

[0076] A (adjacency matrix): encodes the spatial relationship and connection strength between nodes (can be binary adjacency, distance-based weights, flow-based weights, etc.).

[0077] Input: geographic information data, road network topology data, sensor location data.

[0078] Output: The defined graph structure G serves as the basis for subsequent graph convolution operations.

[0079] Spatio-Temporal Graph Convolution Module (ST-GCNModule): This module simultaneously captures the spatial dependencies (inter-node influences) and temporal dynamics (node-specific trends) of traffic carbon emissions data. It combines graph convolution and temporal convolution.

[0080] Structure: Spatial Graph Convolutional Layer (Spatial GCN): operates on the graph adjacency matrix A to aggregate carbon emission-related feature information of neighboring nodes. For example, graph convolution using Chebyshev polynomial approximation or adaptive adjacency matrix.

[0081] Gated Temporal Convolution: Apply one-dimensional convolution to each node along the time axis. Use a gating mechanism (such as Gated Linear Units-GLUs) or simple convolution + nonlinear activation to learn the time evolution pattern of the node itself.

[0082] Stacking: Multiple ST-GCN modules can be stacked to learn deeper spatiotemporal feature representations; each layer operates on the updated node features.

[0083] Input: graph structure G and carbon emission-related feature vectors of each node at historical time steps, such as traffic volume, speed, vehicle type composition, and weather.

[0084] Output: The high-level feature representation of each node learned at the latest time step, which integrates spatiotemporal information.

[0085] Node-Level Fusion & Pooling Layer: If you need to integrate information from different ST-GCN layers or other sources (such as global features), you can perform fusion here (such as splicing, weighted summation, attention mechanism).

[0086] Pooling: If the final prediction target is regional or global carbon emissions (rather than individual nodes), it is necessary to aggregate the features of all nodes or key node groups. Common methods include global average pooling, global maximum pooling, or attention-weighted pooling. If the prediction target is still individual nodes (such as intersections), this layer can be omitted or simply fused.

[0087] Input: High-level feature representations of all nodes output by the last ST-GCN module.

[0088] Output: A fused global / regional feature vector (if pooling is done), or still a feature vector for each node (if the goal is node-level prediction).

[0089] Fully Connected Regression Layer:

[0090] Function: Receive the fused / pooled feature vector (or the feature vectors of all nodes), learn the complex nonlinear mapping relationship between the features and the target carbon emissions, and complete the final regression prediction.

[0091] Input: Feature vector from the previous layer.

[0092] Output: predicted future carbon emissions (can be a single value or values ​​for multiple time steps).

[0093] Output Layer: Depending on the task requirements, this layer may include a linear layer (to map the output of the fully connected layer to the final prediction scale) or directly use the output of the fully connected layer to calculate the final carbon emissions prediction value.

[0094] Output: Final carbon emissions forecast.

[0095] A dynamic monitoring system for highway carbon emissions based on multi-source data fusion adopts the above method, including a construction period monitoring module and a construction period statistics module. The construction period statistics module performs statistics based on the data provided by the construction period detection module, the construction period monitoring module identifies and enters data based on the construction ledger, and the construction period statistics module performs statistics according to the statistical method of step S1 and displays it on the display module.

[0096] A dynamic monitoring system for highway carbon emissions based on multi-source data fusion also includes an operation and maintenance period monitoring module, a data docking module, a calculation module and a display module; the operation and maintenance period monitoring module uses sensors and monitoring equipment on the highway to monitor the information of traveling vehicles; the data docking module is used to dock with the data of the target highway toll station; the calculation module uses the methods in steps S201 and S202 to calculate the data provided by the operation and maintenance period monitoring module and the data docking module; the display module is used to display the calculation results of the calculation module and the statistical results of the construction period statistics module.

[0097] The prediction module also performs prediction using the method in step S203 , and the display module also displays the prediction result of the prediction module.

[0098] The display module shows complete information, which facilitates the observation of carbon emissions throughout the life cycle of the highway and provides a basis for timely adjustments.

[0099] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0100] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A dynamic monitoring method for highway carbon emissions based on multi-source data fusion, characterized in that: The following steps are involved: Step S1: Monitoring carbon emission data during the construction period; wherein: Step S101, setting monitoring targets; Step S102: Formulate a carbon emission monitoring plan; Step S103: Real-time monitoring of carbon emissions during the construction period; Step S104: establishing a data collection system; Step S2: Collect and monitor carbon emission data during the operation and maintenance period; Step S201: Real-time monitoring of vehicle information is performed. In combination with a geographic information system, the monitoring data is associated with geographic information of highways to monitor carbon emissions from traffic flows in different geographic areas. Step S202: Connect with the target highway toll station data to fully grasp the traffic flow characteristics of the highway driving section and achieve all-round monitoring of the traffic flow carbon emission data.

2. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 1, characterized in that: Step S2 also includes step S203, modeling and simulating the highway traffic flow, and estimating and predicting the carbon emissions of the traffic flow.

3. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 1, characterized in that: In step S201 , the information of the traveling vehicle includes speed and vehicle type.

4. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 3 is characterized in that: In step S201, the total carbon emissions of the vehicle are calculated using the monitoring information, where CO2 emissions (g / km) = basic fuel consumption rate (L / 100km)*[1+a+b*(|actual speed - economic speed|)+c*slope percentage]*d / 100; a is the vehicle weight correction factor, b is the speed correction factor, c is the slope correction factor; d is the fuel emission factor; Total vehicle carbon emissions (g) = CO2 emissions (g / km) * driving distance (km).

5. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 4 is characterized in that: The weight correction coefficient for sedans / SUVs is 0.4, the weight correction coefficient for trucks / buses is 0.55, the speed correction coefficient for sedans / SUVs is 0.008, the speed correction coefficient for trucks / buses is 0.012, the slope correction coefficient is 0.1, the fuel emission factor for gasoline is 2320gCO2 / L, and the fuel emission factor for diesel is 2660gCO2 / L.

6. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 1, characterized in that: In step S202 , the traffic flow characteristics of the highway driving section include the number of vehicles, vehicle types and mileage.

7. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 6, characterized in that: Different unit-mileage carbon emission factors are matched according to vehicle types, and the unit-mileage carbon emission factors of each vehicle are multiplied by the mileage to obtain the carbon emissions of each vehicle, and the total vehicle carbon emissions are obtained based on the number of vehicles.

8. The method for dynamic monitoring of highway carbon emissions based on multi-source data fusion according to claim 2, characterized in that: Step S203 includes: Graph structure construction layer: abstract the traffic network into a graph structure G = (V, E, A); V: represents the key point in the transportation network; E: represents the connection relationship between nodes; A: Encodes the spatial relationship and connection strength between nodes; Input: geographic information data, road network topology data, sensor location data; Output: The defined graph structure G serves as the basis for subsequent graph convolution operations; Spatiotemporal graph convolution module: Simultaneously captures the spatial dependencies and temporal dynamics of traffic carbon emission data; Structure: Spatial graph convolution layer: operates on the adjacency matrix A of the graph and aggregates the carbon emission-related feature information of neighboring nodes; Gated temporal convolution layer: At each node, a one-dimensional convolution is applied along the time axis; using a gating mechanism or simple convolution + nonlinear activation, the time evolution pattern of the node itself is learned; Stacking: Multiple ST-GCN modules can be stacked to learn deeper spatiotemporal feature representations; each layer operates on the updated node features; Input: graph structure G and the carbon emission-related feature vector of each node at the historical time step; Output: The high-level feature representation of each node learned at the latest time step, which integrates spatiotemporal information; Node-level feature fusion and pooling layer: If you need to integrate information from different ST-GCN layers or other sources, you can perform fusion here; Pooling: If the final prediction target is regional or global carbon emissions, it is necessary to aggregate the features of all nodes or key node groups; Common methods include global average pooling, global maximum pooling, or attention-weighted pooling; Output: The high-level feature representation of all nodes output by the last ST-GCN module; Output: A fused global / regional feature vector, or still a feature vector for each node; Fully connected regression layer: Receives fused / pooled feature vectors, learns the complex nonlinear mapping relationship between features and target carbon emissions, and completes the final regression prediction; Input: feature vector from the previous layer; Output: predicted future carbon emissions; Output layer: Depending on the task requirements, it may include a linear layer or directly use the output of the fully connected layer; calculate the final carbon emissions prediction value; Output: Final carbon emissions forecast.

9. A dynamic monitoring system for highway carbon emissions based on multi-source data fusion, characterized in that: The method according to any one of claims 1 to 8 comprises a construction period monitoring module and a construction period statistics module, wherein the construction period statistics module performs statistics based on the data provided by the construction period detection module; It also includes an operation and maintenance period monitoring module, a data docking module, a calculation module and a display module; the operation and maintenance period monitoring module is used to monitor the information of traveling vehicles; the data docking module is used to dock with the data of the target highway toll station; the calculation module is used to calculate the data provided by the operation and maintenance period monitoring module and the data docking module; the display module is used to display the calculation results of the calculation module and the statistical results of the construction period statistics module.

10. A highway carbon emission dynamic monitoring system based on multi-source data fusion as claimed in claim 9, characterized in that: It also includes a prediction module, and the display module also displays the prediction results of the prediction module.

Images